Docstoc

Disparities

Document Sample
Disparities Powered By Docstoc
					                       Disparities                               Conclusions: Differences in cancer screening discussion
                                                                 along the different dimensions of patient SES may have arisen
                                                                 because of very different mechanisms. More specifically,
                       Call for Papers                           patient education plays an important role in determining what
                                                                 happens in a clinical encounter. Physicians are distributed
   New Approaches to the Determinants of Disparities             unevenly across communities of different socioeconomic
                                                                 profiles.
         Chair: Marshall Chin, University of Chicago             Implications for Policy, Delivery, or Practice: Physicians
                                                                 need to be informed of the persistent disparities by patient
            Sunday, June 25 • 8:30 am – 10:00 am                 education in clinical communication regarding cancer
                                                                 screening and tailor their efforts to the needs of low-education
                                                                 patients. Quality-improvement efforts targeted at physicians
●Socioeconomic and Racial/ethnic Differences in the              practicing in low-income communities may also be effective in
Discussion of Cancer Screening: Treated Differently or           addressing disparities in cancer screening communication by
Treated by Different Physicians?                                 patient income.
Yuhua Bao, Ph.D., Sarah A. Fox, Ed.D., Jose Escarce, M.D.,       Primary Funding Source: NCI, AHRQ
Ph.D.
                                                                 ●Racial Disparities and Socioeconomic Status in
Presented By: Yuhua Bao, Ph.D., Assistant Professor,             Association with Survival in Older Men with
Department of Medicine, UCLA, 1100 Glendon Ave., Suite           Local/Regional Stage Prostate Cancer
2010, Los Angeles, CA 90024; Tel: (310) 794-3081; Fax: (310)     Xianglin L. Du, M.D., Ph.D., Shenying Fang, MS, Ann Coker,
794-8529; Email: yuhuabao@mednet.ucla.edu                        Ph.D., Maureen Sanderson, Ph.D., C Aragaki, J Cormier, M.D.,
Research Objective: To determine the extent to which             Y Xing, B Gor and W Chan, Ph.D.
socioeconomic and racial/ethnic differences in cancer
screening discussion between a patient and his/her primary       Presented By: Xianglin L. Du, M.D., Ph.D., Associate
care physician are due to “within-physician” differences (the    Professor, Division of Epidemiology, University of Texas
fact that patients were treated differently by the same          School of Public Health, 1200 Herman Pressler Dr, Houston,
physicians) versus “between-physician” differences (that they    TX 77030;
were treated by a different group of physicians).                Tel: (713) 500-9956; Fax: (713) 500-9264;
Study Design: Our main outcomes of interest are whether the      Email: Xianglin.L.Du@uth.tmc.edu
physician has ever talked to the patient about fecal occult      Research Objective: Whether racial differences in cancer
blood test (FOBT; for colorectal cancer screening),              survival were due to race/ethnicity, or other factors such as
mammogram (for breast cancer screening, female patients          socioeconomic status and treatment, remained poorly
only) and the Prostate-specific Antigen Test (PSA, male          explored. This study was to determine the role of
patients only). We consider five racial/ethnic groups: non-      socioeconomic factors in racial/ethnic disparities in survival
Hispanic white, non-Hispanic black, Hispanic, Asian and          among a large community-based cohort of older men
other race/ethnicity. We measure socioeconomic status by         diagnosed with prostate cancer.
both income and education. For each type of cancer screening     Study Design: This was a large retrospective cohort study
discussion, we first estimate a probit model that includes       with up to 11 years of follow-up in older men diagnosed with
patient characteristics as the only covariates to assess the     incident prostate cancer between 1992 and 1999 in the 11
overall differences. We then add physician fixed effects to      SEER (Surveillance, Epidemiology and End Results) areas in
derive estimates of “within-“ vs. “between-“ physician           the United States, identified from the SEER-Medicare linked
differences.                                                     data. These linked data not only provided reliable information
Population Studied: We use data from the baseline patient        on tumor stage and grade (Gleason score) at diagnosis and
and physician surveys of two community trials from the           long-term follow-up of the vital status, but also allowed the
Communication in Medical Care (CMC) research series. The         examination of various socioeconomic factors at the level of
two studies combined provide an analysis sample of 5,978         census tract and zip codes. In addition, comorbid conditions
patients ages 50-80 nested within 191 primary care physicians    and the type of treatment (chemotherapy and hormonal
who practiced throughout Southern California.                    therapy) can be uniquely identified from Medicare claims data.
Principal Findings: There was a strong education gradient in     Population Studied: We studied a cohort of 61,228 men
the discussion of all three types of cancer screening and most   diagnosed with local or regional stage prostate cancer at age
of the education differences arose within physicians.            65 years or older in 1992-99 in the 11 SEER areas (the
Disparities by income were less consistent across different      metropolitan areas of San Francisco/Oakland, Detroit, Atlanta,
screening methods, but seemed to have arisen mainly              and Seattle; Los Angeles county; the San Jose-Monterey area;
because of “between- physician” differences. Asians were         and the states of Connecticut, Iowa, New Mexico, Utah and
much less likely, compared to whites, to have received           Hawaii). Of the 61,228 men, 53,764 were Caucasians (non-
discussion about FOBT and PSA and these differences were         Hispanic whites), 6,321 African Americans (non-Hispanic
mainly “within-physician” differences. Black female patients,    blacks), and 1143 Hispanics.
however, were much more likely, compared to whites treated       Principal Findings: Low socioeconomic status was
by the same physicians, to have discussed mammogram with         significantly associated with decreasing survival in all men
their physicians.                                                with prostate cancer. Those living in the community with the
                                                                 lowest quartile of socioeconomic status were 31% more likely
to die than those living in the highest quartile (hazard ratio of   and whose parents continued to participate in the PSID in
all-cause mortality: 1.31, 95% confidence interval: 1.25-1.36)      2001. Data analyzed are from round 2 of the CDS and were
after adjustment for patient age, comorbidity, Gleason score,       collected in 2002 and 2003.
and the type of treatment. Hazard ratio remained almost             Principal Findings: Construct validity of the SEP measure is
unchanged after controlling for race/ethnicity (1.32, 1.26-1.38).   strong. Large SEP and race/ethnicity gradients were found
Compared to Caucasian men, African Americans were 14%               descriptively, but were reduced in the multivariate models.
more likely to die and Hispanic men were 15% less likely to         Compared to children in the highest SEP group, children in
die, after controlling for age, tumor grade, Comorbidity, and       the two lowest classes were 10 and 20 percentage points less
type of treatment. When additionally controlling for the level of   likely to be in excellent health, exhibited 2 and 3 more behavior
poverty in the community or income, there was no significant        problems, and scored up to 9 and 15 points lower on
difference in all-cause mortality between Caucasian and             standardized achievement tests, respectively after adjustment
African American men. For example, after controlling for            for race/ethnicity and other factors. Few racial/ethnic
income, the hazard ratio in African Americans was no longer         differences remained in multivariate models but Hispanic
significantly higher (1.02, 0.98-1.08) compared to Caucasians,      children were in worse health and black, Hispanic, and other
whereas Hispanics still had a significantly lower risk (0.79,       race children had achievement scores that were 10 points
0.71-0.88).                                                         lower than white/non-Hispanic children. Simulations of the
Conclusions: Racial disparities in survival among men with          contribution of reducing SEP disparities and race/ethnicity
local or regional prostate cancer were largely explained by         disparities to population health indicated that addressing SEP
socioeconomic status and other factors. Lower socioeconomic         disparities would result in greater improvements in population
status appeared to be one of the major barriers to achieving        health than would racial disparities, but that addressing both
comparable outcomes for men with prostate cancer.                   would generate the greatest improvement.
Implications for Policy, Delivery, or Practice: These findings      Conclusions: This analysis reports SEP gradients and
have important public health implications if we are to achieve      race/ethnicity disparities for children across a wide range of
the goal of the Healthy People 2010, which is to eliminate          child health and functioning outcomes. The gradients and
health disparities among different segments of the population.      disparities apparent in both psychosocial and cognitive
Further studies may be needed to address whether this               functioning in children may be the origin of adult disparities in
association is true in younger men with prostate cancer.            health if they place children, as they transition into adulthood,
Primary Funding Source: No Funding                                  in social and work environments that are more likely to
                                                                    threaten health.
●Socio-Economic Position, Race, or SEP and Race:                    Implications for Policy, Delivery, or Practice: The exclusive
Understanding Disparities in Children’s Health,                     policy focus on reducing racial and ethnic disparities is
Psychosocial Well-Being and Cognitive Functioning                   unlikely to be successful. Unless the SEP gradient in health is
Lisa Dubay, Sc.M.                                                   reduced, racial gaps will persist and SEP gradients cannot be
                                                                    eliminated without closing the race gap. Eliminating SEP
Presented By: Lisa Dubay, Sc.M., Principal Research                 gradients would have a larger impact on population health
Associate, Health Policy Center, The Urban Institute, 2100 M        than reducing race disparities, but ameliorating both would
Street, NW, Washington, DC 20037; Tel: (202) 261-5667; Fax:         produce the greatest improvement.
(202) 223-1149; Email: ldubay@ui.urban.org                          Primary Funding Source: No Funding
Research Objective: Objectives of the study are to adapt and
validate Britain’s new National Statistics--Socio-economic          ●Exploring the Determinants of Racial and Ethnic
Classification measure for use in the US; to use the measure        Disparities in Total Knee Arthroplasty: Health Insurance,
to document SEP gradients and racial and ethnic disparities in      Income and Assets
child health, emotional well-being, and cognitive functioning;      Amresh Hanchate, Ph.D., Yuqing Zhang, Dsc, M.P.H., MB,
and to assess the contribution of SEP and race/ethnicity to         David T. Felson, M.D., M.P.H., Arlene S. Ash, Ph.D.
disparities.
Study Design: The study uses cross-sectional data from the          Presented By: Amresh Hanchate, Ph.D., Research Associate,
second round of the Child Development Supplement of the             General Internal Medicine, Boston University School of
Panel Survey of Income Dynamics. A measure of family based          Medicine, 720, Harrison Ave, DOB, Suite 1108, Boston, MA
SEP is developed using parents’ employment status and               02118; Tel: (617) 638 8889; Fax: (617) 638 8026;
occupation that categorizes children in to five groups. The         Email: hanchate@bu.edu
validity of the measure is assessed using parent’s education,       Research Objective: To determine the extent to which
income, employment benefits, home ownership, and health             financial constraints – type of health insurance, income and
status. Descriptive and multivariate methods are used to            assets, are associated with the large racial/ethnic disparities in
assess the relationship between SEP, race/ethnicity, and child      total knee arthroplasty.
health outcomes. Outcomes considered include general                Study Design: The outcome measure is the dichotomous
health status, self-esteem and well-being, the presence of          response to whether respondent has Total Knee Arthroplasty
behavioral problems, and Woodcock-Johnson tests of                  (TKA) in the previous two years. Responses from 1998, 2000
achievement. Simulations of the effect on population health         and 2002 survey rounds are pooled. Longitudinal logistic
of closing the SEP gradients and racial and ethnic disparities      regression models estimated to assess association with
are estimated separately and together.                              lagged financial indicators, controlling for lagged demographic
Population Studied: The population studied includes 2907            and illness burden indicators.
children who were ages 0 to 12 at the time of the 1997 PSID
Population Studied: The data source is the longitudinal             availability of services, patients’ access to specialty physicians
Health & Retirement Study (HRS) survey rounds (1994 to              and delivery sites, provider practice styles, and referral
2002) from the Institute of Social Research, University of          patterns within health care markets may provide insights into
Michigan. Analytic sample has a nationally representative           factors that contribute to health disparities. This project
sample of 42,336 observations from 18,376 persons (ages 47          examines how the structure and financing of cardiovascular
to 109) belonging to three racial/ethnic groups – Hispanic,         care in 10 communities with large African American and
non-Hispanic black and non-Hispanic white. 8,979 (49                Hispanic residents contribute to a two-tiered system with
percent) persons are under 65 years of age. Excluding persons       disparities divided along racial and ethnic lines.
with a history of joint arthroplasty, 494 first TKAs were           Study Design: Ten communities with high numbers of
observed.                                                           African American and/or Hispanic residents were selected as
Principal Findings: Compared to white men, TKA rate is only         the focus of the study. Researchers visited delivery sites and
39% among black men and 44% among Hispanic men. The                 conducted structured interviews with 250+ informants to
TKA rates per 1,000 persons/year are: white=5.4 (CI=[4.5, 6.3]),    identify characteristics that could mitigate or exacerbate
black=2.1 (CI=[0.7,3.5]) and Hispanic=2.4 (CI=[0.5,4.3]).           disparities in cardiac care delivery and outcomes.
Among women the differences are smaller: white=6.5,                 Population Studied: Providers who serve large numbers of
black=6.9 and Hispanic=4.5. Differences in insurance                African American and Hispanic patients with cardiac
coverage, household income and assets have sizable and              conditions in ten U.S. communities.
statistically significant association with first TKA. Among those   Principal Findings: Access to hospital-based cardiac care
aged 65+, those having secondary insurance have at least 30%        differs within and across communities depending upon
higher odds of TKA compared to those with only Medicare Fee         coverage and availability of safety net providers (SNPs). We
for Service – in particular those also with Medicaid have Odds      identified three models of hospital-based cardiovascular
Ratio (OR)=1.73 (CI= [1.00-2.73] and those with Medicare            services to help explain variation in disparities: 1) Funded
HMO have OR=1.49 (CI=[1.05-2.10]). Those with household             SNPs received state/local funding to provide free or
annual income less than $20,000 or assets less than $20,000         discounted care for low-income residents, many of whom are
are at least 20% less likely to have TKA. After adjusting for the   African American or Hispanic. Funded SNPs offered access to
financial indicators, a significant reduction in residual           cardiac care that was fairly comparable across coverage
disparities among Hispanics (OR increases from 0.69 to 0.88         categories (and consequently race and ethnicity) but were
for women, and from 0.49 to 0.61 for men). Among blacks,            least likely to offer invasive and interventional cardiac care and
OR increases from 0.80 to 0.97 among women, and from 0.32           more likely to have low cardiac market share relative to other
to 0.37 among men.                                                  providers. 2) Unfunded SNPs did not receive funding for
Conclusions: While alternative explanations for racial and          indigent care and actively competed with other local hospitals
ethnic disparities have been proposed in medical literature         for Medicare and privately insured patients. At unfunded
(including patient aversion to surgery, physician behavior,         SNPs, minority and non-minority patients often accessed the
patient religiosity), this study indicates that much can be         health system through different entry points with different
learned from focusing simply on financial constraints. For          access to specialty cardiac services, largely because of
those with limited income and assets, out-of-pocket expenses        differences in insurance coverage. Of the three models,
from TKA may simply be too high given insurance coverage            unfunded SNPs were most likely to develop a two-tiered
gaps. Further investigation is needed to determine the cause        system of care based on coverage but resulting in different
of such gaps, especially among the poor. How many with such         experiences by race and ethnicity. 3) Non-SNPs were key
gaps are eligible for Medicaid, and how many are eligible but       providers of cardiac services in their community, attracting a
do not enroll?                                                      strong payer mix within cardiac service lines. Access to a full
Implications for Policy, Delivery, or Practice: This study          array of heart services was generally available for minority
indicates that health insurance gaps, especially among those        patients with Medicare and commercial coverage but much
with low incomes and assets, are key mediators in the low           less available for minority patients without these resources.
usage of TKA among blacks and Hispanics.                            Conclusions: The timely delivery of hospital-based
Primary Funding Source: No Funding                                  cardiovascular services is largely determined by organizational
                                                                    decisions that favor commercially insured or Medicare
●Disparities in Cardiac Care: Are We Developing a Two-              patients. To the extent that African American and Hispanic
Tiered System                                                       patients are overrepresented in their communities in
Marsha Regenstein, Ph.D., MCP, Holly Mead, Ph.D., Anthony           uninsured and Medicaid populations, access to care is
Lara, M.H.S.A., Soeurette Cyrprien, Jennifer Trott, Bruce           segmented along racial and ethnic lines.
Siegel, M.D., M.P.H.                                                Implications for Policy, Delivery, or Practice: Disparities in
                                                                    cardiovascular care should be discussed within the context of
Presented By: Marsha Regenstein, Ph.D., MCP, Associate              health insurance coverage and access to health services.
Research Professor, Health Policy, George Washington                Business decisions designed to enhance market share and
University Medical Center, 2021 K Street, NW, Suite 800,            attract a strong payer mix can have unintended consequences
Washington, DC 20006; Tel: (202) 530-2310; Fax: (202) 296-          on quality of care and can result in furthering racial and ethnic
0025; Email: marshar@gwu.edu                                        disparities.
Research Objective: Despite a vast literature establishing the      Primary Funding Source: RWJF
existence of racial and ethnic disparities in access to health
care, delivery of services, and health outcomes, the reasons for
disparities are not well understood. Analysis of the availability
                                                                   to 67% (1998) and 65% (1999) while changing little nationally
                        Call for Papers                            (55% in 1996 versus 60% in 1998 and 57% in 1999).
                                                                   Conclusions: Faced with a shift away from tight regulation
               System & Policy Approaches to                       and growing demand from hospitals to provide cardiac
             Eliminating Healthcare Disparities                    services, NJ took a novel approach that blended market
                                                                   competition with rules and incentives to improve quality and
          Chair: Helen Bursin, Agency for Healthcare               increase access to CA among underserved populations. This
                     Research and Quality                          strategy appears to have significantly increased competition
                                                                   among hospitals for CA patients, including blacks. While the
             Sunday, June 25 • 3:45 pm – 5:15 pm                   new facilities were not located in areas with large minority
                                                                   populations, the black-white disparity in CA use appears to
                                                                   have declined following reform.
●Evaluation of a State Hospital Regulatory Reform                  Implications for Policy, Delivery, or Practice: Few studies
Addressing Racial Disparities in Use of Cardiac                    have documented effective strategies for addressing long-
Angiography                                                        standing racial disparities in cardiac service utilization. In NJ,
Joel Cantor, Sc.D., Amy Tiedemann, Ph.D., Karl Kronebusch,         a strategy promoting market competition and regulatory
Ph.D., Derek DeLia, Ph.D., Ava Stanley, M.D., M.P.H.               incentives may have contributed to reduced disparity in use of
                                                                   CA. Additional analysis underway of use trends in incumbent
Presented By: Joel Cantor, Sc.D., Professor and Director,          facilities and contrasts to trends in states without reform may
Center for State Health Policy, Rutgers University, 55             shed light on the mechanisms by which disparities declined in
Commercial Ave, 3rd Floor, New Brunwsick, NJ 08901; Tel:           NJ.
(732) 932-4653; Fax: (732) 932-0069;                               Primary Funding Source: AHRQ
Email: jcantor@ifh.rutgers.edu
Research Objective: Through the mid-1990s NJ used
Certificate of Need (CoN) to limit diagnostic cardiac              ●Improving Chronic Disease Management for
angiography (CA) capacity. Regulatory reforms beginning in         Underserved Populations: A National Evaluation of The
1997 greatly increased CA capacity, imposed strict minimum         HRSA Health Disparities
volume requirements, required outreach to underserved              Bruce Landon, M.D., M.B.A., Leroi Hicks, M.D., M.P.H., A.
groups, and created incentives for hospitals seeking cardiac       James O'Malley, Ph.D., Tracy Lieu, M.D., M.P.H., Barbara J.
surgery licensure to address CA disparities. The objective of      McNeil, M.D., Ph.D., Edward Guadagnoli, Ph.D.
this study is to evaluate the effect of these reforms on the
black-white CA use disparity.                                      Presented By: Bruce Landon, M.D., M.B.A., Associate
Study Design: Reform is evaluated with document review, key        Professor of Health Care Policy, Department of Health Care
informant interviews, and analysis of utilization trends. Trend    Policy, Harvard Medical School, 180 Longwood Avenue,
data are from hospital billing records (NJ) and the National       Boston, MA 02115; Tel: 617-432-3456;
Hospital Discharge Survey (US). Quantitative analysis focuses      Email: landon@hcp.med.harvard.edu
on relative use of CA, defined as use by blacks per 10,000         Research Objective: Significant differences in quality
divided by the corresponding number for whites. Ongoing            according to race and socioeconomic status are problematic
work will contrast NJ trends with neighboring states, expand       in the US health care system today. While many studies have
the time series, and adjust for demographic changes.               documented these problems, few have evaluated interventions
Population Studied: NJ and US patients receiving CA in             designed to improve upon current performance. In 1998, the
hospitals per population by race.                                  Bureau of Primary Health Care of the Health Services and
Principal Findings: According to NJ informants, the main           Rescources Administration (HRSA) initiated the Health
drivers of CA regulatory reform were growing demand from           Disparities Collaboratives to reduce health disparities and
hospitals seeking licensure for lucrative cardiac procedures       improve the quality of care in Community Health Centers
and an ideological shift in favor of competition over              (CHCs) that collectively serve 15 Million poor and underserved
regulation. NJ policymakers forged a compromise between            patients. Over 700 CHCs have participated in at least one
pro-regulatory and pro-competitive forces by linking CA CoN        Health Disparities Collaborative to date. The objective of this
approvals to new access and quality requirements.                  study is to evaluate the effectiveness of the Health Disparities
Stakeholders reported little faith in the new access regulations   Collaboratives in improving quality of care for chronic medical
but most believed that increasing capacity and linking             conditions.
disparity-reduction to licensure for cardiac surgery would         Study Design: We performed a controlled pre/post
improve CA access. The number of NJ hospitals providing CA         intervention study of CHCs participating in a quality
roughly doubled after reform (from 26 in 1991-1996 to 52 in        improvement collaborative for Diabetes, Asthma, or
1998 and 49 in 1999-2002). Statewide CA volume grew both           Cardiovascular disease (with our assessment focused on
before and after reform, but volume in "incumbent" (licensed       hypertension) that were conducted during 2000-2002 by
before reform) facilities declined following reform (1997-98)      HRSA and the Institute for Health Care Improvement. The
and grew only slowly after 1998. Hospitals with new CoNs to        collaboratives use rapid cycle improvement techniques based
provide CA were not located in markets with as many minority       on "plan, do, study, act" cycles and the Chronic Care Model.
patients as incumbent hospitals. Still, preliminary evidence       Quality of care measures involving processes and
suggests that the black-white disparity may have declined          intermediate outcomes of care (e.g., control of hypertension)
following reform. In the years spanning reform, relative use of    were abstracted from medical records, and an overall
CA by blacks compared to whites in NJ grew from 52% (1996)         composite measure was created for each condition by taking
the mean of all measures (including both process and              Study Design: An extensive qualitative study consisting of 10
outcomes measures) after normalization of each of the             case studies of U.S. hospitals conducted between September
individauls scores. Changes in quality were evaluated using       2005 and January 2006. Case studies included key informant
hierarchical logistic regression models that controlled for       interviews, surveys, and the collection of organizational
patient characteristics.                                          documents to investigate readiness to change in hospitals
Population Studied: We enrolled 44 CHCs participating in a        committed to address the issues of reducing disparities and
collaborative (13 for asthma, 17 for diabetes, and 14 for         improving cardiovascular care for minority Americans. First,
hypertension) and 20 non-participating centers (“external”        hour- and half-hour-long in-person and telephone interviews
controls) that had not participated in a collaborative after      were taped, transcribed, and analyzed used deductive and
matching on region, size, and location. Each participating        inductive methods, including a grounded theory approach to
clinic also served as a control (“internal” control) for one of   explore emergent themes. Second, a 130-item mailed and
the other targeted conditions. From each center, random           electronic survey asked about organizational change, racial
samples of eligible patients from the one year periods before     and ethnic disparities, cultural competency, and related topics.
and after the intervention were selected for chart review.        Survey data analyses included descriptive statistics, chi-
Principal Findings: We studied 11,153 patients with one of the    squared analyses, linear regression, and logistic regression.
three target conditions in the experimental and control groups    Organizational change theory provided a conceptual
(3,887 with asthma, 2,904 with diabetes, and 3,362 with           framework for the study.
hypertension). Intervention clinics showed significant            Population Studied: Case study organizations included
improvement in the overall composite measure of quality           academic medical centers, community hospital systems, and
when compared to both external controls and internal controls     free-standing hospitals. Key informant interviews were held
for asthma and diabetes but not for hypertension. For             with over 200 organizational representatives and physicians.
instance, participating asthma centers improved the overall       Survey respondents included physicians and hospital
percentage met for the composite quality measure by 15% as        employees from all areas of the organizations.
compared to 5% for internal controls and 1% for external          Principal Findings: Across hospitals, an overwhelming
controls (p<.001 for both comparisons). Individual measures       majority of respondents reported that while they believed that
in the areas of screening, treatment, and self-management         disparities are a nationwide problem, they were skeptical that
such as asthma severity assessment, hemoglobin A1C                the care provided by their own hospital was an issue. Further,
assessment, and diabetic foot exams showed significant            despite the fact that all hospitals had applied to participate in
improvement. There was no improvement, however, in                this improvement initiative, there was considerable variation
intermediate outcomes such as control of diabetes (as             among the hospitals with respect to institutional readiness to
measured by hemoglobin A1C), hypertension, or cholesterol,        change. Smaller, less complex organizations tended to be
or the number of urgent care, emergency room, or hospital         categorized at a higher stage of change readiness than larger
visits for asthma patients.                                       academic centers. Across institutions, factors such as
Conclusions: We found that the Heatlh Disparities                 executive leadership support, availability of clinical information
Collaboratives significantly improved the extent to which         systems, experience with quality improvement initiatives, and
several processes of care were followed for two of the three      transparency of data reporting appeared related to
conditions studied, without any improvement in intermediate       organizations’ abilities to both innovate and implement
outcomes.                                                         needed changes associated with addressing the issues of
Implications for Policy, Delivery, or Practice: Collaborative     disparities and quality improvement in cardiovascular care.
improvement programs continue to be a cornerstone of              Conclusions: Efforts to change institutional patterns of care
quality improvement efforts for CHCs and other health care        to reduce disparities and improve quality can be guided by
organizations. Our findings suggest that while the                organizational change and change management models that
collaboratives have successfully improved the processes of        focus program development and implementation. Health
care, improved methods are required to achieve more               care organizations attempting to address the issue of
significant improvements in intermediate outcomes that are        disparities in care can build support for their initiatives by
the most important determinants of long term chronic disease      promoting transparency in data collection and dissemination,
outcomes.                                                         and then work to support institutional change by facilitating
Primary Funding Source: AHRQ, the Commonwealth Fund               rapid cycle improvements that can be spread throughout the
                                                                  hospital.
●Improving Institutional Readiness to Change:                     Implications for Policy, Delivery, or Practice: Engaging a
Addressing Disparities and Quality Improvement in U.S.            hospital’s multiple internal and external stakeholders hospitals
Health Care Organizations                                         in efforts to reduce disparities and improve quality of care will
Ann Scheck McAlearney, Sc.D., M.S., Bruce Siegel, M.D.            raise awareness about these issues and can help increase the
                                                                  likelihood of program success. Moreover, our work suggests
Presented By: Ann Scheck McAlearney, Sc.D., M.S., Assistant       that wider adoption of health information technologies may
Professor, Health Services Management and Policy, The Ohio        help address the issue of disparities in care by providing
State University, 1841 Millikin Road, Cunz Hall, 4th Floor,       treatment guidance that is both informed and relatively
Columbus, OH 43210-1229; Tel: 614-292-0662; Fax: 614-438-         neutral with respect to patient race/ethnicity. Finally, this
6859; Email: mcalearney.1@osu.edu                                 research has produced several assessment tools (both
Research Objective: To explore institutional readiness to         qualitative and quantitative) that can be used in future
change around the issues of reducing disparities and              evaluations of institutional readiness to change around the
improving the quality of cardiovascular care in U.S. hospitals.   issues of disparities reduction and quality improvement.
Primary Funding Source: RWJF                                           white patients are being referred to CABG, thus freeing
                                                                       capacity among high quality providers.
●When Racial Disparities Disappear: Is it a Good News                  Implications for Policy, Delivery, or Practice: Minority
Story?                                                                 patients are competing with Whites for access to services and
Dana B. Mukamel, Ph.D., David L. Weimer, Ph.D., Tom                    for access to the best providers. Efforts to equalize access
Buchmueller, Ph.D., Heather Ladd, MS, Alvin I. Mushlin,                may, therefore, require increasing capacity, as well as
M.D., M.S.C.                                                           improvement in the quality of low performing providers and
                                                                       decreasing the variations in quality across providers.
Presented By: Dana B. Mukamel, Ph.D., Professor and Senior             Primary Funding Source: NIA, Commonwealth Fund and the
Fellow, Center for Health Policy Research, University of               Natioanl Center for Minority Health and health Dispairties
California, Irvine, 111 Academy, Suite 220, Irvine, CA 92697-
5800; Tel: (949)824-8873; Fax: (949)824-3388;                          ●National Respite Pilot Initiative: Evaluation Results
Email: dmukamel@uci.edu                                                Suzanne Zerger, MA, Adi Gundlapalli, M.D., Ph.D.
Research Objective: Racial disparities in access to high
quality cardiac surgeons persisted throughout the 1990s.               Presented By: Suzanne Zerger, MA, Research Specialist,
Recent data suggest that these disparities no longer exist. The        National Health Care for the Homeless Council, P.O. Box
objective of this study is to examine reasons for this change          60427, Nashville, TN 37206; Tel: 615-226-2292; Fax: 615-226-
and the resulting equalization in access to high quality care.         1656; Email: szerger@nhchc.org
Study Design: A retrospective analysis of Medicare fee-for-            Research Objective: 1) To identify and document the differing
service claims data augmented with quality measures for                models of care for the delivery of respite services; and,
cardiac surgeons obtained from the New York State Cardiac              2) To assess the effect of respite services on the health of
Surgery report and socioeconomic data from the census. The             homeless people during their stay in respite.
analyses compared referral patterns in the late 1990s to the           Study Design: This prospective evaluation of a National
early 2000s. Hierarchical multivariate regression models were          Respite Pilot Initiative used a community-based, participatory
estimated to examine the relationship between race and CABG            design. Respite Coordinators from each of the 10 respite pilot
surgery performed by a high quality surgeon, while controlling         programs worked collaboratively with a research team to
for patient risk factors, physician and hospital fixed effects,        design all aspects of this study, including specific variables
and heteroscedasticity. The change in the average quality of           and methodologies. Program-level data (pre-post surveys and
surgeons in the two periods for Whites and Blacks was                  interviews with Respite Coordinators) describe the service
decomposed into two components: 1. change due to shifts in             models in detail and the pros and cons associated with each,
patients’ referrals to surgeons of better quality, holding             lessons learned about implementing respite programs, and
surgeon quality constant; and 2. change due to improvement             anecdotal data on success stories and ethical dilemmas
in surgeons’ quality, holding patients’ referrals constant.            confronted in implementing these programs. Client-level data
Multivariate regression models were then estimated to identify         were obtained on all consenting clients (n=1349; 97%
factors associated with the two types of change minority               response rate) and the respite services they received during
patient referrals and surgeon quality. Finally, multinomial            their stay (n=1507 encounters) from each of the 10 respite
models were estimated to examine the use of CABG versus                programs. These data, collected between July 2003-March
angioplasty by race in order to investigate the change in              2005, include: referral sources; demographic characteristics;
capacity of CABG surgeons and its impact on access.                    hospitalization and treatment histories; diagnoses (and
Population Studied: All of the 227,405 New York State                  severity of each) made upon admission, during the respite
Medicare fee-for-service patients with cardiac procedures:             stay and upon discharge; all encounters and treatments
angiography, angioplasty or CABG during 1997-1999 and                  during the respite stay (type and number); pre-post outcome
2001-2003.                                                             measures such as housing, income, and family status; and
Principal Findings: In contrast to findings from the 1990s, in         reasons for exiting the respite program. In addition, data
the early 2000s Blacks no longer experience disparities in             were collected on all individuals who were referred to the
access to high quality surgeons. The elimination of this               respite programs but were not admitted – including their
disparity is traced to both a shift in referrals of Blacks to better   diagnoses and the reasons they could not be admitted.
surgeons and an improvement in the quality of the surgeons             Population Studied: 1349 adult homeless persons accessing
that they accessed historically,p<0.001. About two-thirds of           respite services from 10 programs across the U.S.
the overall change is due to shift in referrals and one-third is       Principal Findings: Most of the individuals these respite
due to improvement in surgeons’ quality. Multivariate                  programs serve have multiple, severe, and complex needs,
analyses suggest that this is explained by the shift from CABG         including long histories with homelessness, multiple
to angioplasty, which opens capacity of the higher quality             diagnoses of high severity, and psycho-social needs.
CABG surgeons, p=0.019 and allows them to accept minority              Respite care improves quality of life. This evaluation shows
patients who may not have had access to them in the past.              clients’ improved not only in health status, but also in other
Other surgeon and patient characteristics are not associated           areas critical to their overall health such as access to health
with this phenomenon. The number of CABG surgeries                     care, health insurance, income, and housing. The need for
performed on Whites has declined between the two periods               respite care is vast, and remains largely unmet. All of these
while the number for Blacks has increased. The number of               programs are struggling to manage rapidly growing need with
angioplasty procedures increased for both races.                       already limited resources. Several of these programs routinely
Conclusions: Blacks seem to have gained increased access to            are unable to admit half to two-thirds of the individuals
high quality cardiac surgeons in recent years because fewer            referred to them for care.
Conclusions: These respite programs are filling a gap in             making of patients and patient perceptions of negative health
services between those hospitals and clinics unable to serve         care encounters. For example, consumers with very low
these clients’ needs and the harsh realities of the street.          expectations of the health system could interpret inadequate
These programs have a unique opportunity to assist                   care or unfair treatment as good care. If expectations for care
individuals not only in addressing their acute health needs,         are set too low, then low quality care may become acceptable
but in helping them improve their overall health and the             or worse un-noticeable to vulnerable consumers. Objective: In
quality of their lives. To be effective in the long-term, however,   this study, expectations represent normative predictions about
the necessary housing, treatment programs, and services need         what the consumer anticipates will happen in their encounters
to be available for clients when they are discharged from            with physicians and health plans. This “normative” dimension
respite.                                                             is a function of the social expectations for physician behavior
Implications for Policy, Delivery, or Practice: An increasing        and the health care industry in general. Thus, the purpose of
managed care environment in hospitals is resulting in shorter        this study was to determine whether minority consumers had
hospital stays and more procedures provided on an outpatient         different evaluations of how well physicians and health plans
basis. The impact on homeless individuals is especially harsh,       meet the social norms/ role expectations for medical
because they are frequently discharged from the hospital with        professionalism and patient trust.
prescriptions for mediation they cannot afford to have filled,       Study Design: Methods: Data from the Yale/ NYAM
and/or with instructions for follow-up care they are unable to       Consumer Trust in Health Care Survey, a nationally
heed, such as bed rest and nutritious food. One response to          representative telephone survey (n=5000), were analyzed
this gap in health care between hospitals and the streets has        using logistic regression. The role expectations included
been the development of medical respite services for                 patient advocacy, caring, racial fairness, honesty and technical
homeless persons. Though respite programs for homeless               competence of physicians and health plans.
persons in the U.S. were first developed in 1985, the increasing     Principal Findings: Compared to whites, Blacks were more
pressure on clinics and hospitals to discharge people quickly        likely to expect limited patient advocacy by physicians even
has resulted in a sharp increase in the number of respite            after adjustment for age, health status and other variables
programs in just the past few years, and even the creation of a      (OR=1.69). Blacks were more likely than whites to endorse the
National Respite Providers Network last year. This National          belief that physicians do not provide fair treatment to all of
Respite Pilot Initiative evaluation provides timely data that are    their patients regardless of race (OR=3.04). Latinos also
greatly needed, including detailed descriptions of models of         anticipated limited racial fairness in the medical care provided
respite care that work, insights into program development and        by physicians (OR=1.54). Contrary to the racial fairness
design, and an understanding of the needs of individuals             findings, Latino and black respondents were less likely to
most likely to require, access, and benefit from these unique        expect uncaring and impersonal treatment among physicians
services.                                                            than whites. Interestingly, no race effects for blacks or Latinos
Primary Funding Source: HRSA                                         were observed for the expectations about the technical
                                                                     competence of physicians despite strong race effects on other
                        Call for Papers                              physician expectations. Blacks and Latinos shared the same
                                                                     belief that health plans were not likely to treat all consumers
           The Effect of Patient & Community on                      fairly regardless of race (OR=2.68, 1.58). Asian/other
                   Healthcare Diparities                             consumers were more likely to expect dishonesty in health
                                                                     plan practices (OR=1.79) compared to whites.
         Chair: Lisa Cooper, Johns Hopkins University                Conclusions: There are significant racial differences in
                                                                     consumer expectations of physicians and health plans. The
              Sunday, June 25 • 5:45 pm – 7:15 pm                    lowered expectations of minority consumers may be indicative
                                                                     of the responsiveness of the health system to their current and
                                                                     past health needs. Results are also discussed in terms of how
●Disparities Start at the Door: Racial Differences in                concurrent health care experiences may influence the updating
Consumer Expectations of Physicians and Health Plans                 of consumer expectations.
Lisa C. Gary, Ph.D., M.P.H., Mark J. Schlesinger, Ph.D.              Implications for Policy, Delivery, or Practice: This study
                                                                     shows that measuring and controlling for expectations is
Presented By: Lisa C. Gary, Ph.D., M.P.H., Assistant                 methodologically important especially when using patient-
Professor, Health Care Organization and Policy, UAB School           reported outcomes. The integration of expectations questions
of Public Health, 1530 3rd Ave. South (RPHB 330),                    into patient satisfaction questionnaires is the focus of future
Birmingham, AL 35294-0022; Tel: 205-975-0506;                        research.
Email: Lgary@uab.edu                                                 Primary Funding Source: No Funding
Research Objective: Background: Barriers to healthcare
access and evidence of lower quality of healthcare have been
documented for Blacks and other minorities across a range of
areas including cardiac care, pain management and preventive
care. Despite these findings, research has shown that Blacks
often have high levels of patient satisfaction. The basis for this
conundrum may in part be explained by racial variations in
consumer expectations about the health system. Exploring
the nature of racial differences in expectations is especially
salient as these differences may affect future clinical decision
●Effects of Medical Mistrust and Perceived Discrimination           Implications for Policy, Delivery, or Practice: By treating
in the Medical Community on Self-management Practices               patient's fairly and earning the personal trust of their patients,
for African American and Hispanic Patients with                     health care providers may overcome the potentially deleterious
Uncontrolled Hypertension                                           effects of mistrust/perceived racism that some minorities
Paul Hebert, Ph.D., Jodi Casabianca, MS, Leah Tuzzio,               report in regards to the medical establishment in general.
M.P.H., Mary Ann McLaughlin, M.D., M.P.H., Christine
Brown, M.P.H., Jane E. Sisk, Ph.D.                                  Primary Funding Source: NIH's National Center on Minority
                                                                    Health and Health Disparities
Presented By: Paul Hebert, Ph.D., Assistant Professor, Health
Policy, Mt. Sinai School of Medicine, One Gustave L. Levy           ●Chronic Disease as Contagion: The Effect of Social
Place, Box 1077, New York, NY 10029; Tel: (212) 659-9191; Fax:      Influence and Grocery Store Locations on Body Mass
(212) 423-2998; Email: paul.hebert@mssm.edu                         Index
Research Objective: Previous studies have shown that                Sanae Inagami, M.D., MPH, Deborah A. Cohen, M.D., MPH,
patients' perceived discrimination in medicine and medical          Henry D. Anaya, Ph.D., Brian K. Finch, Ph.D., Steven M. Asch,
mistrust are correlated with race and contributes to lower          M.D., MPH
satisfaction with care among African Americans. The purpose
of this study is to explore whether these perceptions interfere     Presented By: Sanae Inagami, M.D., MPH, Fellow/Research
with the provider-patient relationship in a group of African        Associate, GIM/HSR, UCLA/VA Greater Los Angeles Health
American and Hispanic patients with uncontrolled                    Care System, 11301 Wilshire Blvd (111G), Los Angeles, CA
hypertension, leading to lower patient knowledge of                 90025; Tel: 310-478-3711x40385; Fax: 310-268-4933;
hypertension and worse self-reported self-care.                     Email: sinagami@ucla.edu
Study Design: The data come from the baseline survey of a           Research Objective: Residential neighborhood disadvantage
randomized controlled trial of disease management for               has been associated with increased body mass index (BMI) in
uncontrolled hypertension. The survey administered at               individuals, with a disproportionate number of Latinos and
enrollment included validated 5-item trust and 4-item               African-Americans affected. Individuals have higher BMI when
discrimination scales. Subjects were also asked questions           living in areas where residents frequent grocery stores located
regarding their knowledge of and behaviors toward important         in lower socioeconomic areas. (Manuscript in review.) We
self-managment issues for hypertention. Specifically, patients      hypothesized that residents in these areas with greater
were asked if they knew that exercising, weight control, and        interactions with their neighbors may have even higher BMI as
reducing sodium and alcohol, respectively, were critical to         they may be more greatly influenced by their neighbors
controlling HTN. Patients were also asked whether they were         compared to those who interact less with their neighbors. We
engaged in these preventive behaviors, or if they intended to       examined whether greater interactions and familiarity with
do so. We correlated responses regarding perceived                  neighbors influence BMI of the individual and whether these
discrimination and mistrust with knowledge and behavior             may contribute to race/ethnic differences; we also explored
toward preventive self-care.                                        other neighborhood factors that may be associated with
Population Studied: In 2004/5, we enrolled 416 African              greater increased interactions.
American and Hispanic patients with blood pressure >140/90          Study Design: We linked the 2000 US Census data with the
from the four hospital clinics and one community health clinic      Los Angeles Family and Neighborhood Study (L.A.FANS)
in Harlem, New York City. Patients were 59% Non-Hispanic            database, which consists of 2620 adults sampled from 65
Black, 41% Hispanic, 71% female, with mean age 60.7 (range          neighborhoods in Los Angeles County. “Neighbor familiarity
33 to 94). The mean blood pressure at baseline was 153/86,          and interactions” was assessed using a five-item scale related
and 58% of the enrollees were obese.                                to the respondent's personal association with his/her
Principal Findings: Perceptions of racism and mistrust were         neighbors. Multilevel linear regressions were used to estimate
varied: 53% (95% c.i.: 0.47, 0.58) agreed that racial               associations between BMI and “neighbor familiarity” and
discrimination in the doctor’s office is common while 26%           quality of grocery stores after adjustment for individual level
(95% c.i.: 0.21, 0.31) agreed that doctors treat minority and       factors (gender, education, age, income, car ownership,
white people differently. When asked if they believed patients      race/ethnicity, immigrant status, marital status, employment)
are misled at hospitals, 67% (95% c.i.: 0.61, 0.72) agreed.         and socioeconomic characteristics of residential
Only 8% (95% c.i.: 0.05, 0.11) responded that they personally       neighborhoods.
had been treated unfairly because of race and 95% (95% c.i.:        Principal Findings: Individual perception of neighborhood
0.93, 0.97) said they trusted their own physician. We found no      safety and sense of neighborhood social cohesion are
significant correlations between mistrust/discrimination and        inversely related to whether individuals interact with their
knowledge of or adherence to diet, exercise, sodium, and            neighbors. Greater interaction with neighbors is associated
alcohol recommendations.                                            with increased BMI (p=0.012). Native Latinos and African-
Conclusions: Perceived racism and mistrust though present           Americans who interact less with neighbors, who frequent
is not correlated with patient knowledge of or adherence to         grocery stores located in lower SES neighborhoods, have lower
important preventive behaviors. A possible reason for this          BMI than would be expected (p=0.042,p=0.002) from the
may be the trust that patients feel for their own doctor, and       combined independent effects of each. Immigrant Latinos and
their belief that they are treated fairly. Patients in this study   African-Americans who travel farther for their groceries have
followed the advice of their trusted providers, even if they        lower BMI (-1.33 BMI units, p=0.01; -1.80 BMI units, p=0.04).
perceived discrimination and mistrust in the larger medical         Native Latinos and African-Americans who are more familiar
community.                                                          with their neighbors, who frequent grocery stores in lower SES
areas, had higher BMI than would be expected from the              and other single/multiple races (OR 2.25). A one unit increase
combined independent effects of each.                              in the dissatisfaction scale at the mean (1.98) was associated
Conclusions: We found complex associations between social          with nearly a 30% increased odds of reporting perceived
processes and BMI that may serve to explain why Latinos and        discrimination. Including the dissatisfaction measure as a
African-Americans are disproportionately affected by obesity in    covariate increased the odds of reporting discrimination for
Los Angeles. Energy balance behaviors, such as where people        each of the groups. In the health outcome models,
shop, what people cook, and what people eat may be the             discrimination was consistently and robustly (p<0.001)
result of neighborhood social influence. The roots of obesity      associated with poorer health. Dissatisfaction was significant
and chronic disease may lie in social patterning of health         at p<0.001 for each outcome except chronic conditions. For
behavior, and identifying how health is influenced by the          each outcome, inclusion of dissatisfaction attenuated the
individual's perception of and reaction to their neighborhood      effect of discrimination: poorer general health status (from
(interacting or not interacting with neighbors) may be             OR 1.56 to 1.32), number of chronic conditions (beta 0.059 to
important in stemming the obesity epidemic.                        0.057), physically unhealthy days (beta 0.36 to 0.24), mentally
Implications for Policy, Delivery, or Practice: Health care        unhealthy days (beta 0.54 to 0.41), and activity limitation days
professionals may need to view obesity within the contagion        (beta 0.44 to 0.32). The perceived discrimination estimate
paradigm and redirect their energies to the level of the           remained significant across all health status measures, and
neighborhood in order to effect change at the level of the         was 62% to 96% larger than the dissatisfaction estimate.
individual.                                                        Conclusions: Racial and ethnic minorities report more
Primary Funding Source: HRSA, VAHSR&D                              discrimination in medical care and this perceived
                                                                   discrimination is strongly associated with poorer health status
●Self-Reported Health Disparities: Discrimination or               and worse health-related quality of life. Although patient
Dissatisfaction?                                                   dissatisfaction of health care in part explains variations in
Ninez Ponce, MPP, Ph.D., Gilbert Gee, Ph.D.                        health, discrimination still explains a larger part of health
                                                                   disparities.
Presented By: Ninez Ponce, MPP, Ph.D., Assistant Professor,        Implications for Policy, Delivery, or Practice: Standards set
Health Services, UCLA School of Public Health, 31-254B Center      for providers to treat patients equally and fairly are deserving
for the Health Sciences, Los Angeles, CA 90095-1772; Tel: 310-     of the same credence given to patient satisfaction ratings as
206-4021; Email: nponce@ucla.edu                                   benchmarks for improving the quality of care.
Research Objective: Racial discrimination in medical care has      Primary Funding Source: NCI, National Institute of Aging
been identified as a source of disparities in health status. But
patient dissatisfaction with care may also influence the           ●Do Ethnic Enclaves Contribute to the Disparities in
perception of discrimination. We explore the extent of racial      Cancer Screening?
differences in reporting discrimination, and examine the           Ninez Ponce, MPP, Ph.D.
independent effects of discrimination and dissatisfaction on
measures of self-reported health status and health-related         Presented By: Ninez Ponce, MPP, Ph.D., Assistant Professor,
quality of life measures.                                          Health Services, UCLA School of Public Health, 31-254 B
Study Design: Perceived discrimination was specified as a          Center for the Health Sciences, Los Angeles, CA 90095-1772;
dichotomous outcome (1= reporting that one would get better        Tel: 310-206-4021; Email: nponce@ucla.edu
medical care if they belonged to a different racial or ethnic      Research Objective: The study focuses on the effects of
group). Patient dissatisfaction of medical care was a              ethnic enclaves in California and tests whether ethnic
continuous variable from 0=least dissatisfied and 10=most          concentration in neighborhoods mitigates or promotes receipt
dissatisfied. Covariates included race/ethnicity, age, income as   of cancer screening tests at the recommended intervals.
a percent of poverty level, education, gender, rural/urban         Study Design: Two-level hierarchical logit models, stratified
residence, family composition and health insurance. We             by race/ethnicity, on 1) Pap tests, 2) mammograms, and 3)
tested the effects of perceived discrimination and                 colorectal cancer tests. At the census tract level (Level-2) we
dissatisfaction on several health outcomes: general health         measured racial/ethnic concentration, social capital,
status, count of chronic conditions, physically unhealthy days,    neighborhood tenure, %population living below poverty,
mentally unhealthy days, and activity limitation days. We fit      %non-citizen, and %limited English proficient; at the
logit models on dichotomous outcomes, ordered logit on             metropolitan-statistical area (MSA) level, we included HMO
general health. Chronic conditions were estimated as OLS           penetration, HMO competition, and %staff/group model
and physically and mentally unhealthy days and activity            HMO. We adjusted for individual characteristics: age, gender,
limitation days were estimated as negative binomial models.        household income, education, family composition, years lived
We used survey weights and jackknifed standard errors to           in the US, citizenship, English proficiency, neighborhood
account for the 2003 CHIS multi-stage survey design.               tenure, individual social capital index, discrimination in
Population Studied: Approximately 39,000 adults who                medical care, usual source of care, and health insurance
reported a medical encounter in the past two years from the        coverage.
population-based 2003 California Health Interview Survey.          Population Studied: Appropriately aged adults, depending on
Principal Findings: Racial and ethnic minority groups              type of cancer screening outcome, residing in metropolitan
significantly reported higher odds of perceived discrimination     areas from the population-based 2003 California Health
than non-Latino whites (p<0.001 for each group): Latinos           Interview Survey (n=42,044). Level-2 data obtained from the
(OR 2.23), Asians (OR 1.82), African Americans (OR 4.03),          2003 CHIS, 2000 US Census, and Interstudy 1999-2000.
American Indians/Alaska Natives (OR 2.87), Pacific Islanders
Principal Findings: Asian concentration decreased Pap test
screening (estimate: -1.040, p-value: <0.001) and breast                                   Call for Papers
cancer screening among Asians (estimate: -0.577, p-value:
0.022). Latino concentration decreased likelihood of any CRC          Next Generation Measures & Outcomes of Disparities
screening among Latinos (estimate: -0.007, p-value: <0.001)
and endoscopy in the past 5 years (estimate: -0.008; p-value:                 Chair: Ernest Moy, Agency for Healthcare
<0.001). However, white concentration improved the                                      Research and Quality
likelihood of all types of cancer screenings for whites
examined (p-value <0.01 for all outcomes). After adjusting for                  Tuesday, June 27 • 8:45 am – 10:15 am
concentrated disadvantage, collective efficacy, and health care
market structure, Latino concentration still decreased the
likelihood of CRC screening (estimate: -0.008, p-value: 0.012)      ●The Impact of Racial and Ethnic Disparities in Influenza
and endoscopy (estimate: -0.008, p-value: 0.014). White             Vaccination on Elderly Minority Deaths
concentration significantly improved the likelihood of              Kevin Fiscella, M.D., M.P.H., Richard Dressler, M.D., M.P.H.,
screening for mammography (estimate: 0.325; p-value: 0.016),        Sean Meldrum, MS, Kathleen Holt, Ph.D.
any CRC test (estimate: 0.414; p-value: 0.001) and endoscopy
(estimate: 0.419; p-value: 0.001). For Asians, high Asian           Presented By: Kevin Fiscella, M.D., MPH, Associate
concentration still significantly predicted lower cervical cancer   Professor, Department of Family Medicine and Community &
screening, but not lower breast cancer screening, after             Preventive Medicine, University of Rochester, 1381 South Ave,
controlling for socio-ecological and individual level               Rochester, NY 14620; Tel: (585) 506-9484 ex 106; Fax: (585)
characteristics. Latino concentration was no longer                 473-2245; Email: kevin_fiscella@urmc.rochester.edu
significant after controlling for socio-ecological and individual   Research Objective: To estimate the annual number of
level characteristics, but linguistic concentration (% Limited      minority deaths that would be avoided if influenza vaccination
English Proficient in a neighborhood) had a significant             rates among elderly African Americans and Hispanics were
negative effect. Black concentration was not significant in any     equal to elderly white rates or if the Healthy People 2010
of the models. Among the cancer screening tests, CRC tests          influenza vaccination objective were achieved.
were the most sensitive to health care market structure: HMO        Study Design: We used age-, sex-, race-, and ethnicity-specific
penetration conferred protective effects for CRC screening          influenza 2002 data from the Medicare Current Beneficiary
among Latinos and whites.                                           Survey, national vital data, and conservative estimates of the
Conclusions: The socio-ecological context of where people           deaths attributed to influenza among vaccinated and
live have differential effects by race/ethnicity and by type of     unvaccinated elderly in our models. We calculated the
cancer screening. A high Asian and Latino concentration             difference in current African American and Hispanic influenza-
reflects a concentrated disadvantage in terms of cancer             related deaths and those modeled based on white vaccination
screening tests. In contrast, we found a protective effect for      rates. We compared these numbers to major causes of death
whites living in mostly white neighborhoods. Health care            among elderly minorities.
market variables became prominent with individual                   Population Studied: All African American and Hispanic
interactions suggesting further study to test whether health        Medicare Beneficiaries aged 65 and older.
care market structure might reduce enclave risk.                    Principal Findings: During 2002, there were 162,490 and
Implications for Policy, Delivery, or Practice: Large gains         64,580 African American and Hispanic all-cause deaths
can be made in reducing cancer screening disparities if cancer      among elderly Medicare beneficiaries. Of these, 5,890 African
control programs focus on high-density Asian and LEP Latino         American and 2,220 Hispanic deaths were attributable to
neighborhoods.                                                      influenza. If the minority influenza vaccination rates equaled
Primary Funding Source: NCI, National Institute of Aging            the white, non-Hispanic rate, there would be 2,110 fewer
                                                                    minority deaths. Achievement of the Healthy People 2010
                                                                    influenza vaccination objective for the elderly would prevent
                                                                    4,450 minority deaths annually. The former estimate
                                                                    substantially exceeds the nearly 1,500 combined deaths
                                                                    among elderly African American and Hispanics from
                                                                    homicide, suicide, and motor vehicle accidents. The latter
                                                                    estimate exceeds 3,900 deaths from 10th leading cause of
                                                                    deaths for elderly African Americans and Hispanics.
                                                                    Conclusions: Based on conservative estimates, elimination of
                                                                    racial and ethnic disparities in influenza vaccination would
                                                                    significantly reduce mortality among elderly African Americans
                                                                    and Hispanics.
                                                                    Implications for Policy, Delivery, or Practice: Given the
                                                                    substantial impact of these disparities on minority deaths,
                                                                    elimination of racial and ethnic disparities in influenza
                                                                    vaccination and achievement of the Healthy People 2010
                                                                    objective should become a national priority.
                                                                    Primary Funding Source: RWJF
●Influenza Vaccination Trends Since 1997: Are We Making               to identify the factors that are driving the white-black
Progress Toward 2010 Goals?                                           differences rates, which may include differences in care-
Sharon K. Long, Ph.D., Genevieve Kenney, Ph.D.                        seeking behavior, in access to providers, or in provider
                                                                      behavior and treatment patterns.
Presented By: Sharon K. Long, Ph.D., Principal Research               Implications for Policy, Delivery, or Practice: The
Associate, , Urban Institute, 2100 M Street NW, Washington,           persistence of low vaccination levels among racial and ethnic
DC 20037; Tel: 202.261.5656; Fax: 202.223.1149;                       minorities suggests that these populations would be
Email: slong@ui.urban.org                                             disproportionately affected by an influenza pandemic. Greater
Research Objective: This study monitors progress toward the           investments in outreach and education and more effective
Healthy People 2010 objectives of raising influenza                   targeting is important because of the complications and costs
vaccination rates among adults in high priority groups and            associated with influenza, particularly in light of a possible
eliminating racial/ethnic disparities. We also examine the            influenza pandemic.
extent to which differences in the characteristics of racial and      Primary Funding Source: The Urban Institute
ethnic minorities explain the racial and ethnic differences in
vaccination rates.                                                    ●Race, Income, and Operative Time
Study Design: We rely on data from in-person surveys                  Jeffrey H. Silber, M.D., Ph.D., Paul R. Rosenbaum, Ph.D.,
collected from 1997 to 2004 as part of the National Health            Xuemei Zhang, MS, Orit Even-Shoshan, MS
Interview Survey, a nationally representative survey of the
civilian non-institutionalized population. Three key subgroups        Presented By: Jeffrey H. Silber, M.D., Ph.D., Professor of
are examined: black non-Hispanic, white non-Hispanic and              Pediatrics and Director, Center for Outcomes Research, The
Hispanic adults. We examine two important high priority               University of Pennsylvania/The Children's Hospital of
groups: adults aged 65 and older and adults under age 65 with         Philadelphia, 3535 Market Street, Suite 1029, Philadelphia, PA
a chronic health condition. We compare targeting                      19104; Tel: 215-590-5635; Fax: 215-590-2378; Email:
effectiveness across the different race and ethnic groups in          Silberj@Wharton.UPENN.edu
2004 relative to that achieved in the seven previous years to         Research Objective: Operative time is often included in
assess how well the vaccine is being targeted to the high             studies of quality and efficiency, and is well known to influence
priority groups and whether that has changed over this time           the rate of infection. However, operative time is rarely studied
period. To better understand the factors that affect receipt of       across institutions because of the need for costly chart review.
an influenza vaccination, we estimate multivariate regression         This paper describes how well the anesthesia claim from
models. The multivariate analyses control for individual              Medicare can estimate chart times, and, using this model,
characteristics such as age, education, health insurance              describes how patient race and income are associated with
status, and income. Finally, we use decomposition methods             operative time, after adjusting for other patient characteristics,
to separate the observed racial/ethnic gaps in influenza              procedure and hospital.
vaccination rates into two parts: the part that is due to             Study Design: To estimate anesthesia minutes in a patient’s
differences in the observed characteristics of the white and          chart, we utilized Medicare claims. For 1931 patients we
black (or white and Hispanic) adults and the part that is due         abstracted charts with information that included the time of
to other factors.                                                     anesthesia induction and the time of entrance to the recovery
Population Studied: The total sample size for the study               room, and fit a model using Theil’s estimate. Using this
consisted of 74,926 adults.                                           model, we analyzed claims data to study the relationship
Principal Findings: There has been no meaningful                      between operative time and race and income. Median
improvement in influenza vaccination rates since 1997 within          household income was defined by Zip code, grouping sample
these high priority groups. In addition, there has been little        size into thirds: higher = above $41,858; lower = below $31,388.
progress toward narrowing racial and ethnic gaps in                   Population Studied: We obtained Medicare claims on all
vaccination rates, and if anything, the gaps have been growing        patients who underwent general surgical and orthopedic
larger for some groups. Achieving the Healthy 2010 objectives         surgical procedures in Pennsylvania between the ages of 65
would require a 22 percentage point increase in influenza             and 85 over a two year period (N=77,638). Anesthesia time
vaccination rates for the elderly and a 24 percentage point           was estimated from Medicare Part B bills. Specifically, the
increase for adults under age 65 with a chronic condition.            dataset included claims with minutes of time provided by the
Even larger increases would be required to bring vaccination          anesthesia provider.
rates for racial and ethnic minorities in line with the objectives,   Principal Findings: Claim time from Medicare was highly
since vaccination rates for whites within each high priority          predictive of chart time (Kendall’s tau = 0.93, p < 0.0001,
group are about 30 percent higher than for blacks and about           median absolute error = 5.4 minutes). Using Huber’s robust
22 percent higher than for Hispanics. Very little of the white-       m-estimation, we then studied the influence of race and
black gaps in influenza vaccination rates can be explained by         income on 77,638 claims involving general surgery and
differences in underlying observed characteristics. By                orthopedics procedures, adjusting for 25 patient
contrast, much of the white-Hispanic gap is explained by              comorbidities, transfer status, MedisGroups Severity Score,
underlying observed characteristics.                                  procedure, and the individual hospital. The shortest
Conclusions: Achieving the Healthy People 2010 objectives             anesthesia time was found in the higher income-white group.
will require large increases in vaccination rates, particularly       Compared to this group, lower income-black patients had, on
among racial and ethnic minorities. The sources of the                average, 8.01 minutes longer anesthesia time (p < 0.0001).
persistent white-black differences in influenza vaccination           Placing this in context, history of paraplegia (an extremely
rates are not well understood. Additional research is needed          difficult comorbidity to care for in the operative suite) typically
contributed 5 additional minutes to anesthesia time. Middle        increasing over time overall, 64%-67% of visits to psychiatrists
income-black patients displayed a 5.20 minute longer               involving white or Latino patients resulted in a diagnosis of
anesthesia time (P = 0.0017) than higher income-whites, while      anxiety or depression, compared with about 55% of visits with
higher income-blacks showed no difference in anesthesia time       black patients. In primary care, 4.3%-6% of visits involving
compared to the higher income-white group (0.27 minutes, P         white patients resulted in an anxiety/depression diagnosis,
= 0.881). Anesthesia time for the lower income-white group         compared with about 2.5%-3.5% of visits with black or Latino
was 1.03 minutes longer than for the higher income-white           patients. Prescriptions for antidepressants were more likely
group (P<0.040) but middle income-whites showed no                 for psychiatric visits during the period 1998-2000 (OR=1.47,
significant difference from higher income-whites, (0.68            CI=1.07, 2.00) (compared with 1995-1997), while for primary
minutes, P = 0.121).                                               care the likelihood was greater during 2001-2003 (OR=1.38,
Conclusions: After adjusting for procedure, cormorbidity,          CI=1.12, 1.70). About 54%-62% of psychiatric visits with
severity and hospital, the race and income of a patient appears    Latinos and whites resulted in a prescription for
to be associated with operative time.                              antidepressants, compared to 50%-59% of visits with black
Implications for Policy, Delivery, or Practice: It is of obvious   patients. About 6.5%-8% of visits with white patients resulted
concern that a most fundamental aspect of surgery--operative       in prescriptions for antidepressants, compared with 4%-5% of
time-- is associated with race and income. Understanding why       primary care visits with black or Latino patients.
this occurs will require further study of patients, providers,     Conclusions: Primary care visits involving black and Latino
and provider style of practice. Finally, using Medicare claims     patients, and psychiatric visits involving black patients, are
from the anesthesia provider enables the study of operative        significantly less likely to result in a diagnosis of depression or
time with potentially vast populations of patients, and            anxiety. Psychiatric specialists are equally likely to prescribe
consequently provides a view on the practice of surgery that       anti-depressants to black or Latino patients (as compared with
has never before been seen.                                        whites), but primary care office visits involving black or Latino
Primary Funding Source: AHRQ, The Leonard Davis Institute          patients are less likely to result in a prescription for
of Health Economics                                                antidepressants. These differences remain remarkably
                                                                   consistent over time.
●Anxiety/Depression Diagnoses and Antidepressant                   Implications for Policy, Delivery, or Practice: While primary
Prescribing Patterns by Primary Care and Psychiatric               care and psychiatric specialists diagnose depression/anxiety
Specialties, NAMCS 1995-2003                                       less frequently in visits with blacks, and primary care
Susan Stockdale, Ph.D., Isabel Lagomasino, M.D., Jeanne            physicians were less likely to diagnose depression/anxiety in
Miranda, Ph.D.                                                     visits with Latinos, differences in prescribing practices by
                                                                   race/ethnicity are apparent only for primary care physicians.
Presented By: Susan Stockdale, Ph.D., Associate Research           This may be a consequence of physician knowledge and ability
Sociologist, Health Services Research Center, UCLA, 10920          to identify the disorder, differences in the pathways to care for
Wilshire Blvd., Ste 300, Los Angeles, CA 90024; Tel: 310-794-      psychiatric specialties versus primary care, and/or differences
3732; Fax: 310-794-3724; Email: sstockdale@mednet.ucla.edu         in patient preferences.
Research Objective: Ethnic minority patients with depressive       Primary Funding Source: MacArthur Foundation
disorders are frequently underdiagnosed and undertreated.
National efforts to improve care for depressive and anxiety        ●Quality and Inequality in Medicare Managed Care
disorders have targeted primary care settings, the setting         Amal N. Trivedi, M.D., M.P.H., Alan M. Zaslavsky, Ph.D., Eric
where minorities are most likely to seek mental health care,       C. Schneider, M.D., MSc, John Z. Ayanian, M.D., M.P.P.
and thus have the potential to reduce disparities in care. This
study examined national trends in depressive and anxiety           Presented By: Amal N. Trivedi, M.D., MPH, Research Fellow,
diagnoses and antidepressant prescriptions by psychiatrists        Department of Health Care Policy, Harvard Medical School,
and primary care physicians for white, Latino, and black           180 Longwood Avenue, Boston, MA 02115; Tel: (617)432-3138;
patients.                                                          Email: trivedi@hcp.med.harvard.edu
Study Design: We used data from the US National                    Research Objective: Public reports about the quality of
Ambulatory Medical Care Survey for 1995 to 2003. Differences       health-care organizations rarely assess whether care is
by race/ethnicity in diagnoses and prescription patterns over      provided equitably to racial subgroups. Our objectives were to
time were examined using standardized predictions based on         assess variations in quality and racial disparity within
parameters of multivariate logistic regression models              Medicare health plans, determine whether high-quality plans
(controlling for age, gender, and insurance). All analyses were    have smaller racial disparities in quality, and develop a model
conducted separately for primary care and psychiatric              for publicly reporting racial disparities in the quality of care of
specialties using weights provided by the National Center for      health plans.
Health Statistics to obtain nationally representative estimates.   Study Design: To determine the correlation between quality
Population Studied: Physician office visits to primary care        and disparity, we fitted multilevel linear models predicting
and psychiatric specialties by patients age 18 and over from       receipt of each HEDIS indicator, adjusted for demographic
1995-2003 (N=52,504).                                              and health-plan characteristics. For each measure, we
Principal Findings: Visits to primary care occurring in 2001-      classified each health plan with twenty or more eligible black
2003 were more likely to result in a depression or anxiety         enrollees as having above average, average, or below average
diagnosis as compared with visits in 1995-1997 (OR=1.43,           quality and disparity by testing whether the plan´s
CI=1.15, 1.79); time trends for the psychiatric specialties were   performance rate was both statistically different (p<0.05) from
not statistically significant. With diagnoses and prescriptions    the average rate (using the t-test) and also in the top or
bottom quartile for performance among all plans. Similarly, for
each measure, we classified plans as having above average,
average or below-average white-black disparity.                                             Call for Papers
Population Studied: We analyzed 448,790 observations for
black and white Medicare enrollees in 164 managed care plans        The Role of Language & Literacy in Healthcare Disparities
from 2001 to 2003 who were eligible for 4 HEDIS outcome
measures assessing control of blood pressure for                       Chair: Ninez Ponce, University of California, Los Angeles
hypertension, glucose for diabetes, and cholesterol for
diabetes or coronary disease.                                                    Sunday, June 25 • 5:45 pm – 7:14 pm
Principal Findings: The mean performance on each HEDIS
measure was significantly lower for black enrollees than white
enrollees (all P<0.001), with absolute differences ranging from     ●Literacy and Mortality among the Elderly
5% for blood pressure control to 15% for LDL control after a        David Baker, M.D., M.P.H., Michael S. Wolf, Ph.D., M.P.H.,
coronary event. For each measure, over three quarters of the        Joseph M. Feinglass, Ph.D., Violeta Carrion, M.S., Julie
total disparity was attributable to differences within plans.       Gazmararian, Ph.D., M.P.H.
Among all plans, the standard deviation of mean clinical
performance rates, derived from multilevel models, ranged           Presented By: David Baker, M.D., M.P.H., Chief, Division of
from 7.3% for the blood pressure measure to 14.2% for               General Internal Medicine; Associate Director, Institute for
cholesterol control after a coronary event. No significant          Healthcare Studies, Medicine, Northwestern University, 676
relation between quality and racial disparity was evident on        N. Saint Clair St., Suite 200, Chicago, IL 60611; Tel: (312) 695-
any of the four outcome measures after controlling for age,         0917; Fax: (312) 695-0951; Email: dwbaker@northwestern.edu
gender, and year of measurement. After also controlling for         Research Objective: Low literacy has been linked to less
other individual and health-plan characteristics, a larger racial   knowledge of medical conditions, lower use of preventive
disparity in hemoglobin A1c control was noted in higher             services, higher hospitalization rates, and poorer physical and
quality plans (r=0.4, p<0.001) but no significant association       mental health. We conducted this study to determine
was found for any of the other three HEDIS measures. In             whether literacy independently predicts mortality.
adjusted analyses, three health plan characteristics predicted      Study Design: We conducted a prospective cohort study
performance for at least 3 of the 4 HEDIS outcome measures          using data from the Literacy and Health of Medicare Managed
we studied. Plans in operation >25 years, plans with >100,000       Care Enrollees (LHMMCE) study and the National Death
Medicare enrollees, and staff- or group-model plans had             Index (NDI). The LHMMCE conducted in-home interviews in
higher performance than younger, smaller and non-                   1997 and measured demographics, health status, and health
staff/group model plans. No plan-level characteristics were         behaviors. Literacy was measured using the shortened
consistently associated with racial disparity. Based on our         version of the Test of Functional Health Literacy in Adults
statistical criteria, only one plan achieved both high quality      (TOFHLA) and patients categorized as having adequate,
and low disparity for more than one performance measure.            marginal, or inadequate literacy. The LHMMCE database was
Conclusions: The results of HEDIS outcome measures vary             merged with all-cause mortality data from the NDI through
widely across Medicare health plans and often reveal                2003. To determine whether inadequate or marginal literacy
disparities within health plans. The magnitude of racial            were independent predictors of mortality, we used multivariate
disparity in outcomes is not associated with the HEDIS              Cox proportional hazards models and sequentially controlled
outcomes achieved by a plan.                                        for demographics (age, gender, race/ethnicity, language
Implications for Policy, Delivery, or Practice: Plan-specific       spoken, site), socioeconomic status (annual income, years of
performance reports stratified by race can provide additional       school completed, former occupation), and baseline health,
information about equity, a dimension of quality not assessed       including physical and mental health (SF-36 summary scales),
by most current public reporting systems.                           limitations in activities of daily living, and chronic conditions
Primary Funding Source: AHRQ                                        (hypertension, heart failure, diabetes, coronary artery disease,
                                                                    arthritis, chronic lung disease, and cancer). In addition, we
                                                                    compared the bivariate and multivariate relationship between
                                                                    years of school completed and mortality. Finally, we added
                                                                    health risk behaviors (smoking status, alcohol use, physical
                                                                    activity, body mass index, seatbelt use) to the model to
                                                                    determine whether these mediated the relationship between
                                                                    literacy and mortality.
                                                                    Population Studied: 3,260 new Medicare managed care
                                                                    enrollees in four U.S. metropolitan areas (Cleveland, OH;
                                                                    Houston, TX; Tampa, FL; Fort Lauderdale-Miami, FL).
                                                                    Principal Findings: The crude mortality rates for participants
                                                                    with adequate (N=2,094), marginal (N=366), and inadequate
                                                                    (N=880) literacy were 18.9%, 28.7%, and 39.4%, respectively;
                                                                    p<0.001. After adjusting for demographics, the hazard ratios
                                                                    (HR) for inadequate and marginal literacy were 1.88 (95% CI
                                                                    1.60-2.22) and 1.33 (95% CI 1.07-1.66), respectively. After
                                                                    adding socioeconomic variables, the HRs decreased to 1.74
                                                                    (95% CI 1.44-2.09) and 1.27 (95% CI 1.02-1.60) respectively. In
the final model that included baseline health variables, the         staff; using phone interpreters; analyzing translation
HRs were 1.61 (95% CI 1.32-1.95) and 1.20 (95% CI 0.95-1.51),        quality/accuracy of labels and information packets; and adding
respectively. The addition of health risk behaviors to the model     more languages to pharmacy software. Model pharmacies
did not substantially change these results. In contrast to           shared a combination of validated NEL computer programs,
literacy, years of school completed was not significantly            bilingual staff, and phone interpreters.
associated with mortality in bivariate analyses and in               Conclusions: About half of the pharmacies surveyed
multivariate models that included literacy; the HRs for 0-8, 9-      never/only sometimes can print NEL prescription labels,
11, some college, and college graduate (12 years of education        prepare NEL information packets, or orally communicate in
as reference group) were 0.83 (95% CI 0.66-1.06), 0.95 (95%          NELs. Almost half are unsatisfied with their LEP patient
CI 0.77-1.17), 0.94 (95% CI 0.74-1.17), and 1.23 (95% CI 0.93-       communication, and 1 in 9 pharmacies that orally
1.60) respectively.                                                  communicate in NELs use patient family members/friends to
Conclusions: Among community-dwelling elderly, inadequate            interpret. Chains and those using phone interpreters are more
health literacy is a strong, independent predictor of mortality.     likely and community pharmacies are less likely to be able to
In contrast, education alone was not predictive. Differences in      orally communicate in NELs. Improvements in pharmacies'
health behaviors do not explain the higher mortality for             LEP patient communication may result by increasing bilingual
individuals with inadequate literacy.                                staff, the quality and number of NELs in existing computer
Implications for Policy, Delivery, or Practice: Health care          programs, phone interpreter use, and continuing education.
providers and payors should consider inadequate literacy as a        Implications for Policy, Delivery, or Practice: The study
marker of high risk for adverse outcomes and examine ways to         findings suggest that many urban pharmacies may not be
identify patients and try to reduce their risk through intensive     providing adequate services to LEP patients, thereby
patient education or case-management.                                increasing the risk of compromised patient safety. Excellent
Primary Funding Source: NIA, National Institute on Aging             communication with LEP patients can be achieved by using
                                                                     validated computer programs to prepare NEL prescription
●A Bitter Pill? Language Barriers to Prescriptions for               labels and information packets, hiring bilingual staff, and
Patients with Limited English Proficiency (LEP): A Survey            using phone interpreters when in-person interpreters are
of Pharmacies                                                        unavailable, as already occurs in certain model pharmacies.
Michael Bradshaw, BS, Richard Hayney, RPh, Emmanuel                  Primary Funding Source: NHLBI
Ngui, DrPH, Glenn Flores, M.D.
                                                                     ●Pediatricians' Use of Language Services for Limited
Presented By: Michael Bradshaw, BS, 11010 W. Meinecke Ave,           English Proficient (LEP) Families: A National Survey
#1, Milwaukee, WI 53226; Tel: (414) 258-7901;                        Dennis Kuo, M.D., Karen G. O'Connor, Glenn Flores, M.D.,
Email: mbradsha@mcw.edu                                              Cynthia S. Minkovitz, M.D., M.P.P.
Research Objective: To contact all Milwaukee County
pharmacies to evaluate the ability to provide prescription           Presented By: Dennis Kuo, M.D., Fellow, General Pediatrics &
labels, information packets, and oral communication in non-          Adolescent Medicine, Johns Hopkins University School of
English language (NEL)s, assess satisfaction with LEP patient        Medicine, 600 North Wolfe Street, Park 364, Baltimore, MD
communication, determine pharmacies’ suggestions for                 21287; Tel: (410) 614-0910; Fax: (410) 502-5440;
improving communication with LEP patients, and identify              Email: dkuo5@jhmi.edu
features of model pharmacies.                                        Research Objective: LEP patients confront multiple barriers
Study Design: In this cross-sectional survey of Milwaukee            to health care access in the United States. Adequate language
County pharmacies, a pharmacist or technician at each                services are essential to providing high quality health care for
pharmacy was asked questions by phone, fax or mail. The              children. Trained professional interpreters and bilingual
survey consisted of 45 Likert scale, yes/no, and open-ended          providers are associated with improved quality of care for LEP
questions that addressed selected socio-demographic and              patients, but little is known about pediatricians' use of
language service characteristics of pharmacies.                      language services. The research objectives were to examine
Population Studied: Pharmacists or technicians of 175                pediatricians' provision of language services to LEP patients
Milwaukee County pharmacies, focusing on interaction with            and the physician, practice, and state characteristics
their LEP patients.                                                  associated with use of these services.
Principal Findings: 73% of 175 pharmacies responded. 47%             Study Design: Data were obtained from the Periodic Survey of
of pharmacies never/only sometimes can print NEL                     Fellows #60, a nationally representative survey of American
prescription labels, 54% never/only sometimes can prepare            Academy of Pediatrics (AAP) members. 1,829 surveys were
NEL information packets, and 64% never/only sometimes can            mailed and responses obtained from 58%. Use of six
orally communicate in NELs. For pharmacies that orally               language services (bilingual physicians, bilingual staff,
communicate in NELs, 11% used patient family                         bilingual family members, professional interpreters, telephone
members/friends to interpret. Only 55% of pharmacies were            interpreters, and translated written materials) was assessed.
satisfied with their LEP patient communication. In multivariate      Multivariate analysis examined factors associated with each
analyses, ability to orally communicate in NELs was                  language service after adjusting for physician (age, gender,
significantly associated with chain pharmacies (OR=3.2 [1.1-         ethnicity), practice (setting, location) and state (LEP
9.0]), phone interpreter use (OR=2.9 [1.1-7.3]), and community       prevalence, Medicaid/SCHIP reimbursement) characteristics.
pharmacies (OR=0.2 [0.1-0.5]). Pharmacies suggestions for            Population Studied: 698 non-trainee US members of the AAP
improving LEP patient communication included continuing              with LEP patients.
education; chain-wide lists of resources to call; hiring bilingual
Principal Findings: Respondents report a mean of 13.4%             of each health site. Watchfire’s automated “Bobby” software
(median=5.0%) of patients with LEP. Female, younger, and           on usability was used to assess how accessible state health
Latino physicians care for higher percentages of LEP patients      websites were to individuals with various types of
(p<.01). Bilingual family members (70%) and bilingual staff        impairments. Language access was evaluated by determining
(58%) are the most frequently reported language services.          whether state health websites provided information in
40% of respondents use professional interpreters and 28%           languages other than English or had links to common
use telephone interpreters. Only 35% of practices provide          translation software. Confidentiality was evaluated by
translated written materials. In bivariate analysis,               examining privacy and security statements.
pediatricians in rural and solo/2-physician practices report       Population Studied: This studied analyzed public health
higher use of family members and lower use of professional         department websites maintained by the 50 state governments.
interpreters (p<.01). Family members are among the most            Principal Findings: The average grade readability level of
frequently used method of interpretation regardless of LEP         state health websites was at 10.9th grade in 2005, with 62
prevalence in states. In multivariate analysis, pediatricians in   percent reading at the 12th grade level, well above the
states with higher proportions of LEP persons are more likely      comprehension level of the typical American. Only 42 percent
to use bilingual staff (OR 4.63; 95% CI 3.22, 6.67) but less       of public heath sites were accessible to the disabled; only 34
likely to use professional interpreters (OR .46; 95% CI .31,       percent provided for non-English translations. Eighty-six
.69). Pediatricians in states with Medicaid/SCHIP                  percent had privacy policies and 62 percent security policies.
reimbursement for language services are more likely to use         Though there was progress on privacy and security over time,
professional interpreters (OR 2.05; 95% CI 1.10, 3.83).            there was little to no progress on the dimensions of
Conclusions: Almost three-quarters of US pediatricians use         accessiblity examined. There was considerable variation across
untrained family member interpreters to communicate with           regions, however, though no single region was strong on all
LEP patients and their families. Smaller and rural practices       aspects of accessibility, privacy, and security.
are at greater risk of using family member interpreters. Most      Conclusions: Ideally, all citizens should share equally in the
pediatricians do not provide translated written materials.         benefits of digital technology. Based on that standard,
Pediatricians in regions with the highest proportions of LEP       however, public e-health clearly has a long way to go. The level
persons are least likely to provide appropriate language           at which state public health sites are written represents a
services. Medicaid/SCHIP reimbursement for professional            major accessibility barrier as do the failure of many
language services may increase use of trained interpreters and     government sites to be usable by the disabled or those who
quality of care for LEP children and their families.               do not speak English. This makes it difficult for many people
Implications for Policy, Delivery, or Practice: Appropriate        to acquire timely health information, compare health care
language services for LEP patients should be improved as part      providers, or access online medical services.
of ongoing efforts to improve quality of care; in particular,      Implications for Policy, Delivery, or Practice: The gap
special attention is needed to providing more translated           between information haves and have-nots should be a major
written materials to patients and reducing the use of family       concern to health care policymakers. Inaccessible websites
members as interpreters. Interventions should be targeted in       hurt the underprivileged and make it difficult to justify the
practices and states with high proportions of LEP patients.        technology investment that has taken place in state
Current efforts to expand third-party reimbursement for            governments around the country. Unless these concerns are
language services may help reduce existing health care             addressed, public e-health will remain the domain of highly-
disparities.                                                       educated and affluent individuals who speak English and do
Primary Funding Source: HRSA                                       not suffer from physical impairments.
                                                                   Primary Funding Source: No Funding
●The Digital Divide in Public E-Health: Accessibility and
Privacy in State Health Department Websites                        ●To Err is Human: Literacy and Misunderstanding of
Edward Miller, Ph.D., M.P.A., Darrell West, Ph.D., Edward          Prescription Drug Labels
Alan Miller, Ph.D., M.P,A.                                         Michael Wolf, MA, M.P.H., Ph.D., Terry Davis, Ph.D., Pat F.
                                                                   Bass, M.D., M.P.H., Hugh Tilson, M.D., Dr.P.H., Ruth M.
Presented By: Edward Miller, Ph.D., M.P.A., Assistant              Parker, M.D.
Professor, Taubman Center for Public Policy, Brown
University, 67 George Street, Box 1977, Providence, RI 02912-      Presented By: Michael Wolf, MA, MPH, Ph.D., Assistant
1977; Tel: 401-863-9311; Email: edward_a_miller@brown.edu          Professor, Institute for Healthcare Studies, Northwestern
Research Objective: With the significant increase in online        University, 676 N. St. Clair Street Suite200, Chicago, IL 60611;
resources at official government health sites have come            Tel: 312-695-0459; Email: mswolf@northwestern.edu
questions concerning equity of access and the confidentiality      Research Objective: Patients’ literacy and their ability to
of electronic medical materials. The purpose of this paper is      understand medication information have increasingly been
to examine accessibility and privacy in the context of state       viewed as a safety issue. We examined whether adult primary
public health department websites.                                 care patients were able to understand instructions found on
Study Design: Using a content analysis of health department        the primary container label of five common prescription
sites undertaken each year from 2000 to 2005, several              medications. We hypothesized that low literacy would be
different dimensions of accessibility and privacy were             associated with higher rates of misunderstanding and
investigated: readability levels, disability access, non-English   incorrect demonstration of instructions.
accessibility, and confidentiality. The Flesch-Kincaid method      Study Design: In-depth cognitive interviews including a
was used to judge the grade-level readability of the front page    literacy assessment were conducted. The main outcome
measure was understanding of medication label instructions,
which was determined by a blinded physician panel review of
patients’ verbatim responses for each of the five prescription
                                                                                         Related Posters
labels. For one of the medication labels, patients were also
asked to calculate how many pills were to be taken daily to test                               Disparities
whether understanding translated to an ability to practically
demonstrate how to take the medication.                                                    Poster Session B
Population Studied: 395 adults who were receiving outpatient
primary care services at two community health centers                            Monday, June 26 • 5:30 pm – 7:00 pm
(Jackson, Michigan; Chicago, Illinois) and at one public
hospital clinic (Shreveport, Louisiana) were recruited for the       ●A Predictive Model of Attrition In a Randomized
study.                                                               Intervention Among African American Women
Principal Findings: Rates of understanding ranged from               Mary K Anthony, Ph.D., Lisann Gittner, M.Sc., Salwa
67.1% to 91.1% for the five medication labels. Forty-six percent     Hassanein, Ph.D. candidate, Jacqueline Martin, RN, PA,
of patients misunderstood one or more of the label                   Pamala Murphy, M.D.
instructions, and the prevalence rates of misunderstanding
one or more labels for adequate, marginal, and low literate          Presented By: Mary K Anthony, Ph.D., Associate Professor,
patients were 37.7%, 51.5%, and 62.1% respectively. Patients         College of Nursing, Kent State University, BO Box 5190, Kent,
reading at the 6th grade level or below (low literacy) were less     OH 44242; Tel: (330) 672-8824; Fax: (330) 672-1564; Email:
able to correctly interpret the meaning of four of the five labels   manthony@kent.edu
compared to patients with adequate literacy (9th grade and           Research Objective: African –American women are more
above), with the exception of the instructions associated with       obese, have greater difficulty in losing weight and maintaining
the drug Furosemide (p=0.09). After controlling for relevant         weight loss. Successful weight loss occurs when interventions
potential confounding variables and risk factors, both low           are culturally adapted and of sufficient length to change
(AOR 2.81, 95% CI 1.38-5.71) and marginal (7th-8th grade; AOR        lifestyle behaviors. Unfortunately, extended intervention
2.13, 95% CI 1.15-3.93) literacy levels were significantly           protocols designed for lifestyle modifications are also the
associated with misunderstanding of label instructions.              most challenging in maintaining subject retention. The
Among patients with low literacy who properly understood the         purpose of this study was to identify what personal
label instructions (70.7%), less than half were able to              demographic and social characteristics, health and illness
demonstrate comprehension (34.7%) of the instructions. In            characteristics and self efficacy were predictive of African-
multivariate logistic regression analysis, low literacy was a        American women’s ongoing participation in a culturally
significant independent predictor of correct demonstration of        adapted weight loss intervention.
the label instruction (AOR 4.32, 95% CI 2.10 – 6.92).                Study Design: One hundred African American middle aged
Conclusions: Patients with limited literacy are at greater risk      obese women were recruited from a Midwestern HMO. Five
for misunderstanding instructions on prescription medication         cohorts of women were recruited over a 5 month time period
labels, and this may be a common source of medication                and randomized into a 6-month Take HEED: Health Eating
errors. In addition, correct interpretation of instructions was      and Exercise Decisions intervention. Each 6-month
not an assurance of practical comprehension of the daily             intervention protocol consisted of 15 sessions that included
administration of a prescribed drug.                                 health behavior changes related to diet, nutrition- NCEP Step 1
Implications for Policy, Delivery, or Practice: New                  diet-, motivation and exercise- CHANGE. The initial six
standards should be considered to ensure medication labels           sessions were held weekly and the subsequent nine sessions
are useful to patients. Provider-patient communication on            were held every other week. Sessions were facilitated by either
medication administration should also be emphasized.                 a dietitian or African-American Registered Nurse. To date,
However,recommended low literacy communication strategies            three cohorts have completed the HEED intervention n = 53.
for physicians, such as the ‘teach back’ technique may be            Preliminary findings are reported, and final model building will
inadequate, and actual demonstration of desired patient              be conducted when all five cohorts have completed the HEED
behaviors should be targeted.                                        intervention March, 2006.
Primary Funding Source: No Funding                                   Population Studied: The average age of women was 53.4
                                                                     years SD = 6.4 with 14.4 years of education. Almost 57% were
                                                                     married, 83% were employed and 50% had an annual income
                                                                     of greater than $50,000. Average body mass index was 37.8
                                                                     SD = 6.1. Women attended on average 55.7% SD= 26.7% of
                                                                     HEED sessions.
                                                                     Principal Findings: Variables from baseline data were
                                                                     assessed for their relationship with the percent of sessions
                                                                     attended. Variables or proxies were chosen based on
                                                                     previously identified barriers to retention. Statistically
                                                                     significant personal and social characteristics included age r =
                                                                     .39, marital status r = -.31, number of people living in the
                                                                     household r = 29. Likewise health and illness characteristics
                                                                     that were significantly associated with attendance at HEED
                                                                     sessions were the number of comorbid conditions r = .38 and
the degree of pain while exercising r = .41. Nonsignificant          Standard measures of demographic and socio-economic
factors were income, employment status, previous weight loss         status are included in our analysis as covariates. The chi-
attempts, BMI and self efficacy. Anecdotal reasons for not           square is used to test bivariate relationships between CAM
attending HEED sessions included change of jobs, caregiving          and the main explanatory variables. Multivariate logistic
for family members, illness or change from HMO insurance.            regression methods are used to analyze the effects of health
Conclusions: Without an adequate dose of an intervention,            status and other predictors on use of CAM.
success of weight loss programs are jeopardized. These               Population Studied: An effective sample size of 4684
preliminary findings suggest that older women who have               weighted cases was used.
social support, and have more health and illness problems are        Principal Findings: Overall, 32 percent of the sample used
more likely to maintain participation in an extended                 CAM therapies. Utilization of CAM therapies varied inversely
intervention protocol. These factors have demonstrated an            with health status. Those who were in poor health were 15 %
initial relationship to retention of participants that affects the   more likely as those in good health to report use of CAM.
dose of the intervention, its probable success and                   Those who were dissatisfied with the overall quality of care
effectiveness.                                                       they received in the last two years were more likely to use
Implications for Policy, Delivery, or Practice: Clinical             alternative medicine. Those who were in poor health or
investigations that encourage lifestyle changes are expensive        chronically ill and satisfied with conventional care used CAM
in cost and time and thus identifying culturally relevant factors    more than those who were chronically ill and dissatisfied with
that affect retention can do much to improve both the                conventional care. The chronically ill who were dissatisfied
participation of minorities in studies to improve health,            with conventional care used CAM as an alternative rather than
address health disparities, and decrease the burden on the           complimentary to conventional care.
health care system.                                                  Conclusions: CAM use among Americans is high. This study
Primary Funding Source: Garfield Memorial Foundation                 shows that the chronically ill, those dissatisfied with
                                                                     conventional care and have difficulties accessing care due to
●Health Status and Complimentary and Alternative                     cost, are likely to use CAM. Use among these groups is
Medicine (CAM): Assessing the Effects of Perceived                   primarily seen as an alternative rather than complementary to
Health Status and Satisfaction with Conventional Care on             conventional care. Perceived health status is related to
the use of CAM                                                       enhanced access to conventional care.
Winfred Avogo, MA, Jemima Frimpong, M.P.H., Sam Kim,                 Implications for Policy, Delivery, or Practice:
MA, Patrick Rivers, Ph.D., M.B.A.                                    Considerations should be given to the development of a
                                                                     delivery system that effectively incorporates conventional care
Presented By: Winfred Avogo, MA, Graduate Research                   and CAM into patient care. The impact of this approach could
Associate, Department of Sociology/Centre for Population             be instrumental in addressing health care cost, satisfaction
Dynamics, Arizona State University, Arizona State                    and quality of care.
University/Department of Sociology, Mesa, AZ 85287-4802;             Primary Funding Source: No Funding
Tel: 480 965 3546; Fax: 480 965 0065; Email:
winfred.avogo@asu.edu                                                ●Obesity, Ethnicity and Cardiovascular Diseases
Research Objective: Research has shown increasing                    Jolene Ayres-Oddiah, John Kwagyan, PH.D
popularity of various forms of complimentary and alternative
care in the US; however, the influence of these therapies have       Presented By: Jolene Ayres-Oddiah, M.P.H. Program, Howard
been largely exploratory without a systematic understanding of       University College of Medicine, 2041 Georgia Ave NW,
the individual and institutional factors driving utilization. In     Washington, DC 20060; Tel: 202- 865-7213; Email:
this paper, we address the need to develop and test explicit         jkwagyan@howard.edu
explanatory hypotheses by examining perceived health status          Research Objective: Cardiovascular disease is the leading
and patient satisfaction with conventional care. We                  cause of death in the United States and it disproportionately
hypothesize that CAM utilization varies inversely with               affects African Americans. Obesity is known to be associated
perceived health status and the level of satisfaction with           with cardiovascular disease risk factors including diabetes
conventional care. This relationship is explained by reasons for     mellitus, high cholesterol, and hypertension. African
use of CAM for both patients in good health and the                  Americans are more likely to be overweight and suffer greater
chronically ill.                                                     risk of these cardiovascular diseases than other ethnic groups.
Study Design: The study uses data from the 2001 Survey on            Genetic factors, improper nutrition and physical inactivity have
Disparities in Quality of Health Care conducted by the               been noted to be the leading cause of these increases in risk
Commonwealth Fund. A stratified sample design was used to            among African Americans. However, little or no data examines
conduct random-digit dial telephone surveys of a nationally          the prevalence of these cardiovascular disease risk factors in
representative sample of adults age 18 and older living in           obese individuals in different ethnic groups. The objective of
continental United States. The sample was designed to target         this study is to examine and compare African Americans and
African-American, Hispanic and Asian Households. Sample              Caucasians, the prevalence of common cardiovascular disease
weights are used to correct for the disproportionate sample          risk factors in the context of obesity.
design and standard errors are adjusted for the complex              Study Design: A cross-sectional sample of 3324 individuals, of
survey design. The main outcome measure is use of four main          which 904 were obese, obtained from the National Health and
CAM therapies in the last two years. Primary independent             Nutrition Examinations Survey, NHANES III database were
variables reflecting the hypothesis of the study are perceived       studied. The NHANES III is a survey of a representative
health status and satisfaction with conventional care.               sample of the civilian, non-institutionalized population of the
United States that was conducted periodically from 1988 to        poverty. The study also explores the correlations of pain
1994 by the National Center for Health Statistics. The survey     measures with other health status measures, including
was designed to estimate the prevalence of common chronic         perceived health status, limitation in usual activity, serious
conditions and associated risk factors for disease control and    psychological distress, and obesity, among these
prevention. Chi-Square tests were used to compare the             socioeconomic population subgroups.
prevalence of hypertension, diabetes and high cholesterol,        Population Studied: Data come from National Health and
separately, among non-obese and obese individuals. To             Nutrition Examination Survey and the National Health
reduce the potential influence due to age, data was restricted    Interview Survey. Both of these surveys are nationally
to an age range from 30 to 55 years.                              representative probability samples of the civilian,
Population Studied: African Americans and Caucasians aged,        noninstitutionalized population of the United States.
30 to 55 years.                                                   Principal Findings: Women and white adults report more of
Principal Findings: The data consist of 2291 African              almost all types of pain than do men and other racial/ethnic
Americans, 36% obese and 957 Caucasians, 39% obese. In            groups. Preliminary results indicate that people of Mexican
general, obese individuals had higher prevalence of               origin report considerably less pain lasting 24 hours or more
hypertension, diabetes and high cholesterol than non-obese        over the past month (15%) than do Black-only (22%) or white-
individuals. Among the non-obese individuals, the results         only Americans (28%). Asians report less pain on all
shows that African Americans has a higher prevalence of           measures examined than do people of other races and
hypertension than Caucasians [odds ratio =1.5, CI=(1.2-1.9),p     ethnicities. Among those reporting pain lasting more than 24
<0.001], diabetes [odds ratio=1.5, CI=(1.01-2.3), p<0.5] and      hours, few differences exist in the percentage reporting that
high cholesterol [odds ratio =1.3, CI=(1.1-1.5), p <0.001].       the pain had lasted for more than one year by race, ethnicity,
However, among obese individuals, there were no significant       or gender. Asian and non-Hispanic black women reported low
difference in hypertension and diabetes between African           back pain about 30% more often, and Hispanic women
Americans and Caucasians. Obese African Americans had a           reported it 40% more often than their male counterparts.
higher cholesterol levels than Caucasians [odds ratio=1.2,        Blacks are somewhat less likely to report any joint pain than
CI=(1.1-1.9), p<0.01].                                            whites, but they are more likely to identify the joint pain they
Conclusions: The study was undertaken to examine and              have as being severe (7 or more on a scale of 1-10). Persons
compare cardiovascular disease risk factors between African       living in poverty reported more back pain, headache, neck pain
Americans and Caucasians. Consistent with other studies, the      than nonpoor or near poor adults, and this relationship holds
prevalence of diabetes, hypertension, and high cholesterol        when analyzed by race and ethnicity.
were higher among the obese population than non-obese             Conclusions: Patterns of pain self-reporting vary considerably
irrespective of ethnicity. Among non-obese individuals, the       by race, ethnicity and poverty, controlling for age and sex, and
prevalence were higher in African Americans than Caucasians.      pain is strongly associated with other health status measures.
However, among the obese individuals, the prevalence of           Women, white adults, and people living in poverty were more
these risk factors appears to be the same in both populations.    likely to report most types of pain than were men and other
Implications for Policy, Delivery, or Practice: Promoting         racial and ethnic groups.
healthy lifestyle is essential in reducing the associated         Implications for Policy, Delivery, or Practice: Results of this
cardiovascular disease risk factors in obese population.          analysis may encourage practitioners to query for pain
Endeavoring in proper cardiovascular exercises and nutritional    symptoms in particular populations that tend to report less
intakes will promote healthy body weight and reduce the           pain, or to learn more about the association of pain reporting
morbidity and mortality associated with obesity.                  with other health status, cultural, and demographic and
Primary Funding Source: Nof Funding                               socioeconomic factors. The relationships between reported
                                                                  pain and race/ethnicity, gender, age and health status are
●Prevalence of Pain and Associations with Socioeconomic           complex and raise important issues for public health
and Health Factors: Results from Two National Surveys.            initiatives to reduce disparities in health care policy and
Amy Bernstein, Sc.D., Alan J. Cohen, M.A.                         practice.
                                                                  Primary Funding Source: No Funding
Presented By: Amy Bernstein, Sc.D., Chief, Analytic Studies
Branch/OAE, , CDC/National Center for Health Statistics, 3311     ●Effect of Neighborhood Racial and Ethnic Composition
Toledo Road, Hyattsville, MD 20782; Tel: (301) 458-4700; Fax:     on Disparities in Imaging Utilization
(301) 458-4036; Email: abernstein@cdc.gov                         Mythreyi Bhargavan, Ph.D., Cristian Meghea, Ph.D., Jonathan
Research Objective: Both perceptions and reporting of pain        H. Sunshine, Ph.D.
are subjective. Although prevalence of many different types of
reported pain differ markedly by gender, age, race/ethnicity,     Presented By: Mythreyi Bhargavan, Ph.D., Director of
and poverty, the reasons underlying these differences are not     Research, Research, American College of Radiology, 1891
well understood. This analysis documents differences in the       Preston White Dr, Reston, VA 20191; Tel: 703-715-4394; Fax:
prevalence of selected pain measures by racial and ethnic         703-264-2443; Email: mbhargavan@acr.org
group, and examines correlations between pain and gender,         Research Objective: Racial and ethnic disparities in health
age, health status, poverty, and obesity.                         care in the United States have been well-documented in over
Study Design: Data examined include prevalence of multiple        600 studies. The small-area variation perspective on these
measures of pain, including pain lasting more than 24 hours       disparities suggests that the long-observed geographic
at some time in the past month; severe headache or migraine;      variations in practice patterns may explain some of these
low back pain; and joint pain, by sex, age, race/ethnicity, and   disparities, because minorities are geographically
concentrated in certain areas. For identical medical                ●Neighborhood Socio-Economic Status and Allostatic
conditions, persons living in high-minority areas may receive a     Load
very different set of medical services than those in mostly         Chloe Bird, Ph.D., Brian Finch, Ph.D., D. Phuong Do, M.Phil.,
white neighborhoods. In this study, we examine the role of          Teresa Seeman, Ph.D., Jose Escarce, M.D., Ph.D., Nicole
neighborhood racial/ethnic composition on the utilization of        Lurie, M.D.
medical imaging, one of the fastest-growing components of
healthcare costs in the United States.                              Presented By: Chloe Bird, Ph.D., Sociologist, Health, RAND,
Study Design: We examine the extent to which differences in         1776 Main Street, Santa Monica, CA 90407-2138; Tel: (310)
characteristics of the zip-code of a patient’s residence, such as   393-0411 x6260; Fax: (310) 260-8159; Email: chloe@rand.org
racial and ethnic composition, population density, percentage       Research Objective: A growing body of research
poor, and other demographic and socio-economic factors,             demonstrates relationships between residential neighborhood
explain differences in imaging utilization, while controlling for   context and health. However, few studies examine potential
individual, provider, and neighborhood characteristics.             biological pathways through which neighborhoods affect
Using claims data from a large national employer plan for five      health. We examined the extent to which neighborhood
years (1999-2003), we identify patients with certain conditions,    characteristics are related to biological markers of stress,
such as acute upper respiratory tract symptoms, headache,           based on a summary index of allostatic load, adjusting for
and knee pain, and analyze each condition independently. For        individual characteristics.
each patient who presents with a condition, we analyze the          Study Design: Using 3-level hierarchical linear regression, we
following outcomes: (a) whether an image was ordered, and           analyzed National Health and Nutritional Examination Survey
(b) the total number of images, (c) the type(s) of image(s)         III (NHANES) interview and laboratory data for 14,413 adults
(CT, MRI, etc.), and (d) total imaging utilization measured in      (ages 19.5 to 90, mean = 48) from 83 counties and 1805
global relative value units (RVUs). We use logistic regression      census tracts, merged with data on sociodemographic
for outcome (a), count-data regression models for (b),              characteristics of their residential census tract. The sample
multinomial logistic for (c), and log-linear regression for (d).    was 47% male; 43% white 27% black, 26% Hispanic, 4%
Explanatory factors include characteristics of the zip-code         other; mean income (or mean income/poverty ratio =2.41),
(such as percent of population in each minority group,              53% employed, and 58% with at least a high school education.
income), physician specialty, practice setting (office, hospital,   Allostatic load was measured as a summary score (range 0-9)
etc.), type of insurance plan, and available patient                based on clinical cut points for 9 indicators from 3 systems:
characteristics (age, gender, actuarial risk score).                metabolic (total cholesterol, HDL cholesterol, glycosylated
Population Studied: Persons enrolled in the health insurance        hemoglobin, waist/hip ratio), cardiac (systolic and diastolic
plan of a large national employer during the years 1999-2003        blood pressure, radial heart rate) and inflammatory (c-reactive
Principal Findings: Preliminary results on patients with knee       protein, serum albumin).
pain indicate that controlling for racial/ethnic and socio-         Population Studied: Our analyses included 14,413 adults
economic characteristics of the zip code and for patient age,       from 83 counties and 1805 census tracts, who completed
gender and risk score, high percent of poor or minority in the      surveys and medical exams, were not missing on key
zip code have no significant effect on probability of the patient   components of the outcome measures, and for whom
receiving a knee MRI, but personal characteristics (age,            residential census tract could be geocoded. The sample was
gender, risk score) and population density (additional 1%           47% male; 43% white 27% black, 26% Hispanic, 4% other.
population density results in 7% higher odds of an MRI,             Subjects ranged in age from 19.5 to 90 (mean = 48); 53% were
p<0.01) do. In contrast, analysis of cardiac imaging indicated      employed and 58% had at least a high school education. The
that patients who lived in high-Hispanic zip-codes had a lower      mean family income/poverty ratio was 2.41.
probability of being imaged than high-white zip-codes.              Principal Findings: Across all models, individual level socio-
Among those who were imaged, patients in high-black zip-            economic controls including Hispanic ethnicity (p < .001),
codes had fewer images than those in high-white zip-codes.          lower family income (p < .03), lower education (p < .05), lack
Conclusions: The extent of variation in the utilization of          of employment (p < .005), age (p < .001) and being male (p <
imaging across minority and white zip-codes is different for        .001) were independently associated with higher allostatic
different conditions. Further analysis of a wider range of          load. Even after adjusting for these individual characteristics,
criteria will allow us to identify the kinds of symptoms for        a higher proportion of adults in the census tract with less than
which neighborhood-based characteristics most affect volume         high school education was associated with a higher allostatic
of imaging.                                                         load (coefficient = 0.46, p < .0001). In a separate analysis,
Implications for Policy, Delivery, or Practice: Our results         higher median household income (in 10k units) was
will assist policy makers and payers in designing effective         associated with lower allostatic load independent of individual
incentive structures and practice guidelines to ensure              characteristics (coefficient = -0.03, p = .01).
equitable utilization across demographic categories and             Conclusions: Allostatic load is a biomarker of early health
geographic areas.                                                   decline. The association of lower socioeconomic status with
Primary Funding Source: No Funding                                  higher allostatic load, controlling for individual characteristics,
                                                                    suggests a potentially crucial pathway through which low-SES
                                                                    neighborhoods may affect both individual and population
                                                                    health.
                                                                    Implications for Policy, Delivery, or Practice: By assessing
                                                                    potential pathways through which health—and health
                                                                    disparities may be generated, this study is part of a larger
effort aimed at developing an understanding of whether               and health care expenditures. Finally, cancer control efforts
changing neighborhood features, such as neighborhood                 are potentially important for reducing the number of dually
quality could improve health and reduce health disparities.          eligible individuals.
The mechanisms through which these effects are produced              Primary Funding Source: NCI
need to be explored.
Primary Funding Source: NIEHS                                        ●Leave No Asthmatic Child Behind: The Cultural
                                                                     Competency of Asthma Educational Materials
●Cancer Incidence in Medicare and Dually Eligible                    Jane M. Brotanek, M.D., M.P.H., Kristen Grimes, Glenn
Beneficiaries: The Intersection of Age, Race, and Poverty            Flores, M.D.
Cathy Bradley, Ph.D., M.P.A., Zhehui Luo, Ph.D., Charles
Given, Ph.D., Caralee Roberts, Ph.D.                                 Presented By: Jane M. Brotanek, M.D., M.P.H., Assistant
                                                                     Professor, Pediatrics, Medical College of Wisconsin, 8701
Presented By: Cathy Bradley, Ph.D., MPA, Professor, Health           Watertown Plank Road, MS 756, Milwaukee, WI 53226; Tel:
Administration, Virginia Commonwealth University, 1008 E             (414) 456-5778; Fax: (414) 456-6385; Email:
Clay Street, Richmond, VA 23298-0203; Tel: 804-828-5217; Fax:        jbrotane@mcw.edu
804-828-1894; Email: cjbradley@vcu.edu                               Research Objective: Minority children are at high risk for
Research Objective: To examine how socioeconomic status,             asthma and its associated morbidity and hospitalizations.
age, and race influence cancer incidence rates within a publicly     Cultural competence leads to higher quality care for asthmatic
insured elderly population.                                          children, yet it is not known whether asthma educational
Study Design: Population-based cohort study individuals              materials targeting minority children are culturally competent.
diagnosed with a first primary cancer.                               The objective of this study was to develop and use a tool to
Population Studied: All Medicare and Medicaid beneficiaries          evaluate the cultural competency of asthma educational
age 65 years and older identified from the Michigan Tumor            materials.
Registry as having a first primary diagnosis of female breast,       Study Design: The Wisconsin Asthma Coalition Disparities
prostate, colorectal, or lung cancer. A total of 125,900             Workgroup, a diverse group of representatives from non-profit
subjects who were age 65 years and older were identified.            organizations, hospitals/clinics, and local/state government,
Principal Findings: Lung cancer was the only cancer that             held a series of meetings to develop, test, and apply a tool for
definitively occurred at a much higher rate in the dually eligible   evaluating the cultural competency of all asthma educational
sample relative to the Medicare sample. Relative to white            materials in the state of Wisconsin targeting minorities. The
elderly Medicare patients, elderly white dually eligible patients    tool, based on the Flores model of cultural competency,
have lower rates of breast and prostate cancer and slightly          consists of 7 domains: language, normative cultural values,
elevated rates of colorectal cancer. Dually eligible Blacks, on      folk illnesses, parent/patient beliefs, provider practices,
the other hand, had higher prostate, colorectal, and lung            illustrations, and an overall assessment. Asthma educational
cancer (men only) incidence rates relative to all white patients,    materials were obtained via multiple letters, emails, and
regardless of eligibility status. Dually eligible Black men had a    phone calls to healthcare organizations, agencies, and clinics.
2.5 higher incidence rate of prostate cancer relative to white       Three reviewers from the appropriate cultural group
dually eligible men.                                                 independently evaluated each material using the tool created
Conclusions: Our findings show that national estimates of            by the workgroup. All 7 domains were scored using a 5-point
disease incidence may underestimate how disparate racial             Likert scale (1=poor, 2=fair, 3=good, 4=very good, 5=excellent),
health inequalities truly are among the lowest income groups.        and consensus was reached for each score.
Poor Blacks have much higher rates of cancer relative to poor        Population Studied: Review of asthma educational materials
whites—so much so that the national trend of white women             Principal Findings: Only 17 asthma educational materials
having a higher incidence of breast cancer is not present in         targeting minorities exist in the state: 16 for Latinos, 1 for
the dually eligible sample. There is considerable room for           Native Americans, 0 for African Americans, and 0 for
improvement in cancer detection among the low-income.                Asians/Pacific Islanders. Total scores for each material ranged
Implications for Policy, Delivery, or Practice: Our findings         from 1-3, with a mean +/- SD = 2.3 +/- 0.5. The highest mean
have three important policy implications. First, policy makers       domain scores were for language (3.5 +/- 1.1) and illustrations
who are concerned with reducing health care disparities need         (3.5 +/- 1.2); the lowest score was for folk illnesses (1.1 +/- 0.2),
appropriate data. Further research using population-based            with significant weaknesses identified for normative cultural
datasets such as the linked Tumor Registry, Medicare, and            values (2.6 +/- 0.7); parent/patient beliefs (1.8 +/- 0.7); and
Medicaid files is needed and should be supplemented, where           provider practices (1.9 +/- 0.6).
possible, with individual data regarding education and               Conclusions: Few asthma educational materials targeting
income. Our findings show that national estimates of disease         minorities exist in the state, and those available are not
incidence may underestimate how disparate racial health              culturally competent. Most materials targeted Latinos; only 1
inequalities truly are among the lowest income groups.               targeted Native Americans, and none targeted African-
Second, interventions focused on African Americans and low-          Americans or Asians/Pacific Islanders. More culturally
income populations are needed to reduce disparities in cancer        competent asthma educational materials for minorities are
incidence. These interventions are needed prior to when              needed.
patients become eligible for Medicare and Medicaid and               Primary Funding Source: The Center for the Advancement of
should encompass environmental and social risk factors in            Underserved Children and the Department of Pediatrics,
addition to health care factors. Such interventions may reduce       Medical College of Wisconsin
mortality, the severity of Medicare’s and Medicaid’s case mix,
●Ethinic Disparties and Health Service Usage Post Stroke           ●Contextual and Health Systems Factors on Health
Paul Brown, Dr, Lorna Dyall, Dr, Ngaire Kerse, Assoc Prof,         Related Quality of Life of HIV-infected Urban African
Linda Cameron, Assoc Prof, Kristie Carter, Ms, Valery Feigin,      American Men
Assoc Prof                                                         Aaron Buseh, Ph.D, M.P.H., MSN, Kelber, Sheryl, T., MSc.,
                                                                   Park, Chang. Gi., Ph.D., Stevens, Patricia, E., Ph.D., FAAN
Presented By: Dr. Paul Brown, Senior Health Economist and
Director, Centre for Health Servcies Research and Policy,          Presented By: Aaron Buseh, Ph.D, M.P.H., MSN, Assistant
School of Population Health, University of Auckland, Private       Professor, College of Nursing, University of Wisconsin-
Bag 92019, Auckland, 0005; Tel: 64-9-373-7599; Fax: 64-9-373-      Milwaukee, 1921 East Hartford Avenue, Room 569, Milwaukee,
7503; Email: pm.brown@auckland.ac.nz                               WI 53211; Tel: (414) 229-5462; Fax: (414) 229-5504; Email:
Research Objective: Given the increasing incidence of stroke       aaronbg@uwm.edu
in developed countries, health care systems must ensure their      Research Objective: The Centers for Disease Control and
services are providing quality and appropriate services to all.    Prevention (CDC) HIV/AIDS surveillance data have identified
In a previous study, we demonstrated that there were               that HIV/AIDS as a health crisis for African American men. In
significant differences in functional status 6 months post         order to develop effective culturally appropriate intervention
stroke between Maori (the indigenous people of New                 programs for black males living with HIV/AIDS, it is essential
Zealand), Pacific Islanders, Asians and New Zealand                to identify the factors associated with quality of life in this
Europeans. The purpose of this study is to examine whether         population. The objectives of this study were to: (a) identify
these differences arise from differences in health service usage   the interpersonal and structural barriers to health care and
and pathway of care post stroke and the extent to which these      social services, (b) examine the relationship between the
differences are associated with individual and community level     contextual variables, the health system factors, and HRQOL
measures of socioeconomic status                                   variables, and, (c) suggest culturally appropriate strategies for
Study Design: A population-based stroke incidence study.           improvement of quality of life among African American men
Cases were obtained from hospitals, general practitioners, and     living with HIV/AIDS.
community workers. All first-ever-in-a-lifetime cases of stroke    Study Design: Cross-sectional descriptive survey design was
(n=1423) were collected in a population-based register in          used to examine selected socio-demographic variables, their
940,000 residents (aged 15 years) over 12 months in 2002-          association and influence on the quality of life among a
2003 in Auckland, New Zealand. Ethnicity was self-identified       sample of urban African American males. Data were analyzed
and grouped in this analysis as New Zealand European,              using descriptive statistics, followed by correlations and
Maori, Pacific Peoples, Asian and other ethnic groups.             hierarchical multiple regressions statistics.
Population Studied: All residents of Auckland, New Zealand         Population Studied: Participants were recruited through an
who suffered a stroke during the 12 month calendar period in       African American community-based organization in an urban
2002-2003.                                                         setting of a large Midwestern state. A snowball sampling
Principal Findings: There were significant differences in          technique was used to obtain a sample of (n=55) African
functional status at 6 months between Maori, Pacific               American men (18 years or older) who self-reported as being
Islanders, Asians and New Zealand Europeans. The results           HIV-infected. The sample represented a homogenous group
show the extent that these differences between ethnicities are     of HIV-infected men in regards to sexual orientation, stage of
explained by the following: likelihood of being admitted to        the HIV disease, and years since diagnosis. The men
residential care post-stroke, use of health services and family    completed self-reported instruments to assess socio-
support. The results also provide evidence on barriers that        demographic and contextual factors; self-rated health status,
exist to accessing services, including information provided at     health systems barriers, social stigma, and health related
the time of stroke, financial concerns and family support.         quality of life which comprised nine domains.
Conclusions: Identifying the pathway of care and use of            Principal Findings: Participants ranged in age from 18-66
health services can help explain differences in health services    years, (M = 48.84, SD= 7.67) and the average length of time
usage between people of different ethnic backgrounds.              since diagnosis with HIV infection was (M = 10.79 years SD=
Implications for Policy, Delivery, or Practice: The results        6.4). Most men (72.1%) had a high school diploma and about
add to our understanding of the reason for differences             half (50.5%) had some college or were college graduates.
between ethnic groups in quality of life and functional status     Majority of the participants (88.9%) were unemployed and
post stroke. The implications for the design and delivery of       single (83.6%). The majority of the men (85%) were enrolled
health services is discussed.                                      either in a federal or state health insurance program.
Primary Funding Source: NZ Health Research Council                 Contextual variables of income, education, age, and residential
                                                                   situation were significantly correlated with the HRQOL sub-
                                                                   scale of provider trust. The better the participants rated their
                                                                   own health status the better they rated their general
                                                                   satisfaction with life, overall life functioning, and were less
                                                                   worried about their health. Clinical stage of their disease was
                                                                   also negatively associated with life satisfaction and with health
                                                                   worries. As barriers to healthcare and social stigma increased,
                                                                   HRQOL decreased. In hierarchical multiple regression
                                                                   analysis, contextual factors, health systems, and interpersonal
                                                                   factors accounted for between 19 and 63% of the variance in
                                                                   health related quality of life (HRQOL) measures.
Conclusions: Critical to improving the nation’s health, is the      migrants, disabled or differently abled and indigenous
understanding of disparities in HIV/AIDS in minority                population.
populations. Furthermore, the high rates of new HIV                 Population Studied: Women,Elderly, Migrants, Disabled,
infections among African American men and differences in            Indigenous population namely the scheduled castes and
survival rates for HIV infected African American men require        scheduled tribes in India
focused research to understand factors leading to disparities       Principal Findings: There are health disparities between the
and to develop interventions to maximize health                     identified groups and the other population in the Indian
consequences associated with HIV infection and co-morbid            society. The dimensions and determinants of health
conditions.                                                         disparities are varied ranging from structural, socio-cultural to
Implications for Policy, Delivery, or Practice: As AIDS             political resulting in their exclusion from the mainstream
becomes a manageable chronic condition it has become                society. High rates of morbidity and mortality are found
necessary to develop culturally appropriate, empirically            among the vulnerable groups; they face restricted and denial
validated models that will assist health practitioners in           of their access and utilization of health care services as
working to improve HRQOL among HIV infected African                 compared to others. The health policies have also excluded
American men. Therefore, it is imperative for health care           them. The trends and patterns of health care financing clearly
researchers and practitioners to increase their understanding       shows a link between low expenditure and poor health
of their contribution to improving the HRQOL in this                outcomes among the vulnerable groups. The health
population. The findings from this study indicated broad and        inequities among the various groups is further worsened in
far reaching effect of the health systems and interpersonal         the context of the global changes and challenges, the growth
factors. The measures of structural barriers and social stigma      of the private sector and the creation of health markets, rising
describe what patients have experienced as the result of their      drug prices,the socio-political structures and the relationships
interaction with the medical and social services systems. The       that differentially affects people’s chances of being healthy
HRQOL framework presented in this study serves as one               within a given society, the interrelationships between factors
paradigm, explaining almost two-thirds of the variance in           at the individual level and within the social context that
some of the HRQOL measures in a community sample of                 increase or decrease the likelihood of achieving and
HIV-infected African American men. Future research with             maintaining good health, characteristics of the healthcare
African American men should also explore whether the                system that influence health equity and ineffective policy
HRQOL model explains QOL in men living with other non-              creating health disparities in terms of health outcomes, access
stigmatizing diseases other than HIV/AIDS.                          and utilization among various groups.The research in addition
Primary Funding Source: Sigma Theta Tau International               also highlights gaps in the National Health Policy that have
Honor Society of Nursing "Eta Nu Chapter"                           restricted access and inverted the health outcomes negatively.
                                                                    Conclusions: There are persistent inequities in health
●Health Inequities among Vulnerable Groups in India                 outcomes and health services in terms of its availability in
Chandrima Chatterjee, Ph.D.                                         sufficient quantity, and accessibility i.e., nondiscrimination,
                                                                    physical accessibility, economic accessibility /affordability,
Presented By: Chandrima Chatterjee, Ph.D, Research Officer,         information accessibility among various groups in India. There
Research, Centre for Enquiry into Health and Allied Themes          is differential distributions of health among different groups
(CEHAT), Survey No 2804 & 2805, Aram Society Road, Vakola,          and that speaks of the low level of realization of rights.
Santacruz (E), Mumbai, 400055; Tel: (+91) (0)9821736315;            Identifying barriers to health and providing insights on multi-
Fax: (+91) (022) 26673156; Email:                                   sectoral approaches in planning and interventions to support
chandrima_bchatterjee@yahoo.com                                     the most vulnerable populations can act as a good strategy in
Research Objective: The main objective of the research on           the movement towards the constitutionalization of the right
vulnerable groups in India was to generate disaggregated            to health as a human right.
information on health disparities among various groups from         Implications for Policy, Delivery, or Practice: The research
the human rights perspective. This was to create equity             makes some specific recommendations towards provisions of
indicators, to integrate equity strata and health indicators into   care and utilization of health services. It suggests a strategy to
a number of existing data sources, including censuses, vital        operationalize right to healthcare in India.
registries, household surveys, small areas data, and                Primary Funding Source: Co-funded by Ford, NOVIB and
administrative data sources. The information is meant to            Rangoonwala Trust
assist the preparation of a framework for a draft bill, inform
public policy, impact the budget process, strengthen                ●Linguistic Concordance in Mental Health Services to
appropriate budget allocations and formulate an operational         Chinese Immigrants: Does It Reduce Disparity?
strategy to move towards the goal of Right to Health Care in        Alice Chen, Ph.D.(C), Arminee Kazanjian, DrSoc
India. It would facilitate advocacy campaign and public
understanding on issues involving health and human rights           Presented By: Alice Chen, Ph.D.(C), , Health Care and
violations in India.                                                Epidemiology, University of British Columbia, 5804 Fairview
Study Design: The information have been collected through           Avenue, Vancouver, B.C., V6T 1Z3; Tel: 604-822-2729; Fax:
the various available sources of macro data ; the National          604-822-4994; Email: achen@chspr.ubc.ca
Family Health Services, Reproductive and Child Health (A            Research Objective: Recruiting bilingual and bicultural
household survey), the Census India, apart from micro studies       professionals is often suggested as a strategy to address the
on health of the vulnerable groups namely women, elderly,           disparity in mental health service utilization reported among
                                                                    Asians or immigrants in North America. The objective of this
study is to examine the relationship between the Chinese           ●Mental Illnesses and Perforating Appendicitis
language skill of physicians and the rate of mental health         Yiing-Jenq Chou, M.D., Ph.D., Cheng-Hua Lee, M.D., DrPH,
consultations and likelihood of mental health diagnoses            Nicole Huang, Ph.D., Yea-Jen Hsu, MS, Pen-Jen Wang, Ph.D.
among recent Chinese immigrants to the Canadian province
of British Columbia (BC).                                          Presented By: Yiing-Jenq Chou, M.D., Ph.D., Associate
Study Design: This study is a secondary analysis of three          Professor, Department of Social Medicine, National Yang
linked administrative databases: (1) immigration database –        Ming University, 155, Li-Nong Street, Section 2, Taipei, 112;
all immigrants who landed in BC in 1985-2000(2) provincial         Tel: 886-2-28201458; Fax: 886-2-28261002; Email:
physician payment database – most physician services to BC         yjchou@ym.edu.tw
residents under Canada’s publicly-funded health care system        Research Objective: Mentally ill patients are postulated to be
(3) College of Physicians and Surgeons register – information      at a higher risk of hindered access to care for their physical
to identify the Chinese language skill of the physicians. All      diseases as the psychosis of their mental illness may impair
services provided to the study population by general               the patient’s capacity to recognize or articulate emerging
practitioners (GPs) and psychiatrists in 1992-2001 were            medical illness. More importantly, not all mentally ill patients
included in the analyses.                                          are equally vulnerable. In addition to other cognitive
Population Studied: The study population consists of               impairment and social manifestations, pain insensitivity in
Chinese immigrants who landed in BC in 1985-2000 and a sex         schizophrenic patients may make them particularly vulnerable
and age matched control group randomly selected from the           to adverse consequences of physical diseases that are typically
enrollees of the provincial health insurance plan. Over            pain-causing. As the National Health Insurance program in
150,000 Chinese immigrants and the same number of                  Taiwan has successfully reduced financial barriers of access to
controls were included in the study.                               care for all including mentally ill patients, the purpose of this
Principal Findings: Chinese-speaking GPs and psychiatrists         study was to investigate whether non-financial disparities still
provided 87% and 67%, respectively, of all the consultations       exist between patients with and without mental illness, and
to Chinese immigrants, and 23% and 8% of all consultations         more specifically, between patients with schizophrenia and
to controls. Relative to consultation with a non-Chinese-          other major mental illness.
speaking GP, consultation with a Chinese-speaking GP was           Study Design: This was a population-based case-control
less likely to be identified as being for mental health reasons,   study.
for both immigrants and controls alike (OR=0.78 and 0.43).         Population Studied: Using the National Health Insurance
Compared to non-Chinese-speaking psychiatrists, Chinese-           hospital-discharge data, we compared the likelihood of
speaking psychiatrists were less likely to make the diagnosis of   perforated appendix among 100,578 adults aged 15 and above
depression in their consultations to both immigrants and           who were hospitalized for acute appendicitis in Taiwan during
controls (OR=0.35 and 0.40). Higher proportion of overall          the years from 1997 to 2001. The outcome measurement was
care received from Chinese-speaking GPs was associated with        appendiceal rupture (ICD-9-CM codes 540.0 and 540.1) in all
lower rates of mental health consultation among immigrants         patients hospitalized for acute appendicitis. Appendicitis, not
and controls (RR =0.65 and 0.32) and lower odds of ever            resulting in ruptured appendix, was defined as ICD-9-CM code
being diagnosed with depressive disorder (OR =0.52 and             540.9. Independent variables included mental disorder status,
0.35).                                                             personal characteristics (age, sex, ethnicity and
Conclusions: Chinese-speaking GPs and psychiatrists provide        socioeconomic status (SES)) and hospital characteristics
the majority of general and mental health care to recent           (ownership, accreditation level, patient volume and
Chinese immigrants in BC. However, Chinese-speaking                geographic location).
physicians are less likely to make a mental health diagnosis or    Principal Findings: After adjusting for age, gender, ethnicity,
treat mental health problems, especially depressive disorder.      socioeconomic status (SES) and hospital characteristics, the
Subjects who more often visit Chinese-speaking GPs receive         presence of schizophrenia was associated with a 2.80 times
fewer mental health services; the impact is stronger among         higher risk of ruptured appendix (odds ratio [OR], 2.80; 95%
Chinese immigrants who rely on Chinese-speaking GPs for            confidence interval [CI], 2.18-3.61). However, the presence of
most of their care.                                                affective psychoses (OR, 1.15; 95% CI: 0.77-1.73) or other
Implications for Policy, Delivery, or Practice: While              mental disorders (OR, 1.64; 95% CI: 0.93-2.91) was not
linguistic concordance may enhance the quality of health care      significant predictor of ruptured appendix.
in many aspects, bilingual practitioners do not necessarily        Conclusions: These findings suggest that given mentally-ill
reduce the disparities in mental health care experienced by        patients having no financial access barriers, they are still in
linguistic minorities. The practice patterns of these bilingual    disadvantage to obtain timely treatments for their physical
professionals must be considered when evaluating cultural          diseases. Of major mental illnesses, schizophrenia, due to its
and linguistic access to care. Cultural values and place of        insensitivity to physical pain associated with illness or injury,
training of the service providers, which are closely related to    is most vulnerable to the diseases whose main presenting
language skills, may be crucial influences on practice             signs typically include pain.
behaviors. These factors need to be addressed in cultural          Implications for Policy, Delivery, or Practice: Taiwan’s
competence education and recruitment programs of foreign-          experience suggests that improving access to care of mentally-
trained professionals if linguistic concordance is to become an    ill patients is more than a financial matter. Health care
effective strategy in reducing disparities.                        personnel in either psychiatric hospitals and general hospitals
Primary Funding Source: Canadian Institutes of Health              who treat patients with severe mental disorders should be
Research                                                           able to recognize unique consequences of major mental
                                                                   illnesses which may confound a diagnosis of a serious,
potentially life-threatening physical disease, and pay more         or providing bilingual staff with training as professional
attention to these patients with minimal pain and objective         interpreters may be necessary to improve the quality of
signs, in order to minimize possible misdiagnosis or delays in      communication with providers.
treatment. Case management programs or a better integrated          Primary Funding Source: RWJF
care delivery system for severe psychiatric patients may also
serve as powerful aids. For mentally-ill patients, not only their   ●Health Characteristics and Health Care Access Among
access to mental health services is important, but also their       Major Racial/Ethnic-Nativity Groups in the United States:
access to non-mental care is equally critical for their health.     Results from the 1998-2003 National Health Interview
Primary Funding Source: No Funding                                  Survey
                                                                    Achintya Dey, M.A., Jacqueline W. Lucas, M.P.H., Jeannine S.
●Low-Income Parent’s Reports of Communication                       Schiller, M.P.H.
Problems with Health Care Providers: Effects of Language
and Insurance                                                       Presented By: Achintya Dey, M.A., Statistician (Demography),
Lisa Clemans-Cope, Ph.D., Genevieve Kenney, M.A., Ph.D.             Center for Disease Control and Prevention, National Center
                                                                    for Health Statistics, 3311 Toledo Road, Room 2330,
Presented By: Lisa Clemans-Cope, Ph.D., Research Associate,         Hyattsville, MD 20782; Tel: (301)458-4251; Fax: (301)458-4961;
Health Policy Center, The Urban Institute, 2100 M Street NW,        Email: aad2@cdc.gov
Washington, DC 20037; Tel: (202) 261-5580; Fax: (202) 223-          Research Objective: The prevalences of selected physical and
1149; Email: lclemans@ui.urban.org                                  mental health characteristics, as well access to care are
Research Objective: This study examines how parental                compared among the U.S.-born and foreign-born adults by
reports of communication problems with health providers vary        four major racial/ethnic subgroups. The association between
among subgroups of low-income children.                             immigrant health status and length of stay in the U.S. for
Study Design: The data were drawn from the 1999 and 2002            these subgroups is also examined.
National Survey of America’s Families (NSAF). Bivariate and         Study Design: Data from the 1998-2003 Sample Adult
multivariate analyses, including linear probability models,         component of the National Health Interview Survey (NHIS)
were used to estimate the relationships between parental            conducted by the Center for Disease Control and Prevention's
report of communication problems with health providers and          (CDC) National Center for Health Statistics (NCHS) were
insurance type, socio-economic, demographic, health factors,        used to assess demographic and health characteristics of four
and provider type.                                                  major racial/ethnic subgroups of U.S.-born and foreign-born
Population Studied: We concentrate our analyses on low-             adults. Data were analyzed using bivariate and multivariate
income families in 1999 and 2002. In addition, we study             techniques with the SUDAAN software package.
publicly insured Latino children with a Spanish-speaking            Population Studied: Four major racial/ethnic subgroups of
parent, a sub-group of low-income children that has not been        U.S.-born and foreign-born adults: Hispanic, non-Hispanic
isolated in prior research, despite being a focus of public         white, non-Hispanic black and non-Hispanic Asian adults
policy.                                                             Principal Findings: Preliminary age-adjusted results show
Principal Findings: Almost one-quarter of low-income                that access to care differs for U.S.-born and immigrant adults.
children had parents who reported poor communication with           However, despite having limited access to health care and
health care providers, and for children whose parents were          being uninsured, immigrant adults had significantly better
interviewed in Spanish, the figure was 40 percent. Parents          health than their U.S.-born counterparts. The results also
who were foreign-born or who had poor mental health scores          indicate that health status, chronic health conditions, and
were significantly more likely to report communication              health behaviors varied across the immigrant racial/ethnic
problems. Almost two-thirds (63 percent) of the sample of           subgroups by length of stay in the United States.
low-income publicly covered Latino children whose parents           Conclusions: This analysis describes differences in various
were interviewed in Spanish (n=1,632) report clinics or             sociodemographic, physical, and mental health characteristics
hospital outpatient departments as their usual source of care.      among four racial/ethnic subgroups of U.S.-born versus
For this population, these provider types were associated with      foreign-born adults. The results highlight important
increased communication problems compared to other                  differences between immigrant and native-born populations,
provider types.                                                     and underscore the importance of nativity as a factor in health
Conclusions: These findings indicate that communication             differences among various racial and ethnic populations
problems are more prevalent for children whose parents are          Implications for Policy, Delivery, or Practice:
foreign-born and for those whose parents chose to be                Understanding these differences will assist policy makers in
interviewed in Spanish relative to other low-income children.       developing culturally diverse community intervention
The growing share of the population that is foreign-born or         programs and allocating resources to fulfill health care needs.
that has limited English proficiency suggests that                  Primary Funding Source: CDC
communication problems between parents and health care
providers are likely to increase over time.
Implications for Policy, Delivery, or Practice: In the short
term, improving access to translation services, particularly at
clinics and hospital outpatient departments could reduce
communication problems and improve access for low-income
families with limited English proficiency. In the long run,
increasing the number of providers who are Spanish speakers
●Differential Reporting of Similar Levels of Health among            Implications for Policy, Delivery, or Practice: Differential
Sub-groups of the U.S. Population.                                   health status reporting by subgroups of the US population
Martey Dodoo, Ph.D.                                                  presents a potential problem for the validity of health equality
                                                                     measurements and health level comparisons. Health survey
Presented By: Martey Dodoo, Ph.D., Senior Economist, The             designers would have to supplement national health survey
Robert Graham Center, 1350 Connecticut Ave, N.W. Ste 201,            questions on self assessed health status with questions other
Washington, DC 20036; Tel: 202-331-3360; Fax: 202-331-3374;          than those that elicit self-reports of health. Suggestions of
Email: mdodoo@aafp.org                                               supplementary questions will be provided in the presentation.
Research Objective: To answer the following questions: Is            Primary Funding Source: No Funding
there differential health status reporting within US population
groups? If yes, how can we assess it, and how is reporting           ●Access and American Indians: Is Rural Better?
different - e.g.: cut-point shift or index function shift or both?   Mark Doescher, M.D., M.S.P.H., J. Elizabeth Jackson, MA,
What are the solutions for problems imposed by such                  Dorothy A. Rhoades, M.D., M.P.H.
differential reporting in health surveys?
Study Design: To assess individual health status, a popular          Presented By: Mark Doescher, M.D., MSPH, Associate
survey question is: “In general, is your health excellent, very      Professor, Family Medicine, University of Washington, 4225
good, good, fair, or poor?” However, researchers have found          Roosevelt Way NE, Suite 308, Seattle, WA 98105; Tel: (206)
mainly in studies outside the US that sub-populations with           616-9207; Fax: (206) 685-0610;
same health levels use different thresholds based on culture,        Email: mdoesche@u.washington.edu
age, gender, education, language, and illness experience, to         Research Objective: The health status of American Indians
choose their health status category. Such reporting                  (AI) is among the worst in the nation. Although the majority of
heterogeneity if found to exist in this country, may invalidate      AI live in urban areas, access to care for urban AI has never
some group comparisons and measures of health inequality.            been reported for a nationally representative sample. Most
We assessed differential reporting of self assessed health           health services research among AI has been conducted
status (SAH) in the Medical Expenditure Panel Survey (MEPS)          among rural or reservation-based populations. Unlike rural AI,
using two methods based on mean comparison hypothesis                most urban AI do not receive health care services from
testing, and ordered response logit model estimation. Mean           facilities operated or fully funded by the Indian Health Service.
comparison hypothesis testing was used to ascertain if there         The objective is to test the hypothesis that urban AI face more
is differential reporting. To determine how reporting is             barriers to care than rural AI or other racial/ethnic groups by
different, we stratified the sample into sub groups and              estimating the prevalence of and trends in: (1) having a
estimated ordered response models for each sub group. For            regular physician; and (2) reporting unmet medical need.
each sub group we obtained estimates of the index function           Study Design: National data from the Behavioral Risk Factor
and the cut points based on the model by Lindeboom and               Surveillance System (BRFSS) were used to examine reports of
Doorslaer (2004). We compared between sub-groups using               having a regular physician and unmet need for care in the
MEPS data by conditioning on the more objective indicators           preceding 12 months among adults aged 18 and older
of health including SF-12 scores.                                    (n=234,315). Race/ethnicity was categorized as American
Lindeboom, Maarten & van Doorslaer, Eddy, 2004. "Cut-point           Indian, African American, Hispanic, Asian/Pacific Islander,
shift and index shift in self-reported health," Journal of Health    and white. The urban-rural continuum was grouped as
Economics, Elsevier, vol. 23(6), pages 1083-1099.                    metropolitan, nonmetropolitan adjacent to a metropolitan
Population Studied: A national probability sample of 19,896          area, rural “micropolitan”, and rural “non-core” by 2003
persons 18-64 years old from the 2002 MEPS.                          Urban Influence Code categories. Analyses were adjusted for
Principal Findings: We found that among US population                sociodemographic characteristics.
group pairs at the same level of health (as measured by SF-12        Population Studied: A national sample of non-
scores), persons in the group with low education or low              institutionalized US adultsm from 50 states and the District of
income tend to under-estimate their health relative to persons       Columbia.
with high education or high income respectively. On the other        Principal Findings: In 2002 for all races combined, 20.8% of
hand, younger persons in the 18-40 year age group tend to            metropolitan, 18.2% of nonadjacent, nonmetropolitan, 20.1%
over-estimate their health relative to persons 41-65 years old.      of micropolitan, and 19.3% of non-core respondents lacked a
We also found that (1) the odds of 18-40 year olds being in a        regular physician. Among AI, 21.6% of metropolitan, 22.6% of
higher health category is about twice the odds of 41-65 year         nonmetropolitan adjacent, 15.7% of micropolitan, and 17.5%
olds, (2) the odds of persons with low education being in a          of non-core respondents lacked a regular physician. For all
higher health category is only about 65% of those with higher        races combined, 6.7% of metropolitan, 6.8% or nonadjacent,
education, and (3) the odds of persons with low income being         nonmetropolitan, 5.8% of micropolitan, and 6.1% of non-core
in higher health category is only about 45% of those with            respondents did not receive needed care. Among AI, 8.0% of
higher income                                                        metropolitan, 8.1% of nonmetropolitan, 4.1% of micropolitan
Conclusions: There is certainly differential health status           and 6.1% of non-core respondents reported unmet care. Also,
reporting within US population groups. The differences are           trend analyses and analyses categorizing county of residence
mainly with respect to education, income and age. We could           by IHS catchment area will be presented.
not ascertain if the differences are due to cut-point shift or       Conclusions: AI residing in the most isolated rural locations
index function shift. Further work within the next month, may        had better access to care than other racial/ethnic groups.
shed more light on that (to be reported at the Annual                However, AI residing in metropolitan counties and rural
Meeting).
counties adjacent to metropolitan ones reported more               1.9,3.7) and respondents with the least trust in their PCPs (OR
difficulty accessing care than their more rural counterparts.      2.8, 95% CI 1.8,4.4) thought it was likely that they might be
Implications for Policy, Delivery, or Practice: While the          used a “guinea pig” without permission. African American
Indian Health Service does provide some support for urban          race (OR 1.9, 95% CI 1.3,2.6) and the lowest level of trust in
Indian health programs, it is already chronically under funded     one’s PCP (OR 1.7, 95% CI 1.1,2.7) were also associated with
and allocates less than 2% of its budget to urban programs.        concerns of unknowingly being involved in an experiment
Strategies for improving access to care for urban AI without       through the receipt of prescription medications. When trust in
compromising existing rural programs are needed.                   one’s PCP was removed from these models, however, there
Primary Funding Source: Federal Office of Rural Health             was no significant change in the coefficient for African
Policy                                                             American race. Neither a history of discrimination in health
                                                                   care or knowledge of the Tuskegee Syphilis Experiment was
●Racial Differences in Distrust in Clinical Research Among         related to concerns about unwanted experimentation in the
Elderly Whites and African Americans                               three societal distrust items in multivariable models.
Raegan Durant, M.D., M.P.H., Anna T. R. Legedza, ScD,              Conclusions: Older African Americans have higher levels of
Giselle Corbie-Smith, M.D., MSc, Marcie B. Freeman, MEd,           societal distrust in clinical research compared to older whites.
Bruce E. Landon, M.D., M.B.A.                                      Race and trust in PCPs are independently related to societal
                                                                   distrust in clinical research.
Presented By: Raegan Durant, M.D., M.P.H., Clinical                Implications for Policy, Delivery, or Practice: Improving
Research Fellow, Division of General Medicine and Primary          African Americans’ trust in their PCPs alone may not eliminate
Care, Beth Israel Deaconess Medical Center, 330 Brookline          racial differences in societal distrust.
Avenue, Rose 111, Boston, MA 02215; Tel: 617-667-0487; Fax:        Primary Funding Source: HRSA
617-667-2751; Email: rdurant@bidmc.harvard.edu
Research Objective: Some previous studies have suggested           ●Health Literacy Measurement Issues for Homeless
that African Americans are more distrustful of clinical research   Young People
than whites. We sought to determine if African Americans’          Josephine Ensign, M.P.H., Dr.P.H.
distrust in clinical research arises from their general
perceptions of physicians and the health care system or from       Presented By: Josephine Ensign, M.P.H., DrP.H., Associate
their relationships with individual physicians.                    Professor, Psychosocial and Community Health, University of
Study Design: We assessed responses to 7 previously                Washington, UW Box 357263, Seattle, WA 98195; Tel: 206-685-
validated items from a distrust index examining attitudes          0816; Fax: 206-685-9551; Email: bjensign@u.washington.edu
about clinical research in terms of societal distrust (e.g.        Research Objective: In order to design appropriate clinic-
general perceptions of physicians and research) and                based measures of and interventions to improve health
interpersonal distrust (e.g. perceptions of research relating to   literacy for homeless young people, the perspectives of youth
one's relationship with an individual physician). Trust in         affected is important. The objective of this study was to
one’s primary care provider (PCP) was assessed as an               document the views of homeless youth on their health
independent variable using the 8-item trust subscale from the      information needs and preferred methods for having health
Primary Care Assessment Survey (PCAS). Respondents were            literacy assessed in a primary care setting.
grouped into quartiles of trust in a PCP. We also examined         Study Design: The research was a pen and paper anonymous
other independent variables including knowledge of the             self-administered clinic-based survey. The survey included
Tuskegee Syphilis Experiment and personal experiences with         forced-choice and open-ended questions, and was readability
discrimination in health care. We performed bivariable and         tested at the 5th grade level using the Flesch-Kincaid formula.
multivariable analyses to examine the associations between         Youth were provided with assistance reading the questions by
these factors and responses to the items in the distrust index.    a youth outreach worker familiar to them but not associated
Population Studied: We conducted a mail survey of                  with the research team.
community-dwelling elderly (age >50) whites and African            Population Studied: The research was conducted over a one
Americans in the Boston area.                                      month time period (September 2005) with homeless youth
Principal Findings: We received responses from 829 eligible        ages 13-24 years in a large West-coast city. The homeless
persons (58% response rate). The study population was 60%          young people who participated were recruited from a
white and 40% African American. For two of the items               homeless youth primary care clinic waiting room (either as
assessing societal distrust, African Americans were more likely    patients or as friends accompanying patients). The total
than whites to think that they could potentially be “used as       sample was 35 homeless youth (20 female and 15 male youth),
guinea pigs” without giving their permission (54% vs. 28%,         age range 14-24 years, with an average age of 20 years.
p<0.0001) or to think that health care providers, in general,      Principal Findings: In response to the question “Sometimes
prescribe medications as a way of “experimenting on people         people need extra help in reading or understanding how to
without your permission” (58% vs. 41%, p<0.0001). However,         take medicines or other health information, but are too shy to
in the 3rd item assessing societal distrust, African Americans     ask. What do you think is the best way for us to find out if
were less likely than whites to believe that they had actually     someone needs help?”, all but two of the respondents circled
ever received treatment from any health care provider as part      “ask people in person in a private room”. Of the two who
of an experiment without giving permission (1.6% vs. 2.9%,         circled “have everyone fill out a short form before they are
p=0.23). There were no racial differences in responses to the      seen in clinic”, one young person wrote in “But I know a lot of
index items focused on interpersonal distrust. In a                people have a hard time reading, so this wouldn’t work so
multivariable model, African Americans (OR 2.6, 95% CI             well”. None circled “don’t ask because it’s too embarrassing”.
In addition, the majority (n=30) stated that they used public       employed, and 58% had at least a high school education. The
libraries, mainly to look up health and job information on the      mean family income/poverty ratio was 2.41.
internet as well as to check e-mail. Most of the young people       Principal Findings: Individual characteristics including
reported that their main sources of health information were         income (p < .001), age (p < .001), education (p < .001), being
the clinic, friends, and the internet. There were no differences    male (p < .001) and being Hispanic (p < .001) were negatively
in responses to the survey by gender.                               associated with sedentary life style. Even after adjusting for
Conclusions: The homeless youth included in this study              these individual characteristics, residing in a census tract with
identified internet access through public libraries as a major      a higher percentage of households with incomes below the
source of health information. In addition, all indicated that       federal poverty line was associated with higher likelihood of
clinicians should attempt to asses a young person’s health          sedentary lifestyle (OR= 5.41, p < .0001). Age (p <.001),
literacy. Most young people thought that health literacy            employment (p < .001), education (p < .05) and being
assessment should occur via a one-on-one assessment by a            Hispanic (p < .03) were negatively associated with binge
clinician in a private room.                                        drinking, whereas being male (p<.001) was positively
Implications for Policy, Delivery, or Practice: The results of      associated. Residing in a census tract with a higher
this research indicate a need for the development of brief          percentage of households in poverty was associated with
health literacy assessments appropriate for use in primary          more frequent binge drinking adjusting for individual
care situations with homeless young people. In addition, the        characteristics (1.5 more days of binge drinking for every 10%
results indicate that internet access through public libraries is   increase in the poverty rate; p < .0001). Results from
an important source of health information for homeless young        additional analyses examining smoking will be presented
people. Health and media literacy programs targeting                along with analyses of two additional tract-level characteristics
homeless youth could benefit from a partnership between             (the percentage of census tract with less than a high school
public librarians and clinicians specializing in primary care for   education and median household income).
homeless adolescents and young adults.                              Conclusions: The association of neighborhood poverty with a
Primary Funding Source: AHRQ                                        lack of physical activity and more binge drinking, controlling
                                                                    for individual characteristics, suggests pathways through
●Neighborhood Socioeconomic Characteristics and                     which neighborhoods may affect health. Potential
Individual Health Behaviors                                         mechanisms for these findings include increased fear of
Jose Escarce, M.D., Chloe E. Bird, Ph.D., Brian Finch, Ph.D., D.    crime, fewer parks, more crowding, and more alcohol outlets
Phuong Do, M.Phil, Teresa Seeman, Ph.D., Nicole Lurie, M.D.,        and advertising in poor neighborhoods.
M.S.P.H.                                                            Implications for Policy, Delivery, or Practice: Neighborhood
                                                                    effects represent one pathway through which neighborhoods
Presented By: Jose Escarce, M.D., Professor of Medicine,            ultimately affect health and thus a potential avenue for
Medicine, UCLA, 911 Broxton Ave, Los Angeles, CA 90024;             effective intervention. The mechanisms through which these
Tel: 310 794 3842; Fax: 310 794 0726; Email:                        effects are produced need to be explored.
jescarce@mednet.ucla.edu                                            Primary Funding Source: NIEHS
Research Objective: A growing body of research
demonstrates relationships between residential neighborhood         ●Poor Health in the South: An Examination of
context and health. However, few studies have examined the          Socioeconomic Factors in Eight Southern States.
effect of neighborhoods on health behaviors. We examined            Rachel Ferencik, MPA, Karen Minyard, Ph.D., Kate Stewart,
the extent to which neighborhood poverty is related to binge        M.D, M.P.H., Mei Zhou, M.A, M.S.
drinking and sedentary lifestyle, independent of individual
characteristics.                                                    Presented By: Rachel Ferencik, MPA, Research Associate, ,
Study Design: Using 3-level hierarchical logistic and linear        Georgia Health Policy Center, 14 Marietta St., Suite 221,
regression, we analyzed National Health and Nutritional             Atlanta, GA 30303; Tel: (404)651-1294; Fax: (404)651-3147;
Examination Survey III (NHANES) interview and laboratory            Email: rachel@gsu.edu
data, merged with data on sociodemographic characteristics          Research Objective: The Southern Regional Health
of their residential census tract. Binge drinking was measured      Consortium is interested in understanding how
as the number of days in the past year that one consumed 5 or       socioeconomic factors impact health and whether the south
more drinks (mean = 14.1, with 25% reporting having done so         carries a disproportionate burden than the rest of the country
once or more). Sedentary lifestyle was measured as whether          in lower socioeconomic attainment. Eight of the most rural
or not the subject participated in any moderate or vigorous         and medically underserved states in the country, primarily
physical activity in the past month (75% report having              located in the south, were funded under Robert Wood
engaged in moderate or vigorous activity at least once in the       Johnson’s Southern Rural Access Program to improve access
past month).                                                        to health care. (These states -- Alabama, Arkansas, Georgia,
Population Studied: Our analyses included 14,626 adults             Louisiana, Mississippi, South Carolina, Texas and West
from 83 counties and 1805 census tracts, who completed              Virginia -- now organize themselves as the Southern Regional
surveys and medical exams, were not missing on key                  Health Consortium, also known as “SRHC”).
components of the outcome measures, and for whom                    Study Design: A comprehensive literature review was
residential census tract could be geocoded. The sample was          conducted to understand the relationship of socioeconomic
47% male; 43% white, 27% black, 26% Hispanic, 4% other.             factors on health. Using BRFSS, CPS and ARF data,
Subjects ranged in age from 19.5 to 90 (mean = 48); 53% were        researchers collected and analyzed data on sociodemographic
                                                                    factors, health access and coverage, and health status of these
eight states to determine differences between the SRHC states       Study Design: We used matching self-report and claims data
and the national average.                                           from the Medicare Current Beneficiary Survey, 1999-2002, for
Population Studied: Data were analyzed from the eight states        6 preventive services: cholesterol testing, colorectal cancer
that comprise the Southern Regional Health Consortium,              screening (fecal occult blood testing or
which include Alabama, Arkansas, Georgia, Louisiana,                sigmoidoscopy/colonoscopy), influenza vaccination,
Mississippi, South Carolina, Texas and West Virginia in             mammography, Pap smear screening, and prostate specific
comparison to the rest of the nation.                               antigen screening. In multivariate analyses, we examined
Principal Findings: We used a causal framework developed            predictors of “over-reporting” (self-report in the absence of a
by Williams and House (2000) to test our hypothesis that            claim) and “under-reporting” (claim in the absence of self-
socioeconomic factors play an important role in health              report).
disparities and that the south, compared to the nation, carries     Population Studied: Medicare Beneficiaries 65 and older.
a disproportionate burden of lower socioeconomic                    Principal Findings: With the exception of PSA screening,
attainment. SRHC states had larger minority populations than        racial and ethnic disparities are larger based on Medicare
the national average (21% v. 29% respectively). Residents in        claims than self-report. Absolute differences in age-adjusted
SRHC states were less educated than the national average.           receipt of preventive services between whites and minorities
Almost 10% of the population 25 years and older did not             for self-report ranged from -3% for cholesterol screening to
complete high school, compared to 8% of the national                18% influenza vaccination. In contrast, differences based on
average. In SRHC states, 21% of the population 25 years of          claims ranged from 6% for cholesterol screening to 20% for
age and older had completed a college education, compared           influenza screening. Multivariate analyses controlling for age,
to 24% of the nation. SRHC states had a smaller white collar        gender, income, educational level, health status, proxy
labor force than the national average. (58% v. 60%                  respondents or supplemental insurance showed that
respectively). More residents in SRHC states live below the         minorities are significantly more likely to report receipt of a
poverty level compared to the national average (15% v. 11%          preventive service in the absence of a supporting claim. Effect
respectively)                                                       sizes ranged from 1.07 (95% CI 0.89- 1.28) for PSA to 1.83
Conclusions: Differences in the socioeconomic factors of            (95% CI 1.49 – 2.25) for Pap smear screening. Rates of
race, education, occupation and income between SRHC states          underreporting were low and didn’t differ by minority status.
and the national average are statistically significant.             Conclusions: These findings are consistent with either
Implications for Policy, Delivery, or Practice: Numerous            racial/ethnic reporting biases in receipt of preventive care (i.e.,
foundations have invested large amounts of money to                 that self-report rates are inflated more in minorities than
improve health status by improving access to health coverage        whites) or differences in billing practices among providers
and care. The academic literature confirms that strategies to       serving majority versus minority patients (i.e., claims are
reduce health disparities are largely outside the control of the    inaccurate to a greater degree in providers with large minority
health sector. The best way to reduce health disparities may        practices).
be to reduce socioeconomic disparities. Access to health care       Implications for Policy, Delivery, or Practice: Assessment of
will not alone close the gap in health disparities if policies do   national progress towards the elimination of racial/ethnic
not also address discrimination and disparities in income,          disparities in preventive care that rely on self-report to monitor
education and occupation. Policies and programs should              success may produce overly optimistic findings. Alternatively,
identify the ways in which broader social, economic and             suboptimal billing practices by providers serving minority
political forces constrain or empower people in their health        patients may yield lower reimbursements and provoke
behaviors. Further research is needed to understand the             spurious assumptions regarding the health care quality they
relative effects on health of race, education, occupation and       provide. Distinguishing these two alternatives will be critical to
income in the south.                                                improving the care of minority patients
Primary Funding Source: The Robert Wood Johnson                     Primary Funding Source: AHRQ
Foundation through the Southern Regional Health
Consortium                                                          ●Disparities in Insurance Coverage, Health Status, and
                                                                    Access Barriers to Care for Children of Immigrant Parents
●Estimating Disparities in Receipt of Preventives:                  Jared Fixmer, BA, Emmanuel Ngui, DrPH, Sandy Tomany-
Comparison of Self-Report and Medicare Claims data.                 Korman, MS, Glenn Flores, M.D.
Kevin Fiscella, M.D., M.P.H., Kathleen Holt, Ph.D., Peter
Franks, M.D.                                                        Presented By: Jared Fixmer, BA, Medical Student, Center for
                                                                    the Advancement of Underserved Children, Medical College of
Presented By: Kevin Fiscella, M.D., M.P.H., Associate               Wisconsin, 8701 Watertown Plank Rd., Milwaukee, WI 53226;
Professor, Family Medicine; Community & Preventive                  Tel: (414)456-8273; Fax: (414)456-6385; Email:
Medicine, University of Rochester, 1381 South Ave, Rochester,       jfixmer@mcw.edu
NY 14620; Tel: (585) 506-9484 ext 106; Fax: (585) 473-2245;         Research Objective: The number of immigrant children in the
Email: kevin_fiscella@urmc.rochester.edu                            US is increasing rapidly, but not enough is known about these
Research Objective: 1) To compare estimates of racial/ethnic        children’s health and access to care. The study objectives were
disparities in receipt of preventive services based on self-        to 1) identify disparities in insurance coverage, health status,
report with estimates of these disparities based on Medicare        and barriers to care (BCs) in children of citizen parents (CPs)
claims data. 2) To assess whether differences in insurance or       and non-citizen parents (NCPs); and 2) determine whether a
socioeconomic factors explain observed discrepancies in these       gradient exists in these outcomes across different NCP
estimates.                                                          children subgroups (legal permanent resident parents
[LPRPs], temporary resident parents [TRPs], and                    ●Racial and Ethnic Disparities in Prenatal Syphilis
undocumented immigrant parents [UIPs]).                            Screening among Florida’s Medicaid Pregnant Women
Study Design: Cross-sectional survey of parents at inner-city,     Norma Gavin, Ph.D., Christina Fowler, Ph.D., E. Kathleen
predominantly Latino community sites in Boston, including          Adams, Ph.D., Monique Chireau, M.D., M.P.H.
supermarkets, hair salons, and laundromats. Participants were
asked 74 questions on insurance coverage, BC, and use of           Presented By: Norma Gavin, Ph.D., Senior Research
health services.                                                   Economist, , RTI International, 3040 Cornwallis Road,
Population Studied: CPs and NCPs of children at urban              Research Triangle Park, NC 27709-2194; Tel: (919) 541-6432;
community sites.                                                   Fax: (919) 990-8454; Email: gavin@rti.org
Principal Findings: Of the 1100 parents interviewed, 40%           Research Objective: Syphilis is one of the nation’s most
were CPs and 60%, NCPs (29% were LPRPs; 24%, UIPs; and             glaring racial disparities in health. Transmission of the disease
7%, TRPs). Compared to CP children, NCP children were              to the fetus in utero can be prevented through routine prenatal
significantly (P<.05) more likely to be in poor/fair/good health   screening and treatment of infected women. Professional
(40 vs. 47%), lack regular doctors (3 vs. 5%), and to be           medical organizations and the CDC recommend prenatal
uninsured (7 vs. 23%), with children of LPRPs (12%), TRPs          syphilis screening in the first trimester for all pregnant women
(27%), and UIPs (37%) significantly more likely to be              and again in the third trimester for women of high risk. This
uninsured. NCP children were more likely than CP children to       study investigated whether the incidence and timing of
not be brought in for needed care due to uninsurance (17 vs.       prenatal syphilis screening differed among Medicaid-covered
9%), inability to afford care (14 vs. 7%), language barriers (9    pregnant women of different races and ethnicity and the
vs. 4%), and immigration issues (8 vs. 1%).                        impact that safety net providers and participation in a primary
In multivariate analyses, NCP children had significantly higher    care case management (PCCM) program had on screening
odds of uninsurance than CP children (OR=2.7 [95% CI, 1.5-         rates among the different racial and ethnic groups.
4.6]); among NCP children there was a gradient, with LPRP          Study Design: We identified syphilis screening tests from
children at lowest risk (OR=1.6 [0.86, 2.98]), followed by TRP     outpatient Medicaid claims data during the 9 months prior to
(OR=3.9 [1.8-8.7]), and UIP children (OR=6.4 [3.3-12.3]). When     delivery. We then computed rates of (1) any syphilis screening
asked to identify the three most important BCs for their           during pregnancy; (2) early syphilis screening, defined as a
children, LPRPs had twice the odds of CPs of citing insurance      test 5 or more months prior to delivery; and (3) repeat syphilis
difficulties (OR=2.1 [1.2-3.6]). NCPs were more likely than CPs    screening, defined as both a test early in pregnancy and a test
to cite language barriers (OR=3.8 [2.3-6.3]), with LPRPs           within 3 months of delivery. For each screening measure, we
(OR=4.4 [2.6-7.5]) and UIPs (OR=2.9 [1.6-5.4]) at highest risk.    ran a logistic regression including race/ethnicity as an
TRPs were more likely to not bring children in for needed care     explanatory variable along with usual source of prenatal care,
due to language barriers (OR=4.1 [1.3-13.1]), while LPRPs were     an indicator for PCCM, and a set of variables to control for
more likely to not bring children for needed care due to long      health status and provider availability. We then ran separate
waits in physicians’ offices (OR=1.5 [1.02-2.2]). LPRP children    regressions for non-Hispanic white, non-Hispanic black and
were less likely to be in poor/fair/good health (OR=0.6 [0.4-      Hispanic women and compared the magnitude and
0.9]). A gradient was observed for certain outcomes, with          significance of the coefficients for the usual source of prenatal
LPRP children experiencing the least severity of these             care and PCCM across the equations.
problems and UIP children the greatest severity.                   Population Studied: We included all women with a Medicaid-
Conclusions: Uninsurance is a major problem and frequent           covered delivery in Florida from October 1, 1999 through
BC for NCP children, with the highest risk among UIP and           September 30, 2000. We excluded women with incomplete
TRP children. Uninsurance, language barriers, and long             claims or enrollment data, including undocumented aliens,
provider waits frequently cause immigrant children to not be       and women with any capitated managed care enrollment or
seen for needed care, and a gradient exists for some               concurrent private insurance or Medicare coverage during
outcomes, with LPRP children experiencing the least severe         pregnancy or delivery. The final sample included 29,542 non-
problems and UIP children the most severe.                         Hispanic white, 16,275 non-Hispanic black, and 27,726
Implications for Policy, Delivery, or Practice: Our findings       Hispanic women.
suggest the need to enhance insurance coverage, reduce BCs,        Principal Findings: Only 57% of women had any prenatal
and provide more efficient pediatric services for NCP children,    syphilis test, 43% of those enrolled in Medicaid 5 or more
particularly UIP children, who are most at risk for uninsurance    months had an early test, and 29% of those with an early test
and BCs.                                                           had a repeat test. Non-Hispanic black women were
Primary Funding Source: No Funding                                 significantly less likely than non-Hispanic white women to
                                                                   have any (odds ratio [OR] = 0.95) or early (OR = 0.85) prenatal
                                                                   syphilis screening; Hispanic women were significantly more
                                                                   likely to have a repeat screen (OR = 1.15). Having a safety net
                                                                   provider as a usual source of prenatal care improved the
                                                                   probability of all three race/ethnicity groups receiving any,
                                                                   early or repeat syphilis screening tests, with the largest impact
                                                                   on non-Hispanic whites. Participation in Florida’s PCCM
                                                                   program improved access to care for only non-Hispanic black
                                                                   and Hispanic women.
                                                                   Conclusions: Among Medicaid pregnant women in Florida,
                                                                   rates of prenatal syphilis screening are generally low and
unexplained racial disparities in screening exist. Despite a      disparities in breast cancer in both the U.S. as well as the
higher rate of syphilis among non-Hispanic blacks, these          individual states. In fact, this study was undertaken to
women have significantly lower screening rates.                   develop estimates of the “cost” of particular health disparities
Implications for Policy, Delivery, or Practice: Clinic-based      for use in evaluating the cost-effectiveness of interventions
safety net providers can play an important role in improving      targeting specific breast cancer disparities – one of the
screening rates generally whereas PCCM may also reduce            priorities of Minnesota’s recently launched cancer control plan
disparities. Effective means of reaching and screening non-       - Cancer Plan Minnesota 2005-2010.
Hispanic black women early in their pregnancies are needed.       Primary Funding Source: Minnesota Department of Health
Primary Funding Source: CDC
                                                                  ●Self Efficacy in Obese African American Women
●Estimating Medical Expenditure Differences Attributable          Beginning a Weight Loss Program
to Breast Cancer Disparities in the U.S. and Minnesota            Lisaann Gittner, MSc, Mary Kay Anthony, RN, Ph.D., Pamala
Greg Gifford, Ph.D., Scott Leitz, MA                              Murphy, M.D., Shirley Moore, RN, Ph.D., Salwa Hassanein,
                                                                  MSN
Presented By: Greg Gifford, Ph.D., Research Scientist 3,
Department of Health, State of Minnesota, 85 East Seventh         Presented By: Lisaann Gittner, MSc, Director of Research,
Place, Suite 300, St. Paul, MN 55101; Tel: (651) 282-5643; Fax:   Division of Research, Kaiser Permanente, 19999 Rockside Rd,
(651) 282-5628; Email: greg.gifford@health.state.mn.us            Cleveland, OH 44146; Tel: (440) 735-6431; Fax: (440) 735-
Research Objective: Estimate calendar year (CY) 2002 health       6429; Email: lisaann.gittner@kp.org
care expenditure differences associated with racial/ethnic        Research Objective: Psychometrics of four Self Efficacy
breast cancer disparities in the U.S. and Minnesota.              Instruments are compared as predictors of weight loss in
Study Design: We developed these estimates in two major           African American (AA) women in a weight loss intervention.
steps. First, we developed estimates of annual aggregate and      Predictive ability and scale stability of the instruments was
per capita expenditures attributable to breast cancer by          determined at baseline and one-month post intervention.
racial/ethnic populations in the U.S. in CY 2002. Second, we      Study Design: This randomized intervention examined the
then used the U.S. per capita estimates, adjusted for regional    effect of a culturally adapted weight loss program in AA
differences to produce Minnesota per capita estimates, and        women. The 21-month program (6-months active intervention
estimates of breast cancer prevalence in Minnesota to             and 15-months maintenance) used behavior change strategies
calculate annual aggregate and per capita expenditure effects     modified to include spirituality, AA cultural materials,
attributable to racial/ethnic differences in breast cancer        interactive teaching, and sustained peer support. The NCEP
prevalence in MN in CY 2002.                                      Step 1 diet was modified to include calorie reductions,
Population Studied: Adult U.S. and Minnesota women in             increased fruit, vegetable, calcium intake, and incorporation of
2002 differentiated by race/ethncity categories - Hispanic (and   ethnic recipes. Exercise was culturally adapted to include line-
any race), Black-not Hispanic, Asian-not Hispanic, and Other-     dancing, brisk walking, and included time for socialization.
not Hispanic.                                                     The control group received usual care which includes personal
Principal Findings: Estimated per capita medical spending         dietician visits, nutrition classes, and web based modules.
attributable to breast cancer in the U.S. in CY 2002 was          Descriptive characteristics, self efficacy for exercise adherence,
significantly higher for the Hispanic and Black populations       exercise barriers, and eating control were measured at
compared to the Asian and Other Race/Not Hispanic                 baseline and 1 month after the active intervention.
(includes white) populations. It is noteworthy that medical       Psychometric instruments used were: McAuley Self Efficacy:
spending attributable to breast cancer for the Asian              Exercise Adherence (EA) (Scale 0–100), modified McAuley
population was found to be much lower than for the other          Self Efficacy: Exercise Barriers (EB) (Scale 0–100), Weight
three racial/ethnic populations. After substituting estimates     Efficacy Lifestyle (WEL) (Scale 0–9), Self Efficacy and
of adult female breast cancer prevalence and comparable           Consumption of Fruit and Vegetable (F/V) (Scale 1–5).
population sizes as of CY 2002 in Minnesota into these            Population Studied: Subjects were 190 AA women between
methods, we find a pattern of breast cancer attributable          the ages of 40-65 with BMI >30, and members of a
medical spending in Minnesota similar to the U.S. pattern.        Midwestern HMO.
Conclusions: These and other findings from this study: (1)        Principal Findings: Four self efficacy instruments will have
quantify the greater medical spending burden attributable to      construct validity, reliability, and scale stability assessed to
the treatment of breast cancer for Hispanic and Black women       determine the best scale to predict weight loss. At baseline
than for the Asian or Other/White populations in the U.S. as      there was no correlation of self efficacy with depression, BMI,
well as Minnesota; (2) suggest the possible roles of              age, weight, prior weight loss efforts; and all demonstrated
differences in breast cancer stage, patient complexity, and       good internal consistency (Cronbach’s alpha from 0.943-
rates of relevant preventive services utilization in              0.991). Baseline analysis shows high initial self efficacy for 3 of
understanding those differences; and (3) illustrate the           the 4 instruments [EA x = 76.9+/-25.5SD; Range=0-100; Skew=
potential generalizability of these methods for (a) application   -1.297; n=189; EB x = 62.8+/- 23.9SD; Range=3.85-100; Skew= -
to developing estimates of racial/ethnic disparities in medical   .412; n=186; WEL x = 6.12+/-1.65SD; Range = 1.2-8.8; Skew= -
care expenditures attributable to other diseases, and (b) other   .697; n=16]. For the F/V, self efficacy was x=1.9+/- 0.9SD;
states and regions.                                               Range= 4.8-1.0 Skew= 1.3; n=178. Once Time 1 data collection
Implications for Policy, Delivery, or Practice: The results of    is complete (Feb06) the sensitivity of the instruments will be
this study should be directly applicable to evaluating the cost   examined.
effectiveness of specific interventions targeting racial/ethnic
Conclusions: Interestingly, at baseline there was high               had used an antidepressant. In the multivariate logistic
beginning self efficacy for exercise and eating control exhibited    regression models, Blacks with any lifetime and twelve-month
by most subjects. The self efficacy instruments’ answer              DSM disorders had significantly increased odds of
distribution patterns were compressed towards one end of the         antidepressant use as did those with general medical
scale, high end compression observed for EA, EB, WEL and             disorders, but not their interaction terms.
low end for F/V. This may limit scale usefulness as a predictor      Conclusions: African Americans and Caribbean Blacks had
of successful weight loss. There was no indication that initial      similar estimates of past twelve-month antidepressant use,
BMI, age, depression, and prior weight loss efforts related to       and when combined into one group, few Blacks in this
exercise and/or food control self efficacy.                          nationally representative sample had used an antidepressant.
Implications for Policy, Delivery, or Practice: These                Small proportions of Blacks with MDD, dysthymia and anxiety
preliminary results do not indicate that a pattern of high self      disorders had used antidepressants in the past twelve
efficacy for healthy eating and exercise behaviors translate into    months. Although the proportion of Blacks using
healthy lifestyles. In fact, high self efficacy may interfere with   antidepressants that had no psychiatric disorders were smaller
dieter’s efforts by giving them a false sense that weight loss is    than those with psychiatric disorders; the absolute number of
easier than it actually is leading to a pattern of unsuccessful      Blacks with no psychiatric disorders who had used an
weight loss attempts. Setting realistic expectations about the       antidepressant in the past year was much larger than those
difficulty of weight loss and weight maintenance in those            with psychiatric disorders. Further, Blacks with general
subjects with high initial self efficacy may lead to better weight   medical disorders were likely to receive antidepressant
loss.                                                                pharmacotherapy independent of the presence of psychiatric
Primary Funding Source: Garfield Memorial Foundation                 disorders.
                                                                     Implications for Policy, Delivery, or Practice: These findings
●Predictors of Antidepressant Use in Nationally                      suggest widespread misuse of antidepressants among Black
Representative Sample of U.S. Community-Dwelling                     Americans, with significant unmet need coexisting with
Blacks (African Americans and Caribbean Blacks): Results             considerable unnecessary treatment.
from the National Survey of American Life                            Primary Funding Source: NIMH
Hector Gonzalez, Ph.D., Thomas W. Croghan, M.D., David R.
Williams, Ph.D., Brady T. West, MA, Willie Underwood III,            ●Admission Diagnosis and Sociodemographic Factors
M.D., James S. Jackson, Ph.D.                                        Influence Patient Outcomes Following Surgical Resection
                                                                     of Colorectal Cancer at Low Volume Hospitals in New
Presented By: Hector Gonzalez, Ph.D., Assistant Professor,           York City 2000
Gerontology and Family Medicine, Wayne State University, 87          Dennis Graham, MS, RN, NP
East Ferry Street, Knapp Bldg, Detroit, MI 48202; Tel: (313)
577-2297; Fax: (313) 875-0127; Email:                                Presented By: Dennis Graham, MS, RN, NP, Nurse
hmgonzalez@med.wayne.edu                                             Practitioner Clinical Program Director, Nursing, Memorial
Research Objective: To determine the prevalence of and               Sloan-Kettering Cancer Center, 1275 York Avenue, New York,
predictors of twelve-month antidepressant use among Black            NY 10021; Email: grahamd@mskcc.org
Americans in a nationally representative sample of Blacks            Research Objective: This study characterized the differences
(African Americans and Caribbean Blacks).                            in admission diagnosis and sociodemographic characteristics
Study Design: In-person household survey from 2002 to                of patients undergoing colorectal cancer surgery on at New
2003. The National Survey of American Life (NSAL), along             York City Hospitals in 2000 performed in low versus high
with the National Comorbidity Study-Replication and National         volume hospitals.
Latino Asian American Study, is part of the Collaborative            Study Design: The New York Statewide Planning and
Psychiatric Epidemiologic Surveys. Setting: The 48 contiguous        Research Cooperative System database was queried for
United States.                                                       patient admission diagnosis; demographic factors; an
Population Studied: African American and Caribbean Black             economic profile, determined by the average adjusted gross
household residents ages 18 years and older (n = 5,004) who          income by zip code; and the surgical volume of treating
responded to the NSAL survey.                                        hospital. The 60 hospitals included in the study were grouped
Principal Findings: African Americans and Caribbean Blacks           by colorectal cancer surgical volume into 4 groups: low,
did not differ significantly in past twelve month antidepressant     medium, high and highest. Distribution of factors within
use. In a combined sample, 4.5% of Blacks had used                   hospital groups was analyzed by an unadjusted Chi-square.
antidepressants. Few Blacks with lifetime and twelve-month           Population Studied: All patients in New York City hospitals in
MDD (10.8%, 13.4%, respectively), dysthymia (11.1%, 11.9%,           2000 with the principle diagnosis by ICD-9 code for colorectal
respectively) and anxiety (10.3%, 14.1%, respectively), any of       cancer and who had an ICD-9 code for surgical resection of
the aforementioned psychiatric disorders (9.7%, 13.5%,               the colorectal cancer during this admission were queried.
respectively) had used an antidepressant in the past twelve          Principal Findings: Lower-volume hospitals had the highest
months. Although a small proportion (1.4%) of those with             number of non-colorectal admissions, Medicaid, Medicare
none of the above psychiatric disorders used an                      and minority patients as well as the highest inpatient mortality
antidepressant in the past twelve months, about half of all          rate and older and poorer patient population. To the contrary,
antidepressant use in the entire sample occurred in those who        the higher-volume hospitals had lower inpatient mortality,
did not meet criteria for any twelve-month or lifetime               treated patients of younger age and higher AGI and had fewer
psychiatric disorders. A large proportion of Blacks with general     minority and Medicaid patients than the lower-volume
medical disorders such as diabetes and hypertension (7.1%)           hospitals.
Conclusions: Higher inpatient mortality rate in lower-volume        housed. For ages 12-19, obesity prevalence was 28% homeless
hospitals may, in part, be related to differences in the            vs. 22% housed. The mean BMI for obese patients was similar
sociodemographics and admission diagnosis from those                for homeless and housed patients (27.7 vs. 27.5; not
patient treated at the at the high-volume centers. These            significant). There were no statistically significant differences
differences should be considered when comparing outcomes            in obesity prevalence based on age, gender or race-ethnicity.
across institutions.                                                These data are consistent data for adult obesity in Hunts
Implications for Policy, Delivery, or Practice: This study          Point-Mott Haven (27%, NYC Department of Health). Overall,
revealed that 26 percent of patients who had a surgical             12% of the homeless patients were overweight (BMI 85th-94th
resection for colorectal cancer in 2000 in New York City were       percentile) and 3% underweight (BMI=<5th percentile). For
admitted with a non-colorectal diagnosis. Public policy needs       the housed patients, 19% were overweight and 3%
to better methods to support early detection of colorectal          underweight. Only 55% of homeless and 54% of housed
cancer for the elderly, poor and under-insured.                     patients had a BMI within normal range.
Primary Funding Source: No Funding                                  Conclusions: The health burden of obesity and associated
                                                                    risk appears to be evenly distributed in inner city minority
●Obesity Among Inner City Children, Homeless and                    populations of New York City. Prevalence in these high-risk,
Housed                                                              medically underserved populations exceeds national data for
Roy Grant, MA, Sandra Goldsmith, MS, RD, Alan Shapiro,              minority children and youth.
M.D., Marian Larkin, M.D., Irwin Redlener, M.D.                     Implications for Policy, Delivery, or Practice: These results
                                                                    emphasize the importance of integrating BMI calculation in
Presented By: Roy Grant, MA, Director of Research, Policy,          pediatric primary care for high-risk populations. Weight
The Children's Health Fund, 215 West 125th Street, New York,        management interventions, including nutrition counseling,
NY 10027; Tel: 212-535-9400; Fax: 212-535-7699; Email:              should be available as clinically indicated.
rgrant@chfund.org                                                   Primary Funding Source: HRSA, The Picower Foundation;
Research Objective: Data from the Centers for Disease               Bristol-Myers Squibb
Control and Prevention (2000) show that obesity (BMI =>95th
percentile) in the U.S. doubled from 1980-2000 among                ●Prevalence and Variation of Literacy Skills Among
children 6-11 years old (to 16%) and tripled for youth 12-19        Veterans
(also to 16%). Obese adolescents have an approximately 80%          Joan Griffin, Ph.D., Annamay Snyder, M.P.H., Melissa Partin,
chance of becoming obese adults, with increased risk of type 2      Ph.D., Sean Nugent, BA, Siamak Noorbaloochi, Michelle
diabetes, hypertension, dyslipidemia, nonalcoholic fatty liver      vanRyn, Ph.D., Joseph Grill, MS
disease, orthopedic disorders, low self-esteem and other
potentially preventable conditions. Racial-ethnic disparities are   Presented By: Joan Griffin, Ph.D., Research Investigator,
marked. National data show that for children 6-11, 12% of           Center for Chronic Disease Outcomes Research, Minneapolis
white children are obese compared to 17.6% African-American         VA Medical Center, One Veterans Drive, Minneapolis, MN
and 19.6% Hispanic (Mexican). For youth 12-19, these data are       55417; Tel: 612-467-4232; Fax: 612 725 2118; Email:
12.8% for white youth compared to 20.5% for African-                joan.griffin2@med.va.gov
American and 19.4% for Hispanic (Mexican). This study               Research Objective: Studies show that people with poor
compares the prevalence of obesity among inner city minority        literacy skills have significantly worse health, are less likely to
children and youth age 6-19 years who were housed to those          engage in preventative health measures, and have less
who were homeless.                                                  knowledge about their health. Studies also report that those
Study Design: The method was retrospective chart review of          with the poorest skills do not speak English as a first language
patients of the New York Program of The Children’s Health           or are elderly, poor, African American or Hispanic, and have
Fund, a service of Community Pediatrics, Children’s Hospital        completed fewer years of schooling. Although poor literacy
at Montefiore. Clinical data were coded and analyzed for            skills have been suggested as being one reason for disparities
potential differences based on homeless/housed status, age,         in health outcomes, little is known about the prevalence and
gender and race-ethnicity. Data analysis was done in SPSS           distribution of poor literacy skills among health care users.
using chi square, two-tailed Pearson correlation, and ANOVA         This study’s objective was to determine the prevalence of
calculations as appropriate.                                        literacy skills in a large patient population receiving care from
Population Studied: The sample was comprised of housed              the Veterans Health Administration and to assess variation in
pediatric patients of a federally qualified health center in a      skills by age, education, income, race, urban/rural and marital
medically underserved New York City community, Hunts                status.
Point-Mott Haven in the South Bronx (N=195), and homeless           Study Design: The study design was observational, based on
patients domiciled in the city’s family shelter system (N=200).     cross-sectional interviews with patients. The sample was
The patients were 45% African-American, 53% Hispanic                randomly drawn from a list of all patients with upcoming
(predominantly Puerto Rican or Mexican); 49% male; mean             appointments in primary care clinics at one of the four
age for 6-11 year olds was 8.5 years, mean age for 12-19 year       participating medical centers (Minneapolis, MN; Portland,
olds was 15.5 years.                                                OR; Durham, NC; and, West Los Angeles, CA).
Principal Findings: Obesity prevalence for homeless children        Population Studied: Randomized patients were sent
and youth compared to housed was higher but not to a                introductory letters and then recruited by phone. Patients
statistically significant degree. Overall obesity prevalence was    willing to participate in the study completed a face-to-face
30% for homeless patients vs. 25% for housed patients. For          survey with a trained interviewer. Interviewers collected
ages 6-11, obesity prevalence was 32% homeless vs. 27%              demographic information and conducted an assessment of
literacy (Short-Test of Functional Health Literacy in Adults, S-   replacement, with higher scores indicating more optimistic
TOFHLA). The S-TOFHLA tests reading comprehension and              expectations of surgical outcomes. In ordinary least squares
numeracy skills and classifies literacy by inadequate, marginal    regression controlling for survey site, whites scored 5.8 points
and adequate levels of skill.                                      higher than blacks for knee replacement expectations (95%
Principal Findings: Of the 1632 veterans assessed with the S-      confidence interval[CI] 2.3–9.4, p=0.001) and 8.1 points higher
TOFHLA, the majority (83%) had adequate literacy skills.           for hip replacement expectations (95% CI 3.3-13.0, p=0.001).
Slightly more than 16% had either inadequate (8%) or               Twenty-two percent of black patients, but only 13% of white
marginal skills (8%). In bivariate analyses, those with either     patients, scored below 30 on the knee expectations survey
inadequate and marginal skills were significantly more likely to   (p=0.02), and 25% of blacks, but only 10% of whites, scored
be over 75, have less than a high school education, an annual      below 30 on the hip expectations survey (p=0.01).
household income less than $20,000, and be widowed. In             Conclusions: Among veterans with moderate-to-severe
bivariate analyses the S-TOFHLA did not vary significantly by      osteoarthritis, whites have substantially higher expectations
race. In multivariate analyses, age, education, income, and        for the physical, functional, and psychological benefits of joint
African American race were significantly associated with           replacement surgery.
inadequate or marginal skills after adjusting for study site.      Implications for Policy, Delivery, or Practice: It is possible
Conclusions: A potentially clinically significant proportion of    that this difference in therapeutic expectations is an important
veterans assessed--1 in 6--had inadequate or marginal literacy     cause of the racial disparity in joint replacement procedure
skills. Consistent with other studies, veterans with poor          rates.
literacy were more likely to be poor, elderly, African American    Primary Funding Source: VA
and have fewer years of formal education.
Implications for Policy, Delivery, or Practice: Vulnerable         ●Smoke-Free Workplace Policies and Health Care
populations within the VA are more likely to have poor literacy    Coverage from 1998-2004: Evidence of Racial Disparities
skills than others. Developing effective strategies to improve     Jonathon Halbesleben, Ph.D.
outcomes for people with poor literacy may help reduce
disparities among veterans using VA services.                      Presented By: Jonathon Halbesleben, Ph.D., Research
Primary Funding Source: VA                                         Assistant Professor, Department of Health Management &
                                                                   Informatics, University of Missouri, Columbia, 324 Clark Hall,
●Racial Differences in Expectations of Joint Replacement           Columbia, MO 65211; Tel: (573) 884-1723; Fax: (573) 882-6158;
Surgery Outcomes among Veterans with Osteoarthritis                Email: halbeslebenj@health.missouri.edu
Peter Groeneveld, M.D., MS, C. Kent Kwoh, M.D., Cathleen J.        Research Objective: Tobacco use, including exposure to
Appelt, Ph.D., Jennifer Gutierrez, BA, Abigail L. Resnick, MS,     environmental tobacco smoke (second-hand smoke) in the
Said A. Ibrahim, M.D., M.P.H.                                      workplace, remains a critical public health concern. Previous
                                                                   research has found early evidence that there are racial
Presented By: Peter Groeneveld, M.D., MS, Assistant                disparities in smoke-free workplace policies. Similarly, health
Professor of Medicine, Ctr for Health Equity Research and          services researchers and policymakers have consistently
Promotion, Philadelphia VA Medical Center, 3900 Woodland           drawn attention to concerns with disparities in health care
Ave - 9East, Philadelphia, PA 19104-4155; Tel: (215) 898-2569;     access and insurance. The objective of this research is to
Fax: (215) 573-8778; Email: peter.groeneveld@va.gov                understand how disparities in smoke-free work policies
Research Objective: There are substantial differences in the       intersect with racial disparities in health care coverage.
rates that black and white veterans with osteoarthritis undergo    Study Design: Data from the 1998-2004 Behavioral Risk
joint replacement surgery. Expectations of the outcomes of         Factor Surveillance System (BRFSS) optional Tobacco Use
surgery such as functional status, pain, and quality of life are   Module were used to assess the extent to which individuals in
likely to influence the willingness of patients to undergo         different racial groups were covered by a smoke-free workplace
surgery, yet it is unknown if blacks and white differ in their     policy and health insurance. The data were analyzed across
beliefs about the likely benefits of elective joint replacement.   the seven year period in order to discern trends in racial
Study Design: We surveyed 365 (179 black, 186 white)               disparities in smoke-free workplace policies.
veterans with knee osteoarthritis and to 156 (72 black, 84         Population Studied: Civilian adults living in 34 of the United
white) veterans with hip osteoarthritis to assess their            States and the District of Columbia from 1998-2004.
expectations for the outcomes of joint replacement surgery.        Principal Findings: This research supports previous findings
Patients were eligible to enroll in the study if they indicated    regarding smoke-free workplace policies, finding that minority
moderate or severe joint pain and disability, as indicated by      workers are less likely to work in a workplace that has a
their Western Ontario and McMaster Universities                    smoke-free workplace policy. Specifically, white, non-Hispanic
Osteoarthritis Index. We used Mancuso's previously-validated       individuals were most likely to be covered by a smoke-free
Hospital for Special Surgery Joint Replacement Expectations        workplace policy, followed by black, non-Hispanics, followed
Survey (HSS-JRES) to assess patients’ expectations for pain        by workers of Hispanic descent. Additionally, this research
relief, functional improvement, and psychological well-being       found that while workers in all racial groups were more likely
after joint replacement surgery.                                   to be covered by a smoke-free policy over time, however, the
Population Studied: The study population comprised                 gap between racial groups remained fairly consistent over the
veterans ages 50-79 receiving primary care in the Pittsburgh       seven year period under study. Finally, Hispanics, who were
VA Health Care System or the Philadelphia VA Medical Center.       least likely to be working in an environment with a smoke-free
Principal Findings: Possible scores on the HSS-JRES range          workplace policy, were also the least likely to be covered by
from zero to 68 for knee replacement and zero to 72 for hip        health care insurance.
Conclusions: This research suggests that workers in minority       logistic regression will identify the independent effects of
groups may be doubly at risk, as they are more likely to work in   individual and county-level factors which predict rural
a workplace without a smoke-free work policy and are less          compared to urban avoidable hospitalizations.
likely to have health insurance coverage that might help to        Population Studied: Rural and urban patients hospitalized for
address the subsequent health problems as a result of their        avoidable conditions (including pneumonia, congestive heart
own or others' smoking behavior at work. The correlation           failure, asthma, cellulitis, perforated or bleeding ulcer,
between a lack of smoke-free workplace policies and health         pyelonephritis, diabetes with ketoacidosis or coma, ruptured
care insurance coverage may represent a more broad                 appendix, malignant hypertension, hypokalemia, immunizable
disregard for employee health in some companies; more              conditions, and gangrene) in short-term nonfederal hospitals
research is needed to explore why members of certain racial        in the United States.
groups are more likely to work at such companies.                  Principal Findings: In 2003, the rate of avoidable
Implications for Policy, Delivery, or Practice: This research      hospitalizations per 10,000 population was 198 in rural areas
gives further credence to concerns about the health care of        and 123 in urban areas. Patients with avoidable
members of racial minority groups in the United States. While      hospitalizations who lived in rural areas were older but had
this research suggests that there has been improvement over        similar numbers of diagnoses as their urban counterparts.
the past seven years with regard to voluntary smoke-free           They were likely to live in counties with less education, lower
workplace policies, there remain gaps between racial groups in     income, and less HMO penetration. The rate of hospital beds
coverage by such policies. Moreover, the incremental growth        was similar for rural and urban residents with avoidable
in smoke-free workplace policies suggests that we are unlikely     hospitalizations, but rural residents lived in counties with a
to meet the target of 100% smoke-free workplace policy             lower rate of active MD's. Multivariate analyses will be
coverage detailed in Healthy People 2010. Findings with            conducted to determine the independent contribution of each
regard to smoke-free workplace policies and health care            individual and county-level variable in predicting patients' rural
coverage suggest a need to explore mandated smoke-free             or urban status.
workplace policies and health insurance reform that may            Conclusions: Higher rural rates of avoidable hospitalizations
address racial disparities.                                        and lower rates of active physicians may signify problems with
Primary Funding Source: Centers for Disease Control and            access to appropriate primary care.
Prevention (CDC) and the Missouri Department of Health and         Implications for Policy, Delivery, or Practice: This project
Senior Services                                                    uses nationally representative data on rural and urban
                                                                   residents who had avoidable hospitalizations in order to
●Avoidable Hospitalizations: Disparities between Rural             determine factors which may contribute to disparities between
and Urban Patients                                                 these two groups. This information is crucial for informing
Margaret Hall, Ph.D., Maria F. Owings, Ph.D., Lola Jean            national health policymakers as they consider the need for
Kozak, Ph.D.                                                       further legislative changes designed to improve access to care
                                                                   in rural areas. Also, it is critical to have data on avoidable
Presented By: Margaret Hall, Ph.D., Health Statistician,           hospitalizations for both rural and urban areas for the period
Hospital Care Statistics Branch, National Center for Health        prior to the implementation of the Medicare Drug Benefit,
Statistics, 3311 Toledo Road, Room 3331, Hyattsville, MD           since inability to pay for outpatient drugs may contribute to
20782; Tel: (301) 458-4252; Fax: (301) 458-4032; Email:            avoidable hospitalizations.
mhall@cdc.gov                                                      Primary Funding Source: CDC
Research Objective: Avoidable hospitalizations have been
considered an indicator of access to timely and appropriate        ●Characteristics of Early Inter-Hospital Transfers after ED
primary health care. Rural residents face a unique                 Admission
combination of socioeconomic and geographic factors that           Daniel Handel, M.D., M.P.H., K. John McConnell, Ph.D.,
may create barriers to access. Rural/urban disparities in          Christopher J. Lindsell, Ph.D.
avoidable hospitalization rates may be a reflection of these
barriers. Most research literature that reports rural/urban        Presented By: Daniel Handel, M.D., M.P.H., resident,
disparities in rates of avoidable hospitalization is based on      Emergency Medicine, University of Cincinnati Medical Center,
small-scale or regional studies. The proposed research uses        7790 Fancycab Ct., Cincinnati, OH 45231; Tel: 513-226-5035;
recent, nationally representative data to explore this issue. It   Fax: 513-558-6434; Email: d.handel@fuse.net
will describe rural and urban patients hospitalized for            Research Objective: To determine characteristics predictive
avoidable conditions, and examine factors which might              of early inter-hospital transfer for identifying disparities in
explain rate disparities.                                          access to care after ED admission.
Study Design: Data are from the 2003 National Hospital             Study Design: This retrospective study used data from the
Discharge Survey (NHDS), a nationally representative survey        2002 Healthcare Cost and Utilization Project Nationwide
of short-stay, nonfederal hospitals in the US. County FIPS         Inpatient Sample, a nationally representative sample of 7.8
codes for patients are used to obtain information from the         million admissions and 1,000 hospitals. Admitted ED patients
Area Resource File on the socioeconomic characteristics and        who had not died or been transferred to a non-acute care
the health care capacity within the patient's county of            facility within the first day were included. Age, race, sex,
residence. The OMB definition of metropolitan status based         socioeconomic status and diagnoses requiring specialist
on the 2000 Census is used to define rural and urban.              services (e.g. cardiac) were used as patient level predictors of
Metropolitan counties are classified as urban and all others       early transfer, while hospital ownership, size, geographic
(micropolitan and noncore counties) are rural. Multivariate        location, and teaching status were explored at the hospital
level. Multivariable logistic regression was used to model           Population Studied: Admitted ED patients who had not died
predictors of transfer to an acute care facility within one day of   or been transferred to a non-acute care facility within the first
admission, with weighted data to provide national estimates.         day were included.
Population Studied: A nationally representative sample of 7.8        Principal Findings: The number of inter-hospital transfers
million admissions and 1,000 hospitals.                              within 24 hours of ED admission declined from 2001 to 2003,
Principal Findings: 36% of all hospitalizations met inclusion        from 1.55% (CI, 1.47-1.63) of patients in 2001 to 1.38% (CI 1.30-
criteria; 45% were female, 17% were black, and mean age was          1.46) in 2003. The probability of transfer for Medicaid patients
59.7 years. 1.4% (95% CI 1.3-1.5%) of included hospitalizations      decreased significantly, from 1.9% (CI 1.8-2.1) in 2001, to 1.3%
resulted in early transfer. Compared to 18-34 year olds,             (CI 1.2-1.4) in 2003. Declining trends in transfer were seen in
pediatric patients (OR 2.15, 1.87-2.46) and patients between 35      patients 45-64 years old and from larger hospitals. Patients
and 84 years-old were more likely transferred while patients         with diagnoses that required specialty services also showed
over 85 (OR 0.64, 0.57-0.71) were less likely transferred.           significant decreases in the probability of transfer, from 2.4%
Compared to whites, blacks (OR 0.73, 0.66-0.80) and                  (CI 2.5-2.2) in 2001, to 2.1% (CI 2.0-2.2) in 2003.
Hispanics (OR 0.80, 0.71-0.89) were less likely transferred.         Conclusions: Early inter-hospital transfers declined from 2001
Compared to privately insured patients, Medicare (OR 0.64,           to 2003, with the largest decrease in transfers occurring
0.60-0.69), Medicaid (OR 0.63, 0.57-0.70), and self-pay              among Medicaid patients. Furthermore, patients in need of
patients (OR 0.72, 0.65-0.80) were less likely transferred.          specialty services were less likely to be transferred once
Increasing median household income increased the likelihood          admitted.
of transfer. Transfer was more likely from small compared to         Implications for Policy, Delivery, or Practice: This may be
large hospitals (OR 3.26, 2.82-3.76), and rural compared to          supportive of observed trends towards increasing transfer of
urban teaching hospitals (OR 9.95, 7.93-12.49).                      patients direct from the ED and more appropriate early
Conclusions: Affluence increases the odds of early transfer,         disposition by emergency physicians.
suggesting possible disparities in access to care that could be      Primary Funding Source: No Funding
mitigated by direct transfer from the ED.
Implications for Policy, Delivery, or Practice: Assuming that        ●Beyond Formal Plans: Using Complex Adaptive Planning
early inter-hospital transfers are medically indicated and in the    and Evaluation (CAPE) Methods to Address Minnesota's
patients' best interests, results from this study are suggestive     Health Care Disparities
that patients who have traditionally not had adequate access         Margaret Hargreaves, M.P.P.
to care may be further disadvantaged due to a decrease
probability of early and appropriate transfers. EMTALA and           Presented By: Margaret Hargreaves, M.P.P, Senior Associate,
other federal regulations must be reviewed to see what health        Abt Associates Inc., Email: meg_hargreaves@abtassoc.com
policy implications can be modified to rectify this discrepancy.     Research Objective: In December 2003, The Minnesota
Primary Funding Source: No Funding                                   Department of Health (MDH) and the Minnesota Department
                                                                     of Human Services (DHS) hosted the Governor's Health Care
●Trends of Inter-Hospital Transfer within One Day of ED              Disparities Summit, at which the governor charged the leaders
Admission: 2001-2003                                                 of the state's health care industry to take action to deliver
Daniel Handel, M.D., M.P.H., K. John McConnell, Ph.D.,               culturally and linguistically appropriate health care (CLAS)
Christopher J. Lindsell, Ph.D.                                       services to all Minnesotans. The summit's participants
                                                                     formed the Minnesota Health Care Disparities Task Force in
Presented By: Daniel Handel, M.D., M.P.H., Resident                  March 2004, to address the issue.With staff support from
Physician, Emergency Medicine, University of Cincinnati              MDH and DHS, by early 2005 the task force had developed
Medical Center, 7790 Fancycab Ct., Cincinnati, OH 45231; Tel:        and started implementing statewide work plans in three areas:
513-226-5035; Email: d.handel@fuse.net                               provider collection and use of racial and ethnic health care
Research Objective: There is concern of an increasing rate of        data for quality improvement purposes; the improvement of
inter-hospital patient transfers. The purpose of this study was      interpreter services and other patient access issues, and
to determine regional, demographic, insurance, and hospital          provider cultural competence. At the same time, the task
factors in trends of early inter-hospital transfers over three       force became involved politically in the state's legislative
years.                                                               budget debate over whether to cut state public health
Study Design: This retrospective study used data from the            insurance programs. This research study examined the
2001-2003 Healthcare Cost and Utilization Project Nationwide         strategic use of a government task force to address statewide
Inpatient Sample, a nationally representative sample of 7.8          health care disparities, and specifically the task force's use of
million admissions and approximately 1,000 hospitals.                new Complex Adaptive Planning and Evaluation (CAPE)
Transfer rates were compared between years overall and               methods to facilitate system change, at both organizational
within the following categories: age, gender, race, type of          and political levels. The findings from this task force case
insurance, income, hospital size, type, and location, region,        study contribute to the health care disparities system change
and diagnoses requiring specialist services (i.e. neurosurgery,      and the health planning and evaluation literature.
cardiac catheterization). Proportions of patients transferred        Study Design: This task force was selected for a
within one day of admission were estimated, with 95%                 comprehensive case study because of its use of CAPE
confidence intervals (CI) adjusted for the complex sampling          methods, based on complex adaptive systems (CAS), agent-
scheme and subgroup analysis.                                        based modeling (ABM), conditions of self-organization, and
                                                                     emergent co-evolution theory. This study describes the task
                                                                     force's use of CAPE methods and subsequent actions to
intervene politically in the state's budget process and           increase in diabetes prevalence from 5.0% in 2001 to 6.4% in
organizationally, in the state's health care delivery system.     2003. In addition, Latinos have the highest prevalence in each
Research methods used included in-depth interviews with task      of the following age groups: 18-49, 50-64 and 65 and over.
force members and stakeholders, a review of relevant              Diabetes prevalence decreases with additional education from
documentation, and participant observation of task force          12.0% for those who did not complete eighth grade to 4.5%
methods, decisions, and processes.                                among college graduates. Prevalence also decreases with
Population Studied: The Minnesota Health Care Disparities         increasing income from 8.8% for those with household
Task Force, a statewide coalition of about 40 health care         incomes below 200% of the federal poverty level (FPL) to 5.1%
leaders from all health sectors.                                  for those with household incomes of at least 300% FPL. High
Principal Findings: Unlike traditional health task forces that    blood pressure and smoking have been linked to increased
publish reports, this task force took unprecedented action in     risk of serious complications such as cardiovascular disease,
both political and organizational arenas, impacting the state's   kidney disease, stroke, and amputation, and among those
health care system. The task force's efforts helped prevent       diagnosed with diabetes, these risk factors vary by
budget cuts in state public health insurance programs and         race/ethnicity. Overall, 60.9% of those with diabetes have
supported rapid diffusion of innovative CLAS practices            high blood pressure. More African Americans with diabetes
throughout the state's health care industry.                      have high blood pressure (74.0%) than any other group,
Conclusions: CAPE was shown to be an effective tool for           followed by Asians (67.2%). Overall, 15.0% of those with
helping government-based health task forces move beyond           diabetes are current smokers, but 39.8% of AI/ANs with
the publication of formal strategic plans to begin taking         diabetes are current smokers, followed by African Americans
action, both collectively and as individual task force members,   (20.7%). Disparities also exist in risk factors for developing
with system-wide results.                                         diabetes. Overweight and obesity are major diabetes risk
Implications for Policy, Delivery, or Practice: The CAPE          factors. Among all adults not diagnosed with diabetes, 54.0%
model complements traditional planning efforts by pairing         are either overweight or obese; however, significantly more
formal planning elements with opportunities for less scripted,    Latinos (64.8%), African Americans (63.7%) and AI/ANs
more adaptive action across multiple sectors and levels. The      (62.7%) are overweight or obese compared to whites (52.5%)
CAPE model provides a viable alterative to tightly controlled     or Asians (31.4%), placing these groups at increased risk for
government planning ? a middle ground of complex, adaptive        developing diabetes.
negotiation, collaboration, and co-evolution.                     Conclusions: African Americans, American Indian/Alaska
Primary Funding Source: The Minnesota Health Care                 Natives and Latinos have the highest diabetes prevalence, and
Disparities Task Force                                            are at greatest risk for developing diabetes and complications
                                                                  from diabetes. African Americans with diabetes have the
●Disparities in Diabetes Prevalence, Risk Factors and Risk        greatest prevalence of hypertension putting them at increased
of Comorbidities in California Adults                             risk for end-stage-renal-disease, hemodialysis and kidney
Theresa Hastert, M.P.P, Allison Diamant, M.D., MSHS, Susan        transplantation, as well as high rates of smoking putting them
Babey, Ph.D., E. Richard Brown, Ph.D.                             at increased risk for amputation.
                                                                  Implications for Policy, Delivery, or Practice: African
Presented By: Theresa Hastert, M.P.P, Research Associate,         Americans, American Indian/Alaska Natives and Latinos stand
UCLA Center for Health Policy Research, 10911 Weyburn Ave.,       to benefit the most from interventions targeted at diabetes
Suite 300, Los Angeles, CA 90024; Tel: (310) 794-2827; Email:     prevention and management.
thastert@ucla.edu                                                 Primary Funding Source: The California Endowment
Research Objective: To estimate disparities in diabetes
prevalence, comorbidities and risk factors among California       ●Are There Age Differences in Lipid and Glycemic Care
adults. To estimate disparities in diabetes prevalence,           Among Elderly Veterans with Diabetes?
diabetes risk, and comorbidity risk factors such as tobacco use   Drew Helmer, M.D., MS, Mangala Rajan, M.B.A., Usha
and hypertension among those diagnosed with diabetes.             Sambamoorthi, Ph.D., Leonard Pogach, M.D., M.B.A.
Study Design: We used data from the 2003 California Health
Interview Survey (CHIS). CHIS, a random-digit dial (RDD)          Presented By: Drew Helmer, M.D., MS, Physician,
telephone survey of households drawn from every county in         Researcher, VA-NJHCS, 385 Tremont Ave (129), East Orange,
California, completed interviews with over 42,000 households      NJ 07018; Email: helmer@njneuromed.org
in 2003. Bivariate analyses were used to examine diabetes         Research Objective: There is little research supporting age-
prevalence, comorbidities and risk factors among California       specific diabetes care standards for the elderly and guidelines
adults.                                                           recommend making diabetes care decisions on an individual
Population Studied: We analyzed responses from over               basis. The primary objective of the study is to compare
42,000 adults interviewed for CHIS 2003.                          patterns of diabetes care between the “young-old” (65-74
Principal Findings: In 2003 6.6% of California adults had         years) and “old-old” (>or= 75 years) using a nation-wide
been diagnosed with diabetes, up from 6.2% in 2001                database of veterans.
(p<0.10). American Indians/Alaska Natives (9.9%) and              Study Design: We tested for association between age and
African Americans (9.3%) had the highest prevalence.              diabetes care among the elderly in 2000 in this retrospective
Prevalence among African Americans was significantly higher       analysis of merged Veterans Health Administration (VHA) and
than among Latinos (7.5%), Asians (6.4%) or Whites (5.6%).        Medicare administrative data. Diabetes care measures
Although Asians have a comparatively low rate of diabetes,        consisted of hemoglobin A1c (HA1c) and low-density
they were the only racial/ethnic group to show a significant      lipoprotein (LDL) testing in VHA or Medicare data, and HA1c
greater than 9% (HA1c>9) and LDL greater than 130mg/dL              indexes as well as AMC-specific DES indexes were derived.
(LDL>130) in VHA data. We tested the null hypothesis of no          Persistence of race disparity based on DES index was
differences in care between the “young-old” (65-74 years) and       examined by age, gender, payer, comorbidities, severity of
“old-old” (>= 75 years) using logistic regression adjusting for     illness (SOI), admission status, LOS, and individual AMC
socioeconomic status (gender, race/ethnicity, marital status),      stent volume.
comorbid conditions (mental health disorder, comorbidity            Population Studied: 63,933 patients discharged from AMCs
score (DxCG)), access (priority group status, Medicare Part B       who had coronary stent implants between April 2002 and June
enrollment, Medicaid enrolled, distance to nearest VHA              2005. The data was from the University HealthSystem
facility), and dual-system use (VHA vs. non-VHA). We                Consortium Clinical Data Base which contains UB-92 billing
included cardiology care for LDL outcomes and cardiology and        abstract information.
endocrinology care for HA1c outcomes. We also examined              Principal Findings: There are statistically significant
interaction terms for age and specialty care variables.             differences proportions of patients having bare metal stent
Population Studied: Veterans with diabetes enrolled in the          (BMS) vs. DES implants by race. The DES index for white
VHA and Medicare fee-for-service in 1999 and 2000 who had           patients increased from 32.6% to 91.2% between April 2002
one or more face-to-face visit in the VHA in 2000 and were          and June 2005 while it increased from 20.8% to 88.7% among
alive at year’s end (n=194,772).                                    African Americans in the same period. While the gap
Principal Findings: Overall, 71% of young-old (n=120,000)           narrowed, a statistically significant difference remained. The
and 70% of old-old (n=74,772) had HA1c measured and 66%             disparity between white and African American DES indexes
of young-old and 62% of old-old had LDL measured. Among             persisted by age, gender, payer, comorbidities, severity of
those who had LDL tested (90,592) 18.6% of the young-old            illness (SOI), and admission status, the difference being about
and 18.3% of old-old had LDL>130. Among those who had               10% (and statistically signficant) across all comparisons. The
HA1c tested (112,168) 12.8% of young-old and 9.4% of old-old        smallest difference was between white and African American
had HA1c>9. The age differences in care were statistically          patients with the highest severity of illness score, 52.0%
significant for HA1c control and LDL testing. Cardiology care       among African American patients versus 55.0% among white
was associated with increased likelihood of LDL testing. The        patients. Institutional DES indexes for white and African
age-cardiologist interaction term was significant and positive      American patients were highly correlated, i.e., higher AMC
for LDL testing. The age-endocrinology interaction term was         DES indexes for white patients were associated with higher
significant and positive for HA1c>9.                                AMC DES indexes for African American patients. There was
Conclusions: Compared to the young-old, the old-old are             more variation in this relationship among AMCs with lower
significantly less likely to have LDL testing and poor glycemic     overall DES indexes. When AMCs were sorted by their African
control. However, the magnitude of the differences may not          American DES indexes in decreasing order, the percent of
be considered clinically significant. The association between       white patients receiving DES iimplants varied extensively, even
older age and better glycemic control is likely due to survival     though proportionally, white patients received more DES
bias or unadjusted socioeconomic or disease factors.                implants. There were differences in rate of adoption of DES
Cardiologist care is associated with better lipid management        technology across AMCs. Within three months, a third of the
and attenuates the age discrepancy in LDL testing.                  AMCs already had DES indexes greater than the overall
Implications for Policy, Delivery, or Practice: It appears that     average for the year. Another third did not cross that
clinicians are treating elderly patients with diabetes to similar   threshold until 11 or 12 months after introduction, and about
standards.                                                          20 percent did not reach that threshold in the first year after
Primary Funding Source: VA                                          availability of the new technology. Some of the disparity may
                                                                    be a consequence of availability of DES at AMCs with high
●Race Disparities in Delivery of Invasive Cardiology                proportions of African Americans and low percentages of DES
Procedures at Academic Health Centers                               implants.
Samuel Hohmann, Ph.D., MSHSM, Sandra Magnetti, Dr.P.H.              Conclusions: This bivariate analysis suggests a large,
                                                                    significant race disparity in providing DES implants. The
Presented By: Samuel Hohmann, Ph.D., MSHSM, Senior                  proportions are affected by proportions of African Americans
Research Analyst, Information Architecture, University              receiving care at an AMC as well as the number of patients
HealthSystem Consortium, 2001 Spring Road, Suite 700, Oak           receiving staents at the AMC. While disparities exist, the issue
Brook, IL 60523; Tel: (630) 942-1740; Email:                        may be explained by access/availability, and access barriers
hohmann@uhc.edu                                                     were not considered in this study. Differential adoption of the
Research Objective: The evidence is clear that minority             use of DES implants may result in unequal access to DES.
patients are less likely to receive invasive cardiac procedures.    Implications for Policy, Delivery, or Practice: Race
These disparitites remain after adjustments for demographics.       disparities can be identified and better understood through
Providing drug eluting stent (DES) implants, a rapidly adopted      access to hospital discharge data. Results like these can be
practice since 2003, has varied by race. The objective of this      used to examine outcomes of care and provide insights to
research project was to determine if there are race disparities     improve quality of care. The results can drive policy
in providing new invasive cardiology technologies to patients       development to address disparities and support the need for
at academic medical centers (AMCs).                                 resources to eliminate or reduce disparities. To do this we
Study Design: Retrospective cohort study. An index was              must determine what incentives/disincentives there are for
developed to represent the proportion of patients receiving         use of new technologies, create awareness of unequal access
coronary DES implants versus the number of patients                 to new technology, and develop strategies to ensure equal
receiving any type of coronary stent implant. Overall DES           access.
Primary Funding Source: No Funding                                  ●A National Sutdy of Asthma Prevalence and Trends in
                                                                    Urban and Rural Counties, 2000-20003
●Racial Differences in Hospital Very Low Birthweight                J. Elizabeth Jackson, MA, Mark Doescher, M.D., MSPH, Gary
Neonatal Mortality                                                  Hart, Ph.D.
Elizabeth Howell, M.D., M.P.P., Paul Hebert, Ph.D., Samprit
Chatterjee, Ph.D., Mark R. Chassin, M.D., M.P.H,. M.P.P.            Presented By: J. Elizabeth Jackson, MA, Research Assistant,
                                                                    Family Medicine, University of Washington, 4225 Roosevelt
Presented By: Elizabeth Howell, M.D., M.P.P., Assistant             Way NE, Suite 308, Seattle, WA 98105; Tel: (206) 685-1990;
Professor, Health Policy, Mount Sinai Medical Center, One           Fax: (206) 685-0610; Email: bjackson@u.washington.edu
Gustave L. Levy Place, New York, NY ; Tel: (212) 659 9567; Fax:     Research Objective: Asthma has long been considered a
(212) 423 2998; Email: elizabeth.howell@msnyuhealth.org             problem of urban populations, but it is not an insignificant
Research Objective: To rank New York City hospitals                 problem in rural areas. Unfortunately, detailed information on
delivering very low birthweight infants by risk-adjusted            asthma prevalence and recent trends in rural locations is
neonatal mortality. To assess differences in the racial             limited. The objective of this study is to estimate the
distribution of very low birthweight births among New York          prevalence of and recent trends in asthma among US adults
City hospitals. To estimate the extent to which any such            residing in metropolitan and non-metropolitan counties, as
differences are associated with the disparity in neonatal           well as to identify specific rural groups at particular risk.
mortality between Blacks and Whites in New York City.               Study Design: Analysis of data from the Behavioral Risk
Study Design: This was a population-based cohort study of           Factor Surveillance System (BRFSS) national sample for the
New York City very low birthweight infants. The primary             years 2000 (n=184,450), 2001 (n=212,510), 2002 (n=247,964),
outcome was risk-adjusted neonatal mortality. We ranked             and 2003 (n=264,684). The outcomes measured were asthma
New York City hospitals by risk-adjusted neonatal mortality         diagnosis and current asthma symptoms, from self-report.
and then assessed differences in the distribution of Black and      Population Studied: Non-institutionalized US adults, using
White births among these hospitals. We estimated the effect         repeated samples drawn from all states and the District of
of racial differences in distribution of hospital births on Black   Columbia.
very low birthweight neonatal mortality.                            Principal Findings: In 2003, adjusted prevalence of asthma
Population Studied: We used New York City Vital Statistics          diagnosis ever was 12.0% for metropolitan counties and
records on all live births and deaths of infants weighing 500 to    11.0% for non-metropolitan counties (p<.001). The adjusted
1499 grams born in 45 hospitals from January 1, 1996 through        prevalence of asthma symptoms given a prior diagnosis was
December 31, 2001 (N=11,828). We excluded infants who               65.8% in metropolitan counties vs. 66.5% in non-metropolitan
were delivered outside of a hospital setting or had missing         counties (p=.554). Symptoms were most prevalent among
birthweights.                                                       non-metropolitan African Americans, Native Americans, those
Principal Findings: Risk-adjusted neonatal mortality for very       over 65, and those with less than a high school diploma.
low birth weight infants in New York City hospitals ranged          Prevalence of asthma diagnosis has trended upwards for both
from 9.6 to 27.2 per 1000 births. Distribution of births among      metropolitan and non-metropolitan counties between 2000
hospitals ranked from lowest to highest risk-adjusted mortality     and 2003; prevalence of current symptoms has declined over
differed for Black and White mothers (p=0.006). White               the same period.
infants were more likely to be delivered in low mortality           Conclusions: The prevalence of asthma is increasing at a
hospitals: 49% of White versus 29% of Black very low                similar rate among residents of both metropolitan and non-
birthweight infants delivered in the lowest mortality tertile of    metropolitan counties. However, until 2003 the prevalence of
hospitals. If there were no racial difference in hospital choice    current asthma symptoms was higher for residents of non-
of birth we estimate that Black VLBW mortality would be             metropolitan counties and remains especially problematic for
reduced by 6.6 deaths per 1000 VLBW births per year. This           rural African Americans, Native Americans, seniors and those
number represents 34.2% of the racial disparity in VLBW             with low educational attainment.
neonatal mortality in New York City.                                Implications for Policy, Delivery, or Practice: Asthma is a
Conclusions: New York City hospitals vary widely in VLBW            national problem requiring increased attention in both rural
risk-adjusted neonatal mortality. Differences in distribution of    and urban areas. However, particular efforts are needed to
very low birthweight Black and White births in New York City        improve the use of asthma controller medications in several
hospitals partially explain the racial disparity in neonatal very   high-risk, rural groups.
low birthweight mortality.                                          Primary Funding Source: Federal Office Of Rural Health
Implications for Policy, Delivery, or Practice: Our data            Policy
suggests that some of the racial disparities in neonatal
mortality rates may be related to where Black and White VLBW
infants are born. Variations among hospitals in the use of
effective treatments for VLBW babies may be responsible for
some of the racial disparity in neonatal mortality. Detailed
research examining practice patterns for the management of
prematurity is needed.
Primary Funding Source: AHRQ, Commonwealth Fund;
National Center for Minority Health and Health Disparities
●Does Having More Information Explain Observed Racial              amount of information black and white patients receive, but in
Disparities in Kidney Transplant Preferences?                      how that information is processed.
Cara James                                                         Implications for Policy, Delivery, or Practice: Increasing the
                                                                   amount of information all patients receive can improve
Presented By: Cara James, Health Care Policy, Harvard              treatment decisions and reduce disparities. However, more
Unversity, 180 Longwood Ave., Boston, MA 02115; Tel: 617-          research is needed to understand how the information is
493-1701; Email: cjames@fas.harvard.edu                            processed, and which sources are most influential.
Research Objective: Research suggests that African-                Primary Funding Source: WKK
Americans with end-stage renal disease (ESRD) are less likely
to want a kidney transplant and are less certain about their       ●The Future of Health Care for People with Disabilities
decision. The primary objective of this study is to determine      Who Do Not Receive Social Insurance Benefits
whether differences in sources of information explain              Tricia Johnson, Ph.D., William G. Johnson, Ph.D.
observed racial disparities in kidney transplant decision
certainty. This study also examines whether increasing the         Presented By: Tricia Johnson, Ph.D., Assistant Professor,
number of sources from which a patient receives information        Health Systems Management, Rush University, 1700 West
increases decision certainty, and whether it raises expectations   Van Buren Street, TOB Suite 126B, Chicago, IL 60612; Tel:
for transplant outcomes.                                           (312) 942-7107; Fax: (312) 942-4957; Email:
Study Design: This was a cross-sectional study of                  tricia_j_johnson@rush.edu
hemodialysis patients in four geographic areas of the United       Research Objective: The increasing number of people with
States (Alabama, southern California, Michigan, and the mid-       disabilities under age 65 will profoundly affect public policy
Atlantic region of Maryland, Virginia, and the District of         towards disability by increasing the administrative burden on
Columbia). Telephone interviews were conducted on a                programs such as Social Security Disability Insurance (SSDI)
stratified random sample of black and white men and women          and Medicare. The aging of the baby boom population will
in these regions to obtain information about transplant            substantially increase the number of people with disabilities
preferences and information sources used in the decision           who do not qualify for disability benefits. Failure to address
process. Patients who began dialysis between May 1996 and          the health care needs for persons with disabilities prior to
June 1997 were interviewed at home close to 10 months after        retirement is likely to increase long terms costs to Medicare by
initiation of dialysis. Medical records were reviewed by renal-    increasing the severity of illness once these persons reach
network staff to collect information on comorbidities and          retirement age. This paper estimates the number of people
referrals for transplant evaluations. Logistic regressions were    with disabilities under age 65 who do not receive disability
run to examine the relationship between information and            benefits and projects the number without benefits through
decision certainty, as well as racial differences in the number    2025 to quantify the uncompensated health care burden of
of information sources used in the transplantation process.        disability. Simulations are conducted to understand how
All analyses were adjusted for demographic characteristics         changes in the underlying characteristics of the population will
including gender, age, education and income.                       affect these projections.
Population Studied: The population studied consisted of            Study Design: This study estimates health care expenditures
hemodialysis patients between the ages of 18 and 54.               for people with disabilities who do not receive benefits using a
Pediatric patients were not sampled. The initial sampling          two-part generalized linear model. A simulation model is
process yielded 1933 patients. After excluding 254 patients for    developed to explore the effects of different assumptions
reasons including death, relocation, and cognitive                 about the demographic characteristics and health conditions
impairments, 1679 patients were eligible for the study.            of the population on the prevalence of disability and health
Principal Findings: After controlling for demographic              care expenditures. The 1996 panel of the Survey of Income
characteristics, patients with more sources of information         and Program Participation is used to estimate the number of
were more likely to want a transplant (OR=1.16, p=.01), be         people with disabilities who do not receive benefits. The 1997
certain about their decision (OR=1.15, p<.01), and have better     to 2000 Cost and Use Files of the Medicare Current
expectations regarding quality and length of life following        Beneficiary Survey are used to estimate health care
transplant (OR=1.07, p=.03 and OR=1.11, p<.01 respectively)        expenditures for people with disabilities who do not receive
than patients with fewer sources. Black patients had fewer         benefits, simulating costs as if they received benefits.
sources than white patients (6.34 vs. 6.58, p=.02), and were       Population Studied: Persons with disabilities under the
less likely to say they received enough information (73% vs.       normal retirement age who do not receive Social Security
82%, p<.01). They also had lower expectations than whites for      Disability Insurance or Medicare.
a much better quality of life following transplant (OR=.74,        Principal Findings: Nearly 23% (2.8 million [M]) of people
p=.01), but were as likely as whites to believe they would live    with disabilities did not receive SSDI benefits and had no
much longer. Blacks are less certain in their transplant           recent work experience in 2000, and 67% (8.9M) of those with
decision than whites (OR=0.67, p=.01). However, this               disabilities did not receive SSDI benefits overall. These
difference decreases after adjusting for sources of information    estimates will increase to 3.7M and 11.4M respectively by 2025.
(OR=0.73, p=.04). The interaction term for race and sources        Health care expenditures for people with disabilities without
was also significant (p=.02).                                      recent work experience and no SSDI benefits were estimated
Conclusions: Observed racial disparities in transplant             at $20.7 billion (B) in 2000 and will increase to $29.3B in
decision certainty decreased after adjustment for the amount       2025.
of information patients had. However, disparities persist.         Conclusions: People with disabilities who do not receive SSDI
Findings also suggest that disparities exist not just in the       benefits represent the hidden population of Americans with
disabilities. The reported prevalence of disability is sharply      visits=1.26, 7-13=1.45, >13=2.17), dying in hospital (RR=2.09),
reduced if only Social Security statistics are used. Results        dying outside the major urban centre, long term care facility
emphasize that the problem of uninsured losses among                Family Physician (FP) visit (RR=1.28), PCP admission
Americans with disabilities who do not qualify for social           (RR=1.16), palliative radiation (RR=1.13), and older age. Shorter
insurance benefits will be one of the most important,               stays were associated with more recent year of death, regular
unsolved dilemmas for society created by the baby boom              office hour FP visit(s) (RR for 1-4 visits=0.75, >4=0.58), FP
generation.                                                         home visit(s) (RR=0.87) and after hours FP office visit(s)
Implications for Policy, Delivery, or Practice: These results       (RR=0.81). Length of stay varied by cancer type. Income was
provide a unique insight into the impact of the baby boom           not a significant predictor.
generation on the health care industry and the problems to be       Conclusions: End of life care may be improved by focusing on
faced by aging baby boomers who will not have access to             PCP, nursing home and oncology services as well as
SSDI and Medicare. Estimates suggest that the burden of             community and clinical factors.
dealing with both the health care and income needs of this          Implications for Policy, Delivery, or Practice: Most cancer
very large population will be borne by the families of persons      patients die in hospital, yet up to 90% prefer to die at home.
with disabilities or by the Medicaid system.                        Historic decisions in Canada contributed to this situation.
Primary Funding Source: Disability Research Institute at the        With no national comprehensive palliative care plan, end-of-
University of Illinois at Urbana-Champaign                          life care evolved in response to local leadership. Determining
                                                                    the factors associated with spending less time in hospital is
●10 Years of Research on Access to and Predictors of End            important in developing policies and programs that optimize
of Life Cancer Care for more than 20,000 Adults in Nova             the use of health care resources and help people spend their
Scotia, Canada                                                      last days at home. The methods and data we are developing
Grace Johnston, Ph.D., Meaghan O'Brien, MHSA, Jun Gao,              will help evaluate and develop models of care for the expected
MSc, Ron Dewar, MSc, Beverley Lawson, MSc, Frederick                increase in need as the baby boomer cohort ages.
Burge, M.D., MSc                                                    Primary Funding Source: Canadian Institutes for Health
                                                                    Research
Presented By: Grace Johnston, Ph.D., Associate Professor,
School of Health Services Administration, Dalhousie                 ●Public Hospitals, Quality and Disparities: Are We There
University, 5599 Fenwick Street, Halifax, Nova Scotia, B3H          Yet?
1R2; Tel: (902)494-1309; Fax: (902)494-6849; Email:                 Karen Jones, MS, Bruce Siegel, M.D., M.P.H., Marsha
Grace.Johnston@Dal.Ca                                               Regenstein, Ph.D., MCP, Anne Beal, M.D., M.P.H.
Research Objective: To determine the percentage of adult
cancer patients dying at home and receiving various forms of        Presented By: Karen Jones, MS, Senior Research Scientist,
end of life care in Nova Scotia (NS) Canada, and to identify        Health Policy, George Washington University School of Public
predictors of their health services provision.                      Health and Health Services, 2021 K St, NW, Ste 800,
Study Design: Administrative data for 24,000 NS adults who          Washington, DC 20006; Tel: 202-530-2349; Fax: 202-296-
died of cancer between 1994 and 2003 were obtained by               0025; Email: kcjones@gwu.edu
linking NS Cancer Centre, Halifax and Cape Breton Palliative        Research Objective: The Institute of Medicine report Unequal
Care Program (PCP), and 1996 and 2001 census information.           Treatment recommended the use of evidence-based
For deaths from 1992 to 1998, hospital and physician billing        guidelines to reduce health care disparities and promote
data have also been linked, and these data are now being            quality. With the Hospital Quality Alliance (HQA) there is
updated. Multivariate logistic regression analysis was used to      now regular public reporting of quality data. We sought to
identify the odds and predictors of dying out-of-hospital.          understand 1) whether hospitals treating many minority
Population Studied: 24,000 NS adults who died of cancer             patients are measuring quality by race and ethnicity; 2) the
between 1994 and 2003                                               challenges they face in such analysis; 3) whether they are
Principal Findings: Only 26.9% died at home. Those                  finding disparities in care; and 4) their views on the public
significantly more likely to die at home were female (Odds          reporting of quality by race and ethnicity.
ratio [OR]=1.3), over 85 years (2.4), residing in a nursing home    Study Design: We selected public hospitals around the nation
(12.3), registered in a PCP: Cape Breton (1.5), Halifax (1.1),      that serve many minority cardiac patients. The hospitals were
living in an upper income community (1.2), dying from breast        asked to provide HQA measures by race and ethnicity from
(1.2), colorectal (1.2) or prostate (1.1) cancer, and living >120   the first and second quarters of 2004. This was followed by
days after their cancer diagnosis (2.6). Those less likely to die   structured interviews with 32 clinical and administrative staff
at home were living outside the largest metropolitan area in        at these hospitals, examining their process of obtaining and
Cape Breton (0.7) or other non-Halifax county (0.7), received       reporting quality, race and ethnicity data, their use of such
palliative radiation (0.9), a medical oncology consultation in      data to improve quality, and their views on public reporting.
their last year (0.9) or lived in a non-official language           Population Studied: Our sample consisted of six
community (0.8). Time in hospital prior to death varies but is      geographically dispersed large public hospitals. Three of these
generally decreasing. From 1992-8 in Nova Scotia, the               hospitals had a majority of Hispanic patients, two had a
percentage of persons dying in hospital decreased from 80.2%        majority of black patients and the sixth had approximately
to 69.8%. In 1992, adults diagnosed with cancer spent an            equal percentages of Hispanic and black patients. Five
average of 30 days in hospital in the last six months of life. By   hospitals completed all phases of the study.
1996, this dropped to 22 days. Longer stays were associated         Principal Findings: Performance on quality measures did not
with medical specialist visits (Rate Ratio [RR] for 3-6             vary substantially by race or ethnicity within these hospitals.
Performance was high on select measures across hospitals,            Women used home health and hospice more often than men;
but at some institutions as few as five percent of all patients      and married men had the lowest rates of utilization for both
received recommended care such as heart failure discharge            services. Less than half of patients used any services prior to
instructions. No hospital had examined its HQA measures by           death. However, patients who used home health services
race or ethnicity prior to participation in this study. Obtaining    were twice as likely to use hospice services.
these data by race and ethnicity was difficult for some sites        Conclusions: There is a relative underutilization of
and variation in race and ethnicity classification and collection    community-based services among older cancer patients,
precluded inter-hospital comparisons. Hospital interviewees          especially prior to death, particularly among married men and
were divided over whether HQA data should be publicly                non-whites.
reported by race and ethnicity.                                      Implications for Policy, Delivery, or Practice: Future studies
Conclusions: There was no evidence of disparate care using           should address optimal use of home health and hospice in
the HQA measures, and hospitals are not routinely using              this population, and how the patterns observed in our study
quality data to detect any such patterns. This analysis is           would be influenced by how a dual-provider divides services
difficult for them given their frequent reliance on contracted       among two separate billing opportunities, either home health
data vendors. The use of quality improvement techniques to           or hospice.
improve care is essentially universal, but is not viewed as a        Primary Funding Source: NCI, UAB Center for Aging
strategy to address disparities. Both technical and perceptual
challenges will need to be overcome if quality data by race and      ●Effect of State Medicaid and Medigap Eligibility Policies
ethnicity is to be publicly reported.                                on the Enrollment and Health Outcomes of Disabled
Implications for Policy, Delivery, or Practice: While the            Beneficiaries
HQA measures may not be the most sensitive tools for                 Jill Klingner, MS, RN, Ira Moscovice, Ph.D., Robert Town,
detecting disparities within institutions, reducing disparities      Ph.D., Kathleen Call, Ph.D.
may entail improving overall quality at providers with large
minority populations. From this limited sample it seems that         Presented By: Jill Klingner, MS RN, Research Assistant,
hospital leaders and clinicians are not viewing disparities as a     Health Services Research Policy and Administration,
quality of care issue. Given increasing competing demands            University of Minnesota, 2221 University Ave SE Suite 112,
for quality reporting coupled with technical obstacles,              Minneapolis, MN 55414; Tel: 612-626-6251; Fax: 612-626-6270;
hospitals are not likely to begin routine analysis of quality data   Email: klin0089@umn.edu
by race and ethnicity in the absence of some external                Research Objective: To determine the effect of state
mandate.                                                             Medicaid eligibility and Medigap guarantee offer policies on
Primary Funding Source: CWF                                          the enrollment in Medicaid or Medigap supplemental policies
                                                                     for the disabled under-65 Medicare beneficiaries and the effect
●Patterns and Predictors of Home Health and Hospice                  of supplemental insurance on their self-reported health and
Utilization Among Older Adults with Cancer                           functional status.
Meredith Kilgore, Ph.D., Locher Julie, Ph.D., Micheal Morrisey,      Study Design: This study is a natural experiment using
Ph.D., Christine Ritchie, M.D.                                       changes in state Medicaid and Medigap offering and eligibility
                                                                     policies over the 5-year period of 1997-2001, to estimate the
Presented By: Meredith Kilgore, Ph.D., Assistant Professor,          effect of these policies on Medicaid and Medigap enrollment
Health Care Organization and Policy, UAB School of Public            and subsequent health outcomes in the disabled under-65
Health, 1665 University Blvd. RPHB 330, Birmingham, AL               Medicare beneficiaries. Multinomial logit with fixed state and
35294-0022; Tel: (205) 975-8840; Fax: (205) 934-3347; Email:         year effects and robust standard errors clustered by state was
mkilgore@uab.edu                                                     used to estimate the effect of the state policies on enrollment
Research Objective: The primary aim of this paper is to              in Medigap and Medicaid. Linear regression with instrumental
describe patterns of home health and hospice use among               variables and state and year fixed effects was used to estimate
older cancer patients. A secondary aim is to identify                the supplemental insurance policy effect on health and
predictors of home care and hospice utilization. Finally, this       functional status.
study examines the extent to which home health and hospice           Population Studied: The study subjects are disabled, under-
are substitutes or complementary services.                           65 Medicare beneficiaries in the Medicare Current Beneficiary
Study Design: Retrospective analysis using the SEER-                 Survey (MCBS) sample from 1997-2001.
Medicare Database, a linkage of the Surveillance,                    Principal Findings: Medicaid eligibility and Medigap
Epidemiology and End Results Program of the National                 guarantee-offer policies have significant effects on enrollment
Cancer Institute and Medicare Claims.                                in Medicaid and Medigap, though not entirely as expected.
Population Studied: Three analytic samples were. The first           Medicaid enrollment is decreased by the income limit or
consisted of all cases who had ever had a diagnosis of cancer        percent of Federal Poverty Level required for Medicaid
and were eligible for services in calendar year 1999                 eligibility. The number of prescription Medigap plans required
(n=120,072). The second and third were sub-samples of these          to be offered has a negative affect on Medicaid enrollment,
and consisted of cases with a new cancer diagnosis in 1999           suggesting that the Medigap prescription plans compete with
(n=46,373) and cases who died in 1999 (n=41,483).                    Medicaid for enrollment. State prescription drug programs
Principal Findings: For all persons with a diagnosis of cancer,      also decrease Medicaid enrollment. Medigap enrollment is
29.0% accessed home health services and 10.7% used                   affected by one of the Medicaid related state policies, the
hospice services. Home health utilization is higher among            Medicaid buy-in policy, suggesting that not all disabled with
non-whites, and hospice utilization is higher among whites.          Medicaid eligibility select Medicaid. An increase of number of
Medigap prescription plans required to be offered increases        Population Studied: Hospital discharge records obtained
Medigap enrollment, suggesting that the availability of            from the State Inpatient Data of the Health Care Utilization
prescription drug plans increases the selection of Medigap         Project (HCUP) for 1995-96 and 2001-02 from four states
supplemental coverage. The effect of supplemental                  (Arizona, Florida, New Jersey, Wisconsin) chosen to reflect a
insurances on disabled health outcomes is mixed. Subsets of        mix of racial and ethnic population groups in the U.S. and the
disabled Medicare beneficiaries appear to have improved or         range of urban and rural settings.
declined health outcomes related to different types of             Principal Findings: We examine two main policy approaches:
supplemental health insurance coverage. This is not                First, state policies that concentrate services, including
surprising given the diversity of the disabled population.         certificate-of-need programs that limit the number of hospitals
Medicare HMOs appear to have a positive effect on functional       providing particular services, and efforts to promote centers of
status change. Medigap coverage neared significance for            excellence. In general, concentration of services should tend
improving health status change. Prescription drug coverage         to increase the likelihood that patients use high-volume
improved heath status for those with only one condition.           providers, and to reduce the extent of racial/ethnic disparities.
Conclusions: Beneficiaries’ enrollment in supplemental             Second, information strategies designed to facilitate
health insurance policies appears to be strongly influenced by     consumer choice of high-quality hospitals. We theorize that
the availability of prescription drug coverage. Future             quality-informed decisionmaking will arise as information
estimations of the health effect of supplemental insurance         about the relationship between volume and quality has
coverage for this diverse sample will require a more complete      become more widely known. Due to their comparatively
understanding of their healthcare utilization needs.               lower levels of income and education, racial/ethnic minorities
Implications for Policy, Delivery, or Practice: States may be      are less likely to be able to engage in quality-informed
able to decrease their Medicaid enrollment by supporting           decisions, and so we hypothesize that this information by
prescription drug plans for the disabled or Medigap                itself may actually lead to increases in the extent of disparities
prescription drug guarantee-offer plan requirements for the        in the use of high-volume hospitals. We also examine state
disabled. Examination of the effect of the guarantee-offer of      policies concerning payments for charity care and Medicaid
prescription plans (H, I, J) on Medicaid enrollment of elderly     managed care contracting, which will affect hospital use for
beneficiaries is indicated. More studies of the health outcome     those racial/ethnic minorities who are uninsured or insured
effect of supplemental coverage for subsets of disabled            through Medicaid.
Medicare beneficiaries are indicated.                              Conclusions: We discuss how state policies might affect high-
Primary Funding Source: No Funding                                 volume hospital use, and how our empirical results contribute
                                                                   to understanding the interactions between consumer
●Impacts of State Policy on Racial/Ethnic Disparities in           decisionmaking and public policy as they relate to disparities
Use of High-Volume Hospitals                                       in hospital use.
Karl Kronebusch, Ph.D., Bradford Gray, Ph.D., Mark                 Implications for Policy, Delivery, or Practice: The results
Schlesinger, Ph.D., Tracey Thomas, M.P.H., Edward Miller,          will aid policymakers in the design of policies to address
Ph.D.                                                              racial/ethnic disparities in utilization of high quality providers.
                                                                   Primary Funding Source: RWJF
Presented By: Karl Kronebusch, Ph.D., Associate Professor,
School of Public Affairs, Baruch College / CUNY, 1 Bernard         ●Access to Medical and Dental Care: Does Travel Burden
Baruch Way / D-901, New York, NY 10010; Tel: 646-660-6809;         Differ by Race and Ethnicity?
Email: karl_kronebusch@baruch.cuny.edu                             Sarah Laditka, Ph.D., M.B.A., Janice C. Probst, Ph.D., Jong-Yi
Research Objective: Most studies of the impact of policy on        Wang, Ph.D., MS, Andy O. Johnson, M.P.H.
disparities focus on policies that are deliberately designed to
alter disparities. Disparities, however, may be equally affected   Presented By: Sarah Laditka, Ph.D., M.B.A., Associate
by policies designed for other ends, but which nonetheless         Professor, Health Services Policy and Management, University
have important and often unintended consequences on the            of South Carolina, 800 Sumter Street, Columbia, SC 29208;
magnitude of racial/ethnic disparities concerning medical          Tel: (803) 777-1496; Fax: (803) 777-1836; Email:
care. The objective of this study is to determine whether and      sladitka@gwm.sc.edu
how these policies affect racial/ethnic disparities in use of      Research Objective: Greater travel time is often associated
high quality medical providers, defined as use of high-volume      with reduced access to care. A better understanding of
hospitals for those medical services for which there is            distances and mode of travel for individuals seeking health
evidence of a positive relationship between volume and             care is particularly useful for vulnerable populations, such
outcomes.                                                          racial and ethnic minorities, who experience transportation
Study Design: The study analyzes hospital discharge data for       barriers. We examined travel to care patterns by race and
20 services for which there is a demonstrated volume-              ethnicity. We used a broad category of “medical/dental” as a
outcome relationship compared to 6 services with no such           measure of health related travel. This measure provides
demonstrated relationship. Multivariate models control for         actual travel distance, in both miles and minutes, from the
patient characteristics that affect utilization, including         persons actually making the trip.
demographic characteristics, distance to hospitals, and            Study Design: A cross-sectional analysis of the National
insurance status. The analysis utilizes difference-in-difference   Household Travel Survey (NHTS). The NHTS is a nationally
contrasts between these services over time and the policy          representative sample of households, recruited by telephone,
variation between and within these states.                         who kept travel logs for one week between March 2001 and
                                                                   May 2002. We subset the analysis to individuals making at
least one trip for medical/dental care. The outcome variables       services needed to make appropriate health decisions--has
were distance to care (miles) and travel time (minutes). The        been identified as a major health care problem. Based on one
independent variables were residence (rural/urban) and race         estimate, the cost of low health literacy ranged from $30
(white, Black, Hispanic, American Indian/Alaska Natives,            billion to $73 billion in 1998 US dollars. The adverse
AI/AN, Asian/Pacific Islanders, A/PI). Covariates included          consequences of low health literacy are particularly
sociodemographics (sex, age, education, occupation, income),        pronounced in the elderly population because of the
medical conditions limiting driving (yes/no), travel                prevalence of chronic illnesses that require the elderly to
characteristics (perceived road conditions, day of week,            follow an intensive and complex medical regime (medications,
personal vehicle/other), and ecological factors (population         daily monitoring, routine physician visits, tests, etc.).
density, job density). Descriptive and multivariate analyses        However, decisions of health care arrangements for the elderly
were conducted in SAS-callable SUDAAN to account for the            are often made jointly with family members or loved ones.
NHTS sampling design. Logistic regression also examined             Social support also has a positive impact on the health of the
the odds of spending more than 30 minutes traveling for care,       elderly. In this study, we examine whether social support
holding other factors constant.                                     interacts with health literacy in affecting the health status of
Population Studied: Respondents to the 2001 NHTS: 3,926             older adults. We expect that high social support may alleviate
trips, made by 2,433 households.                                    the adverse effect of low health literacy in the elderly
Principal Findings: A routine medical/dental care trip              population.
averaged 11.0 miles (median, 5.4 miles) and took 22.3 minutes       Study Design: A cross-sectional survey of Medicare enrollees
(median, 14.9 minutes). Mean travel to care distances were          aged 65 and over in 2003-04. Health literacy was assessed
nearly identical for whites and blacks, at 9.82 and 10.1 miles.     using the Test of Functional Health Literacy in Adults. Social
AI/AN traveled furthest for care (18.1 miles), followed by          support was measured by a composite score constructed
Hispanics (12.0 miles). A/PI reported the shortest travel           using the MOS scale. Health literacy was indicated by self-
distances, 6.7 miles. Whites (20.4 minutes) and A/PI (17.0)         reported general health, SF-12 physical health, and SF-12
reported the shortest travel times, with Hispanics (23.0),          mental health. The interactive effect was tested using three
Blacks (29.4), and AI/AN (38.3) spending more time in travel.       dummy variables in OLS regression: (1) high health literacy
Rural residence was associated with increased travel distance       and high social support (HHL-HSS), (2) high health literacy
and time. In the analyses of odds spending more than 30             and low social support (HHL-LSS), and (3) low health literacy
minutes traveling for care, Black and Hispanics were more           and high social support (LHL-HSS), with the low health
likely to have extended travel time than were whites. Rural         literacy and low social support (LHL-LSS) group as the
residence, perceived poor driving conditions, being a               reference. Control variables included gender, ethnicity, age,
passenger rather than driver, and night travel were associated      education, and marital status.
with longer travel times.                                           Population Studied: 489 elderly Medicare enrollees.
Conclusions: Based on an analysis of completed trips,               Principal Findings: The average level of health status was
minorities invest more time to travel for medical/dental care       highest in the HHL-HSS group and lowest in the LHL-LSS
than whites, especially those spending more than 30 minutes         group. Results of the multivariate analysis showed: (1) HHL-
traveling for care. AI/AN populations, in particular, are           HSS (p<.001) and HHL-LSS (p<.01) were positively associated
disadvantaged. Anticipated travel burden, not captured in this      with general health status, (2) HHL-HSS (p<.01) was
study, may serve as another barrier to accessing                    positively associated with physical health status, and (3) HHL-
medical/dental care among minorities.                               HSS (p<.001), HHL-LSS (p<.05), and LHL-HSS (p<.001) were
Implications for Policy, Delivery, or Practice: Providers           positively related to mental health status.
should be aware of transportation patterns when locating            Conclusions: In general, there was an interaction between
safety net sites. Reliance on public transportation can result      health literacy and social support in influencing the health
in greater travel burdens for minorities, even when distances       status of Medicare enrollees. High social support appeared to
are similar. Policies should explore additional transportation      alleviate the adverse effect of low health literacy on mental
strategies in rural areas. Further research is needed to            health. High social support appeared to enhance the positive
determine effects of distance on compliance with specific           effect of high health literacy on general and physical health.
therapeutic regimens including provider visits and                  Implications for Policy, Delivery, or Practice: Both social
prescription filling.                                               support and health literacy are important determinants for
Primary Funding Source: HRSA                                        health status in Medicare enrollees. It is necessary to
                                                                    consider the individual’s social conditions while assessing the
●The Interactive Effect of Health Literacy and Social               impact of health literacy on health status. Furthermore, it may
Support on Health Status among Medicare Enrollees                   be useful to design intervention programs on individual
Shoou-Yih Daniel Lee, Ph.D., Ahsan M. Arozullah, M.D.,              differences in health literacy as well as differences in social
M.P.H., Young I. Cho, Ph.D., Kathleen S. Crittenden, Ph.D.          support.
                                                                    Primary Funding Source: AHRQ
Presented By: Shoou-Yih Daniel Lee, Ph.D., Assistant
Professor, Health Policy and Administration, University of
North Carolina at Chapel Hill, 1101 McGavran-Greenberg Hall,
Chapel Hill, NC 27599-7411; Tel: (919) 966-7770; Fax: (919)
966-6961; Email: sylee@email.unc.edu
Research Objective: Low health literacy--i.e., lack of ability to
obtain, process, and understand basic health information and
●Evaluation of the Tailored Patient Navigator Programs            our understanding of the barriers to cancer care for minorities
for the Underserved Populations with Cancer                       and underserved populations and how to reduce disparities in
Chyongchiou J. Lin, Ph.D., Karen Schwaderer, RN, BSN, Keith       cancer care. Finances and transportation issues appear to be
Morgenlander, M.P.H., Edmund Ricci, Ph.D., Dwight E.              the cornerstone of our patient’s non-medical barriers and
Heron, M.D.                                                       future research should address these specific needs in this
                                                                  population.
Presented By: Chyongchiou J. Lin, Ph.D., Visiting Associate       Implications for Policy, Delivery, or Practice: Our evaluation
Professor, Radiation Oncology, University of Pittsburgh,          of the PNP assessed what services are most needed by cancer
Shadyside Place 110, 580 South Aiken Ave, Pittsburgh, PA          patients and how much time is required for each service. This
15232; Tel: 412-235-1060; Fax: 412-235-1313; Email:               provides a model for future replication and comparison.
cjlin@pitt.edu                                                    Primary Funding Source: NCI
Research Objective: Cancer is the second leading cause of
death in the United States and in Pennsylvania. Cancer            ●Reducing Global Health Disparities: Emerging Initiatives
mortality rates for African American men and women in             for Developing Drugs for Diseases of Poverty
Pennsylvania are higher than their white counterparts and         Helene Levens Lipton, Ph.D., Jacob Doll, Corey Langenbach,
racial disparities are increasing. The Patient Navigator          M.P.H., Janet M. Berreman, M.D., M.P.H.
Program (PNP) was designed to overcome barriers to cancer
treatment in minorities and underserved populations. The          Presented By: Helene Levens Lipton, Ph.D., Professor Health
purpose of this study is to describe and evaluate the PNP for     Policy and Pharmacy, Schools of Pharmacy and Medicine,
the Radiation Oncology Cooperative Outreach Group                 University of California, San Francisco, 3333 California Street,
(ROCOG), a 5-hospital Cancer Disparities Research                 Suite 420, Box 0613, San Francisco, CA 94118; Tel: (415)476-
Partnership (CDRP) in Pennsylvania.                               2964; Fax: (415)502-0792; Email: liptonh@pharmacy.ucsf.edu
Study Design: The PNP is an innovative case management            Research Objective: Few disparities illustrate more starkly the
and outreach program in which “navigators” help patients          health challenges faced by developing countries than the so-
overcome barriers to cancer treatment. The PNP is designed        called 90/10 gap: only 10 percent of the annual worldwide
to enhance timely entrance into the healthcare system for         health R&D expenditure of US105 billion dollars is allocated
patients with positive pathology for cancer, through              for diseases that cause 90 percent of the global disease
coordination of support services, transportation and              burden. Between 1975 and 1999, only 1 percent of nearly 1400
assistance with insurance and other financial issues. The         newly marketed chemical entities were for such diseases.
PNP, working with two community hospitals in Pennsylvania,        Because few data are available in the literature to evaluate
as well as community and faith-based organizations, is            innovation in drug development for diseases of poverty, we
targeting two populations, inner-city African Americans and       systematically review and evaluate emerging initiatives and
residents of non-urban and socio-economically deprived areas.     suggest future policy directions.
The types of barriers for participants, types of assistance       Study Design: We selected four case-studies after: 1,
provided, and the amount of time the patient navigators           performing a literature review on global health and
spend on each issue raised or needed by patients are              development; 2, reviewing initiatives cataloged in an online
recorded. Patient and physician satisfaction surveys regarding    database of public-private global health partnerships; and 3,
the PNP were conducted.                                           conducting a series of key-informant interviews. Selected case
Population Studied: To date, the study has enrolled 23            studies evaluated were advance purchase commitments; novel
African Americans and 77 non-African Americans from the           regulatory approval of a rotavirus vaccine; a pharmaceutical
target areas who have a confirmed diagnosis of cancer.            peace corps; and a nonprofit drug company. For each
Principal Findings: Among the 100 participants enrolled from      initiative we identify potential strengths, discuss challenges
January through November, 2005, 30.5% of enrollees within         faced, and examine policy implications.
the urban facility were African Americans versus 12.2% at the     Population Studied: Not applicable.
non-urban facility. Among those who requested services, the       Principal Findings: Successful initiatives foster collaboration
navigator spent an average of 64.7 minutes per patient in the     among key stakeholders and have the potential to further drug
urban facility and 76 minutes per patient in the non-urban        development for neglected diseases. Advance Purchase
facility. The majority of requests for services were related to   Agreements, APCs,--arrangements in which a sponsor, e.g.
finance (54.2%), transportation (42.4%) or referral to the        global health organization, developed country government,
American Cancer Society (ACS) Patient Services (35.6%) in the     and/or philanthropic organization, guarantees a price for a
urban facility. In the non-urban facility, the distribution of    new product needed in the developing world, e.g. a malaria
service requests was 41.5% for referrals to the ACS Patient       vaccine,--potentially benefit all involved parties; challenges,
Services, 24.4% transportation issues, and 22% financial          however, include determining affordable pricing, setting the
issues. Health insurance approval, out-of-pocket payment,         purchase commitment size, enforcing the agreement, and
and transportation were listed as the primary barriers to         anticipating global health priorities. The novel regulatory
treatment in both facilities. All patients participating in the   approval of Rotarix vaccine in Mexico without prior FDA
PNP reported that they would use the service again and would      approval brought a needed vaccine rapidly to market in a
recommend the program to others.                                  developing country, but raises ethical concerns and may be an
Conclusions: While the PNP is a new initiative, there has         inefficient path to international regulatory approval. Pfizer’s
been a steady patient demand for PNP services in both             Global Health Fellows program enables volunteer
hospitals. All patients participating in the Program were         pharmaceutical professionals to provide expertise to
satisfied with the services. Our findings may further improve     developing countries and strengthens drug industry ties with
nonprofit groups and governmental organizations. Cross-            tuberculosis was also dichotomized (a lot to some knowledge
cultural differences in expectations and abilities, however,       versus a little to no knowledge).
have posed challenges and long-term effects are unclear. A         Population Studied: A nationally representative sample of
nonprofit drug company, OneWorld Health, identifies                adults who participated in the National Health Interview
previously developed promising drug candidates for neglected       Survey. The number observed varied by survey year and
diseases. Paromomycin is being developed as a new                  region from a high of 33, 000 to a low of 20,000.
treatment for visceral leishmaniasis, and will be submitted for    Principal Findings: Knowledge about tuberculosis and race of
regulatory approval in India. Despite early success in forming     the respondent were strongly associated with perceived risk of
collaborations with biotechnology, academe and philanthropic       tuberculosis in every region. Respondents with a lot or some
organizations, engaging the for-profit pharmaceutical industry     knowledge of the disease were 1.6 to 2.8 times more likely to
remains a challenge.                                               perceive a high to medium risk as compared to respondents
Conclusions: These innovations show promising early                with a little or no knowledge. Black respondents were at least
achievements in drug development for diseases of poverty.          twice as likely to perceive a high to moderate risk compared to
However, serious challenges confront each model.                   other races in all regions except the South (AOR=1.56). In
Additionally, they do not address other important barriers to      addition, knowing someone with tuberculosis was strongly
redressing global health disparities.                              associated with perceived risk in all regions except the South.
Implications for Policy, Delivery, or Practice:                    However, knowing there is a cure for tuberculosis was related
Considerations of drug distribution systems, health care           to perceived risk in the South but not anywhere else in the
infrastructure, ease of drug delivery and administration, and      nation. Nationally, a respondents knowledge about
affordable pricing need to be addressed. Evaluations of these      tuberculosis, knowing someone with tuberculosis, and being
emerging programs’ impact, e.g. cost-effectiveness analyses        Black were most strongly associated with perceived high to
and outcomes assessment, are needed to facilitate evidence-        medium risk of getting the disease.
based policy decisions designed to reduce global disparities in    Conclusions: Racial and socioeconomic disparities in
access to essential drugs.                                         knowledge and perceive risk of tuberculosis remain higher for
Primary Funding Source: University of California at San            African Americans than any other racial group. Knowledge of
Francisco                                                          tuberculosis is closely associated with perceived risk of the
                                                                   disease providing information form which to guide prevention
●Tuberculosis: The Shameful Truth About Health                     and treatment initiatives.
Disparities                                                        Implications for Policy, Delivery, or Practice: This study has
Ana Lopez - De fede, Ph.D., Karen Kirtland, Ph.D., Muriel          clinical and public health related implications. Although
Harris, Phs                                                        tuberculosis has been declining in the United States, it is
                                                                   increasing among certain high risk groups with regional
Presented By: Ana Lopez - De fede, Ph.D., Research Associate       variations reflecting demographic trends. Tuberculosis is
Professor, Institute for Families in Society, Universty of South   treatable and curable. Knowledge of tuberculosis is a key
Carolina, 1600 Hampton Street - 5th Floor, Columbia, SC            factor in the fight to eliminate health disparities among
20208; Tel: 803-777-5294; Fax: 803-777-1120; Email:                African Americans.
adefede@sc.edu                                                     Primary Funding Source: CDC
Research Objective: Tuberculosis continues to be a major
health threat in the United States with minimal efforts placed     ●Health Insurance Coverage Along the US-Mexico Border
on public education to increase knowledge about the disease        Jill A. Marsteller, Ph.D. M.P.P, Juan Albertorio, M.D.
among native-born US residents. The objective of this study
is to examine the role of race and place of residence residency    Presented By: Jill A. Marsteller, Ph.D. M.P.P, Contractor,
on knowledge and perceived risk of tuberculosis                    DHCS/HCSB, National Center for Health Statistics, 3311
Study Design: The study uses data from the 1994, 2002, and         Toledo Rd. Suite 3312, Hyattsville, MD 20782; Tel:
2004 National Health Interview Survey, the only nationwide         301.458.4098; Fax: 301.458.4032; Email: JMarsteller@cdc.gov
survey documenting knowledge about TB and its risk factors.        Research Objective: To address health disparities issues
Socio-demographic and geographic variables (age, sex,              along the US-Mexico Border, in 2000 the US-Mexico Border
education, annual household income, employment status, and         Health Commission (USMBHC) established a 10-year disease
census region) were taken from the sample adult                    prevention and health promotion program called Healthy
questionnaire. A measure of perceived risk and knowledge of        Border 2010. This program, like the Healthy People 2010 US
tuberculosis was obtained from NHIS. All analysis were             program, sets goals for improvement and monitors progress.
stratified by race (African American ,White and Hispanic)and       One selected health indicator is health insurance coverage
census regions (South, Northeast, West and Midwest). For           among Border populations. Estimates of health insurance
each census region, multiple logistic regression analysis was      coverage for regions smaller than state level are quite limited,
used to calculate the adjusted odds ratio (AOR, with 95%           and estimates based on ethnicity for Border region residents
confidence intervals) of the association between perceived risk    have been previously unavailable. This presentation will
of tuberculosis and knowledge of TB after adjusting for socio-     describe the USMBHC and the National Center for Health
demographic characteristics to include race and variables that     Statistics‘ Healthy Border 2010 Program, estimate the current
differed by race. For all models, perceived risk of tuberculosis   lack of health insurance coverage along the Border, compare
was treated as a dichotomized variable (high to medium risk        coverage rates among Hispanics and Non-Hispanics and
versus low to no risk), and the independent variable that          identify covariates of health insurance coverage on the Border.
measures how much knowledge the respondent had of
Study Design: The authors merged four years of the National        race/ethnicity or socio-economic status. We instead measure
Health Interview Survey (NHIS), 2000-2003, and calculated          disparities between high minority neighborhoods and white
lack of health insurance coverage for the population in US         neighborhoods. Advantages of the MEPS surveys are that
counties along the US-Mexico Border. Estimates were                they are nationally representative and they have detailed
stratified by ethnicity and geographic area (the US, the four      information on individual and family race/ethnicity and socio-
states on the US-Mexico Border (AZ, TX, NM, CA), and the           economic characteristics. Two measures are used: percent of
Border counties within these states). A logistic regression will   women age 40 and older who had a mammogram in 2003,
test the influence of demographic and health related variables     from MEPS’s medical records; and percent of MEPS
on insurance status among residents of the Border region,          respondents who said they had a mammogram in the past
with particular attention to differences between Hispanics and     two years. We also use multivariate regression analysis to
Non-Hispanics.                                                     isolate the individual effects of race and other characteristics
Population Studied: The civilian noninstitutionalized              on variations in mammography screening.
population residing in sampled primary statistical units along     Population Studied: Women age 40 or older.
the US-Mexico Border.                                              Principal Findings: Overall, 38% of women age 40 or older
Principal Findings: Within counties on the US-Mexico               living in predominantly white neighborhoods had a
Border, 22.9% were uninsured and 44.1% of these had lacked         mammogram in 2003; only 33% of enrollees living in black
coverage for a year or more. In the Border counties, 38.2% of      neighborhoods had screening mammograms in 2003, while
Hispanics were uninsured compared to 12% of Non-                   the statistic for Hispanics and Asians were 29% and 31%.
Hispanics. At all geographic levels examined, uninsurance was      Using LNEP, we will also examine trends since 1999. MEPS
considerably higher for Hispanics. The difference between          medical records show that 28% of all white women age 40 or
Hispanics and Non-Hispanics was larger in the Border states        older in US had a screening mammogram in 1998, while only
than nationally and larger still in the Border counties than in    18% of black women had one. The percentages for Hispanic
the Border states.                                                 women and Asian women were 15% and 21%. The disparities
Conclusions: In comparison with US estimates, disparity in         persisted in 2003: 30% of white women age 40 or older had a
insurance coverage along the Border is strongly evidenced.         mammogram, while only 19% of black women, 17% of
Specifically, the Hispanic population at the Border are            Hispanic women, and 16% of Asian women had one. When
substantially more likely than Hispanics and Non-Hispanics in      asked if they had a mammogram in the past two years, 70% of
any other US region examined to lack health insurance              white women answered “yes” in 1998, and 69% of black
coverage at the time of the interview.                             women answered the same. The percentages for Hispanic
Implications for Policy, Delivery, or Practice: Addressing         women and Asian women were 61% and 48%. The numbers
access to health insurance for the population residing along       in 2003 were 72% for white women, 72% for black women,
the US-Mexico Border may be particularly difficult because, as     62% for Hispanic women, and 56% for Asian women.
we know from the literature, the Border populations is mobile,     Regression results show that racial and ethnic disparities
multicultural and faces a challenging socio-economic reality.      remain even after controlling for other factors.
Coverage goals may best be addressed through international         Conclusions: Estimates of disparities in mammography
cooperative activities such as the USMBHC.                         screening between white and black women depend on how
Primary Funding Source: No Funding                                 the question is asked. Our finding persists over time, and is
                                                                   present in two different datasets. However, there are clear
●Racial and Ethnic Disparities in Mammography                      disparities in mammography screening between white and
Screening Over the Period 1996 - 2003                              Hispanic and Asian women.
Cristian Meghea, Ph.D., Mythreyi Bhargavan, Ph.D., Jonathan        Implications for Policy, Delivery, or Practice: The results of
H. Sunshine, Ph.D., Cristian Meghea, Ph.D.                         this paper will help policy makers understand the extent of
                                                                   disparities in mammography screening between various
Presented By: Cristian Meghea, Ph.D., Senior Researcher,           demographic groups, and design the appropriate policies to
Research, American College of Radiology, 1891 Preston White        ensure equitable utilization.
Drive, Reston, VA 20191; Tel: 703-648-8983; Email:                 Primary Funding Source: No Funding
cmeghea@acr.org
Research Objective: Breast cancer mortality decreased by           ●The Association of Depression with Missed Days of
24% between 1990 and 2000, but there are persistent                Employment Among persons with Chronic Conditions. Is
variations in mortality rates by race and ethnicity. Early         there a Race/Ethnicity Effect?
detection by mammography screening had a sizeable                  David Mosen, Ph.D., M.P.H., Rachel Gold, Ph.D., M.P.H.,
contribution to the reduction in breast cancer mortality.          Renee N Saris-Baglama, Ph.D., Chris Jentz, MHSA, Winston
The objective of this paper is to measure racial and ethnic        Wong, M.D., MS
disparities in mammography screening in 2003 and trends
over time since 1996. The paper will measure disparities           Presented By: David Mosen, Ph.D., M.P.H., Senior Program
between white and black women and expand the analysis to           Evaluation Consultant, , KP Center for Health Research, 3800
Hispanic and Asian women as well.                                  N. Interstate Avenue, Portland, OR 97227-1110; Tel: 503-335-
Study Design: We use claims data for five years, 1999-2003,        6637; Fax: 503-335-2424; Email: david.m.mosen@kpchr.org
on all medical services provided under a large national            Research Objective: Previous research has found that
employer plan (LNEP); and the Medical Expenditure Panel            depression may be associated with a greater likelihood of
Surveys (MEPS) for 1996-2003. LNEP covers millions of              missed employment among persons with chronic conditions.
individuals but has no information on an individual’s              However, less is known about whether the association
between depression status and missed days of employment            Study Design: Data come from the 2005 National Healthcare
varies by race/ethnicity. The objective of this study is to        Disparities Report, which compiled national estimates on 200
twofold: 1) To determine the impact of depression status on        measures of health care quality from 3 dozen data sources.
missed days of employment among a population with chronic          Measures cover the quality dimensions of effectiveness,
conditions and 2) To determine whether this relationship           safety, timeliness, and patient centeredness. Analyses focus
varies between racial/ethnic groups.                               on 46 core measures selected by consensus from the full
Study Design: We examined survey data for 1,202 employed           measure set and comparisons of whites, blacks, and
adults with chronic conditions enrolled in a large staff model     Hispanics. For each core measure, disparities are measured
HMO. Members with on or more of five chronic conditions            in both absolute and relative terms. Disparities for a baseline
(asthma, diabetes, chronic pain, heart failure, coronary artery    year and for the most current year of data are compared to
disease) were identified from 2003 patient data, using HEDIS       assess trends in disparities.
and registry inclusion criteria. In Fall 2004, the same patients   Population Studied: Most measures cover the civilian, non-
were surveyed regarding health status using the SF-8 Health        institutionalized U.S. population. Some measures are
Survey, which yields physical (PCS) and mental (MCS)               appropriate for specific groups defined by age and gender.
component scores. Based on previously published research,          Nursing home measures are specific to residents of these
patients with an MCS < 42 are likely to have depression.           facilities.
Additionally, respondents reported any missed days of              Principal Findings: Disparities are pervasive: Blacks receive
employment in the past 28 days due to their health                 poorer quality of care than whites for over 40% of core
conditions. Logistic Regression was used to examine the            measures; Hispanics receive poorer care for over half of core
independent association of depression status on missed days        measures.
of employment, adjusting for age, gender, educational              Some disparities are diminishing: Over time, more black-white
attainment, and geographic location. Three logistic models         differences in care are getting smaller than are getting larger.
were constructed: 1) overall and 2) stratified by white and non-   However, for Hispanics, the reverse is true; more Hispanic-
white ethnicities.                                                 white differences are getting larger than are getting smaller.
Population Studied: Employed adults with 1 or more chronic         Conclusions: Tracking disparities over time is important to
conditions (mean age = 50.4). About 30% of the population          gauge our progress towards the goal of eliminating health care
was non-white.                                                     disparities and to identify groups and areas in greatest need of
Principal Findings: Depression status was independently            intervention.
associated (OR=2.12, 95% CI = 1.43-4.89) with an increased         Implications for Policy, Delivery, or Practice: Communities
likelihood of missed days of employment. This association          should measure their disparities in health care, track changes
was strongest in the non-white population (OR=2.64, 95% CI         in disparities over time, and compare their local experience
= 1.49-3.00). There was no significant association between         with national benchmarks presented in the annual National
depression status and missed days of employment in the             Healthcare Disparities Reports.
white population (OR=1.25, 95% CI = 0.72-2.15).                    Primary Funding Source: No Funding
Conclusions: We found that likely depression status was
independently associated with increased likelihood of missed       ●Obesity in Native American Elders
days of employment. Our research suggests that depression          Kyle Muus, Ph.D., Leander McDonald, Ph.D., Richard Ludtke,
may impact employment-based outcomes more acutely in               Ph.D., Alan Allery, Ph.D.
minority populations than white populations.
Implications for Policy, Delivery, or Practice: Because            Presented By: Kyle Muus, Ph.D., Research Associate, Center
depression is associated with an increased likelihood of           for Rural Health, University of North Dakota, PO Box 9037,
missed days of employment in minority populations in               Grand Forks, ND 58202; Tel: (701) 777-3848; Fax: (701) 777-
particular, further efforts are needed to better recognize and     6779; Email: klmuus@medicine.nodak.edu
treat depression in this population. Future research should        Research Objective: Research questions include:
evaluate the impact of such interventions on employment-           What is the prevalence of obesity in Native American elders?
based outcomes in minority populations.                            By gender and age cohorts? How do Native obesity rates
Primary Funding Source: KPCMI                                      compare to all U.S. elders? What’s that relationship between
                                                                   rurality and prevalence of obesity among Native American
●Trends in Racial and Ethnic Disparities in Health Care            elders? Does the presence of obesity increase the likelihood of
Quality                                                            having co-morbid health conditions among Native American
Ernest Moy, M.D., M.P.H., Elizabeth Dayton, MA, Karen Ho,          elders?
M.H.S.                                                             Study Design: Descriptive, cross-sectional survey
                                                                   Population Studied: 8,119 Native Americans aged 55 years
Presented By: Ernest Moy, M.D., M.P.H., Senior Service             and older.
Fellow, Center for Quality Improvement and Patient Safety,         Principal Findings: Of the respondents, 43.2% were classified
Agency for Healthcare Research and Quality, 540 Gaither            as obese, about one-third (33.0%) were overweight, and
Road, Rockville, MD 20850; Tel: 301-427-1329; Fax: 301-427-        approximately one-quarter (23.7%) were classified as having
1341; Email: emoy@ahrq.gov                                         normal or low weight. By age, the youngest elders (ages 55-64)
Research Objective: To examine changes over time in racial         were most likely to be obese. By gender, females were more
and ethnic disparities across a broad range of health care         likely to be obese. By geography, urban-based Native elders
quality measures.                                                  were more likely to be obese, compared to rural elders.
                                                                   Obesity prevalence rates among Native elders were
substantially higher than for U.S. all races, particularly for         Conclusions: Antidiscrimination policies will improve access
younger cohorts. Obese Native American elders were more                to care defined in terms of numbers of clients receiving
likely to have certain chronic health conditions, compared to          services. Improving the quality of services received by
their non-obese counterparts. Obesity significantly increased          minorities will take more than antidiscrimination legislation
the likelihood of having hypertension, arthritis, diabetes,            and judicial decisions.
asthma, and congestive heart failure, after controlling for age,       Implications for Policy, Delivery, or Practice:
gender, education, and income.                                         Antidiscrimination legislation would be more effective if it
Conclusions: Native Americans elders, especially females,              included ways of monitoring and ensuring
were substantially more likely to be obese than their U.S. elder       improvements/equity in quality of serviced rendered minority
counterparts. Younger Natives were much more apt to be                 populations.
obese than Natives in the older cohorts. Finally, obese Native         Primary Funding Source: No Funding
elders were significantly more likely to have one or more
chronic diseases than non-obese Native elders.                         ●Gender, Race, and Socioeconomic Disparities in
Implications for Policy, Delivery, or Practice: Based on               Inpatient Quality of Care Measures
these findings, specific recommendations relate to increases           Helen Neikirk, M.A., Raj Behal, M.D., M.P.H.
in: prevention efforts including health promotion, screening
and wellness programs; disease management programs                     Presented By: Helen Neikirk, M.A., Program Director, Analytic
including access to services and a focus on preventing co-             Services, Clinical Informatics, University HealthSystem
morbidity; and Native elders’ access to insurance programs             Consortium, 2001 Spring Rd, Ste 700, Oak Brook, IL 60523;
such as Medicare. Native American elders can prevent obesity           Tel: (630) 954-2462; Fax: (630) 954-5879; Email:
through concerted efforts to improve and maintain good                 neikirk@uhc.edu
nutritional and exercise regimens. Elders who are obese                Research Objective: To identify gender, race, and
should work with their health care provider to lose weight and         socioeconomic disparities in performance on inpatient quality
increase physical activity, which has the following associated         of care measures in National Association of Public Hospitals
health benefits: decreased likelihood of having a stroke or            (NAPH) members and non-NAPH members.
developing diabetes; reduced blood pressure; improved                  Study Design: Performance of NAPH and non-NAPH
cholesterol; reduced blood sugar; and slowed progression of            hospitals on the JCAHO/CMS quality measures for AMI, heart
arthritis and reduction of joint damage/pain.                          failure and pneumonia was compared for three types of
Primary Funding Source: Administration on Aging                        disparity: males vs. females, Whites vs. non-Whites, and low
                                                                       income vs. higher income patients. Low income was based
●Effects of Antidiscrimination Legislation on Minority                 on primary payer being Medicaid, self pay/uninsured, charity,
Access to Quality Care                                                 or medically indigent services. In addition to performance on
Ari Mwachofi, Ph.D., Geraldine Ross, Ph.D.                             individual quality measures, a composite measure was created
                                                                       for each measure set (i.e., AMI, heart failure, pneumonia) to
Presented By: Ari Mwachofi, Ph.D., Assistant Professor,                indicate the percentage of patients who were eligible for at
Health Administration and Policy, OUHSC, 801 NE 13th St,               least one measure in a measure set that received all of the
Oklahoma City, OK 73104; Tel: (405) 271-2114; Fax: (405) 271-          recommended elements of care for which they qualified. Only
1868; Email: amwachof@ouhsc.edu                                        patients who received all of the elements of care for which
Research Objective: The purpose of the study is to determine           they were eligible were counted in the numerator for the
the extent to which legislation and judicial edicts can improve        composite rates.
minority access to quality care.                                       Population Studied: Data for all JCAHO/CMS core quality
Study Design: Using Rehabilitation Services Administration             measures for discharges from University HealthSystem
data on closed cases, the study compares African American              Consortium member hospitals during the time period January
access to rehabilitation services in 2004 to their access in 1937      2004 - June 2005 (N = 144,708). Hospitals were grouped into
before anti-discrimination legislation. It analyzes: total             NAPH members (37,255 discharges) and non-NAPH members
participation in rehabilitation programs relative to needs;            (107,453 discharges).
types of services rendered; relative expenditures; and                 Principal Findings: Statistically significant differences were
outcomes for Whites and for African Americans.                         found for many individual quality measures in both NAPH and
Population Studied: The study uses 291,135 cases from the              non-NAPH hospitals across all disparity types. Non-White
sixteen southern states and 150,018 cases from seven                   patients with LVSD were significantly more likely to have ACE
northern states for a total of 441,153 cases that were closed by       inhibitors prescribed for both AMI and heart failure in NAPH
rehabilitation services in 2004. This sample includes white            hospitals. In non-NAPH hospitals this was true only for heart
and African American clients.                                          failure. For AMI and heart failure a significantly higher
Principal Findings: The study finds a significant                      percentage of White patients received all of the care elements
improvement in the numbers of minority clients accessing               for which they were eligible in both NAPH and non-NAPH
services. There is also an improvement in spending on                  hospitals. For pneumonia a significantly higher percentage of
minority clients. However the study finds that the                     non-White patients received all of the care elements in non-
antidiscrimination legislation has not improved the quality of         NAPH hospitals. Significantly higher proportions of males
care received by minority clients to the same extent as the            received all of the elements of care for AMI and pneumonia
improvement in terms of numbers accessing services. There              and a significantly higher percentage of females received all of
is still a lack of equity in the quality of services rendered and in   the care elements for heart failure in non-NAPH hospitals.
the outcomes.                                                          Low income AMI and heart failure patients were significantly
less likely to receive all of the elements of care for which they   covariates, the odds of having a disability (OR=2.1; 95% CI
were eligible in both NAPH and non-NAPH hospitals. For              [1.16-3.95], fair/poor HS (OR=2.1; [1.12-3.82], poor PH (OR=2.0
pneumonia, this pattern was reversed with a significantly           [1.09-3.61], poor MH (OR=2.2; 1.24-3.87], more ALs (OR=3.1;
higher percentage of low income patients receiving all care         [1.64-5.85], and arthritis (OR=2.6; [1.41-4.97]) were two to three
elements in both NAPH and non-NAPH hospitals.                       times higher among respondents who reported ES/PS due to
Conclusions: Disparities in care exist in both NAPH and non-        PRD. Adjusting for other covariates (including health
NAPH hospitals with regard to race, gender and income/payer         insurance coverage), the odds of having a disability (OR=2.9;
status. Compared to heart failure and pneumonia, a higher           [1.46-5.62], fair/poor HS (OR=2.6; [1.22-5.41], poor PH (OR=3.6
percentage of AMI patients received all of the elements of care     [2.06-6.28], MH (OR=3.5;1.94-6.29], AL (OR=3.6; [1.81-7.26]),
for which they were eligible in both NAPH and non-NAPH              and arthritis (OR=3.0; [1.51-6.14]) in the past month were three
hospitals regardless of their race, gender or income/payer          to four times greater among those who reported experiences
status. Heart failure patients are least likely to receive all of   that were worse than other races when seeking health care,
the elements of care for which they are eligible in both NAPH       compared to those who reported experiences that were the
and non-NAPH hospitals regardless of race, gender or                same as other races.
income.                                                             Conclusions: ES and PS due to PRD and perceived worse
Implications for Policy, Delivery, or Practice: Hospitals that      treatment when seeking care are strongly associated with
are members of NAPH are sometimes referred to as "safety            adverse health outcomes including poor QOL and chronic
net" hospitals whose mission is to care for the disadvantaged       conditions such as arthritis among adults. African-Americans
in society. This study provides evidence that quality of care       are at a greater risk of experiencing ES/PS due to PRD and are
disparities with regard to race, gender and income exist in         more likely to encounter PRD when seeking care than whites.
these "safety net" hospitals as well as in other hospitals.         These associations persist even after adjustment for relevant
Primary Funding Source: No Funding                                  covariates.
                                                                    Implications for Policy, Delivery, or Practice: Our findings
●Is Perceived Racial Discrimination Associated with                 suggest the need for increased awareness of the potential role
Quality of Life and Self-Reported Chronic Conditions?               of PRD on health and health disparities, and more research to
Results from the 2002 North Carolina BRFSS Survey.                  examine the mechanism through which PRD influences health
Emmanuel M. Ngui, DrPH, Ziya Gizlice, Ph.D.                         and QOL. The study findings also underscore the need for
                                                                    policies and strategies that reduce and eliminate racial
Presented By: Emmanuel M. Ngui, DrPH, Assistant Professor           discrimination or unequal racial treatment our society and in
of Pediatrics, Epidemiology and Health Policy, Department of        the healthcare system.
Pediatrics, Medical College of Wisconsin, 8701 Watertown            Primary Funding Source: No Funding
Plank Rd, Milwaukee, WI 53226; Tel: 414 456-4302; Fax: 414
456-6385; Email: engui@mcw.edu                                      ●Are Pediatric Asthmatics Receiving the Necessary
Research Objective: The study objectives were to examine the        Medications?
association of 1) self-reported emotional symptoms (ES) and         Jeannette Oshitoye, Ph.D.
physical symptoms (PS) due to perceived racial discrimination
(PRD); and 2) PRD experiences when seeking healthcare with          Presented By: Jeannette Oshitoye, Ph.D., Adjunct Professor,
quality of life (QOL) and self-reported physician-diagnosed         Social Sciences, University of Texas at Dallas, 2915 Antares
chronic conditions.                                                 Circle, Garland, TX 75044; Tel: 972-925-6428;
Study Design: Cross-sectional data from the 2002 North              Email: Joshitoye@dallasisd.org
Carolina Behavioral Risk Factor Surveillance System (BRFSS).        Research Objective: To determine whether asthmatic
PRD was assessed using self-reported 1) ES(i.e., angry, sad, or     patients are receiving required medications outlined by the
frustrated) and PS (i.e., headache, upset stomach, or               National Asthma Guidelines.
pounding heart) due to racial treatment within the past 30          Study Design: This study analyzes data from the Medical
days; and 2) perceived differential racial treatment when           Expenditure Panel Survey (MEPS). It uses a multinomial logit
seeking healthcare within the past 30 days. Dependent               regression to predict the probability of a pediatric asthma
variables included: 1) five QOL measures (perceived disability,     patient receiving appropriate medications as outlined in the
self-rated fair or poor health status(HS), seven or more days       national guidelines.
of poor physical health (PH), seven or more days of poor            Population Studied: U.S. asthmatics between the ages of 5-17
mental health (MH), and activity limitation (AL) in the past        years old
month chronic), and 2) self-reported physician-diagnosed            Principal Findings: African Americans and Latinos are least
chronic conditions (arthritis, diabetes, and current asthma).       likely to receive the required medications as outlined by the
Multivariate analyses adjusted for race, gender, age, education     National Asthma Guidelines. African Americans and Latinos
level, and household income.                                        are least likely to receive the required medications as outlined
Population Studied: Random sample of North Carolina                 by the National Asthma Guidelines. Having insurance had no
adults.                                                             effect on the results.
Principal Findings: Seven percent of North Carolina adults          Conclusions: African Americans and Latinos are least likely to
reported ES and PS associated with PRD in the past 30 days,         receive the required medications as outlined by the National
with African-Americans disproportionately more likely than          Asthma Guidelines. Having insurance had no effect on the
whites to report ES (17% vs. 3%), PS (8% vs. 1%), and               results. It is proposed that this is due to a lack of cultural
experiences that were worse than other races when seeking           competency or the part of physicians.
health care (7% vs. 1%). Adjusting for other relevant
Implications for Policy, Delivery, or Practice: These findings      treatment levels and life expectancy for U.S. blacks. This
could have significant policy implications especially in regards    differential finding may be a result of VA’s integrated health
to medical school training and continuing education credits         care system which reduces barriers to care through subsidized
for new and practicing physicians.                                  comprehensive health care services. Our objective was to
Primary Funding Source: No Funding                                  examine the role of systems of care in racial health disparities
                                                                    by comparing 30-day mortality for whites and blacks following
●Self-Order Purchasing of Antibacterial in a Developing             hospital admission to VA and non-VA hospitals.
Country – India                                                     Study Design: Retrospective observational study. We
Bhagirathkumar Patel, M.Pharm., Ph.D., Mr Nikhl Patel,              examined 30-day mortality rates for blacks and whites
B.Pharm., Ramesh K. Goyal, Ph.D., Jigneshkumar Patel, M.Sc.         following hospital admission for six medical conditions
                                                                    between 1996 and 2001 in all acute care VA hospitals, and in
Presented By: Bhagirathkumar Patel, M.Pharm., Ph.D.,                acute care non-VA hospitals in Pennsylvania and California.
Pharmacologist, Pharmacology, Shri B M SHah College Of              Discharge data were linked to death certificate data and data
Pharmacy, College Campus, MODASA, 383315; Tel: +91                  from the Beneficiary Identification Record Locator System
9426391015; Email: bhagirath70@yahoo.com                            within the VA for information on out of hospital deaths.
Research Objective: With ever widening medical research             Logistic regression was used to adjust for differences in
and an increasing supply of miracle drugs being discovered          demographic factors, comorbidities, and race within VA and
and made available, antibacterial resistance is global public       non-VA hospitals. Interactions between race, hospital system,
health service problem. In the present study, a profile of 1178     and age were tested.
self orders were monitored at retail pharmacy outlets of North      Population Studied: The dataset consisted of 346,301 VA
Gujarat, India with the aim of investigating status of self         patients and 2,935,543 non VA patients, admitted to VA
purchase of antibiotics amongst the non medical population.         hospitals nationwide or non-VA hospitals in Pennsylvania or
Study Design: Data regarding self-order purchase profiles           California with a principal diagnosis of one of 6 AHRQ
were collected by qualified pharmacist, using specially             Inpatient Quality Indicators: acute myocardial infarction,
prepared and pre-tested questionnaires. The collected data          congestive heart failure, gastrointestinal bleeding, hip fracture,
were tabulated and analyzed by using normal proportion test.        pneumonia, or stroke.
Population Studied: Total population is around 2 millions in        Principal Findings: Among subjects under 65, blacks and
the area of North Gujarat, India.                                   whites had similar 30-day mortality for 5 of 6 conditions in
Principal Findings: It was observed that 45% of patients            both VA and non-VA hospitals. For example, odds ratios for
purchasing over the counter (OTC) antibiotics did not               AMI: VA:1.13, 95%CI: 0.94-1.37; and non-VA:0.95, 95%CI:0.86-
purchase a sufficient quantity of antibiotics to cover the          1.04. Among those over age 65, blacks were less likely to die
minimum duration of therapy. The most often respondents             than whites in both VA and non-VA hospitals. Odds ratio AMI:
indicated following antibiotics for self-medication: Penicillins-   VA:0.81 95%CI:0.72-0.91; non-VA:0.84 95%CI:0.79-0.88. The
42% (Amoxycillin/Ampicillin), Quinolones-20%, Co-                   gap in mortality between blacks and whites was comparable
trimoxozole-16%, Cephalosporines-14%, Erythromycin-13%,             between VA and non-VA hospitals in 4 of the 6 conditions. In
and Tetracyclines-6%. However, since patient in irregular           the other two conditions, CHF and pneumonia, this gap in
dosage schedule purchased them, resistant strains of highly         mortality favored blacks more in non-VA hospitals compared
pathogenic organism may develop.                                    to VA hospitals.
Conclusions: Rationalization of self-medication and                 Conclusions: These findings suggest that factors associated
introduction of extended professional services should be            with better 30-day mortality for blacks are not unique to VA.
directed towards achieving better patient heath care.               Implications for Policy, Delivery, or Practice: These results
Implications for Policy, Delivery, or Practice: Health care         suggest that interventions aimed at reducing health disparities
professional (pharmacist) can play a greater role in health         and improving health outcomes and life expectancy of blacks
education by providing information to the public about the          may need to focus on settings besides hospital care.
detrimental effects on health of the irregular an unnecessary       Primary Funding Source: VA, VA Health Services Research &
use of the broad-spectrum antibacterials.                           Development
Primary Funding Source: No Funding
                                                                    ●A New Approach to Facilitating Substance Use Disorder
●Is Lower Hospital Mortality Among Blacks Unique to the             Treatment Effectiveness, Efficiency and Accountability
Veterans Administration?                                            Janice Pringle, Ph.D., Michael T. Flaherty, Ph.D., Amy
Daniel Polsky, Ph.D., Judith Lave, Ph.D., Ashish Jha, M.D.,         Shanahan, Stephanie Murtaugh
M.P.H., Mark Pauly, Ph.D., Zhen Chen, Ph.D., Kevin Volpp,
M.D., Ph.D.                                                         Presented By: Janice Pringle, Ph.D., Scientific Director, ,
                                                                    Institute for Research, Education, and Training in Addictions
Presented By: Daniel Polsky, Ph.D., Affiliate Investigator/         (IRETA), 425 Sixth Ave, Pittsburgh, PA 15219; Tel: 412-258-8556;
Assistant Professor, Center for Health Equity Research and          Fax: 412-258-8554; Email: pringlej@ireta.org
Promotion, Philadelphia VA Medical Center, Blockley Hall,           Research Objective: Examine the impact of applying a
Rm. 1212, Philadelphia, PA 19104-6021; Tel: (215) 573-5752; Fax:    complex approach addressing organizational/system
(215) 898-0611; Email: polsky@mail.med.upenn.edu                    leadership, real time performance measurement and feedback,
Research Objective: 30-day mortality following hospital             an internal learning strategy using an industry-based process
admission within VA is lower for black patients relative to         improvement strategy, training on an appropriate evidence-
white patients; a finding that contrasts sharply with the lower     based practice, and an application of a continuum of care
approach to substance use disorder treatment (SUD) upon            ●Complementary and Alternative Medicine Use in Chinese
performance indicators such as treatment process (access           and White
and retention) and patient recovery progress (housing status,      Hude Quan, M.D. Ph.D., Daniel Lai, Ph.D., Delaine Johnson,
employment status, alcohol and drug use profile, and criminal      MA, Marja Verhoef, Ph.D., Richard Musto, M.D.
justice involvement).
Study Design: A multi-disciplinary team applied ongoing            Presented By: Hude Quan, M.D. Ph.D., Assistant Professor,
training, mentoring, and implemented a real time                   Community Health Sciences, University of Calgary, 3330
performance measurement and reporting system within a              Hospital Dr. NW, Calgary, T2N 4N1; Tel: (403) 944 8912; Fax:
system of SUD treatment in Western Pennsylvania using              (403) 210 3818; Email: hquan@ucalgary.ca
protocols embedded within complex adaptive systems theory.         Research Objective: Complementary and alternative
The team gathered data describing how the project was              medicine (CAM) is commonly utilized. Health care
implemented, and collected three waves of performance              professionals and health policy makers need to know what
measurement data along various indicators of treatment             CAM therapies are being used by their patients. CAM is
process and recovery progress for patients receiving               closely related to cultural background. Acupuncture and herbal
outpatient and residential SUD treatment. Focus groups, key        therapies originated in China. Therefore, there may be
informant interviews and surveys were used to                      significant differences in characteristics of Chinese and White
describe/improve the implementation process.                       CAM users. Hence, this study aimed to describe the
Population Studied: Over 400 adult (18 and older) persons          differences and similarities in CAM use between the Chinese
with SUDs who sought various levels of inpatient or                and the White Canadian populations. Currently, few data are
outpatient treatment within one system of care.                    available on this topic.
Principal Findings: Process data collected indicated that all      Study Design: A cross-sectional telephone survey was
four elements of the system were key to the success of the         conducted in English and two Chinese dialects (Cantonese
project (leadership, internal/external learning, performance       and Mandarin) in a large Canadian city (Calgary). Information
measurement, and a clinical model). Sustaining and                 on utilization of 12 CAM modalities in the past 12 months was
expanding the effort was supported best by changes in              collected.
leadership style that articulated a clear vision for the work      Population Studied: The target population of this study
being conducted, and a process for achieving that vision.          consisted of Chinese and White aged 18 or over in Calgary. We
Staff buy in was accelerated when the effort launched a            randomly selected telephone numbers listed with a Chinese
successful, concurrent paperwork reduction effort. Results of      surname from the 2004 Calgary telephone directory to obtain
this pilot have been promising. Some of the findings include:      Chinese Canadians. We then randomly selected telephone
Reduced amount of time between first contact and first             numbers from remaining names. Those selected were asked
treatment session from an average of 14 days to 48 hours.          to identify their ethnicity. Only White were included.
Increased program census over one year period by at least          Principal Findings: Of 1727 Chinese and 1948 White
50%. Increased program revenues over one year period by            surveyed, 850 (49.2%) Chinese and 805 (41.3%)White
over 25% (in a fee for service environment). Reduced               responded to the survey. Of the Chinese respondents, 80.5%
provider-developed redundant and unnecessary paperwork             used CAM, mainly vitamins and/or mineral supplements
burden by approximately 40%. Statistically significantly           (58.4%), herbal remedies (48.7%), massage therapy (17.1%),
increased client retention (p< 0.008) over ten month period.       and chiropractic (8.4%). Of the white respondents, 86.3%
Clients who completed treatment and who were referred from         used CAM, mainly vitamins and/or mineral supplements
the County jail - 84% had ceased drug and alcohol use, 60%         (79.1%), herbal remedies (33.7%), massage therapy (30.4%)
were employed, 92% had no probation/parole violation and           and chiropractic (21.2%). Acupuncture use was similar in both
100% were stably housed six months after discharge.                groups: 8.2% in Chinese and 7.9% in White participants. The
Conclusions: The use of complex adaptive systems theory, as        remaining 8 CAM modalities surveyed (including naturopath,
it is applied in other healthcare , which underscores the use of   homeopathy, amino acids or enzyme supplements, Ayurvedic
simultaneous strategies along clearly identified system            medicine, biofeedback training, hypnosis and Reiki) were less
“levers” holds much promise in establishing, sustaining and        frequently used by both Chinese and White participants. White
expanding efforts to facilitate SUD treatment effectiveness,       participants were more likely to use multiple-CAM modalities
efficiency and accountability.                                     than Chinese. For example, 14.0% White used 4 or more CAM
Implications for Policy, Delivery, or Practice: This approach      modalities versus 5.0% Chinese in the last 12 months.
holds promise for use by federal, state and private payers to      Conclusions: Chinese and White Calgarians were similar with
address quality issues in SUD treatment. The authors strongly      respect to overall CAM use. However, the level of CAM
suggest payers consider this approach when applying training,      utilization varied by type of CAM therapy between Chinese and
performance measurement and process improvement                    White. Chinese utilized herbal therapies far more than
approaches currently being supported at the federal level (i.e.,   acupuncture although both are traditional Chinese therapies.
NOMs, NIATx), each of which have had limited impact upon           The variation may be related to patient’s knowledge and
treatment quality. This approach could also have profound          beliefs about effectiveness of individual CAM modularity.
implications for how NIH funds and translates scientific           Primary Funding Source: Alberta Heritage Foundation for
findings into practice, which is often based upon simple           Medical Research
assumptions about organizational/system behavior.
Primary Funding Source: No Funding
●Cancer Patients' Attitudes toward Cancer Trials                   probit (DIHOPIT) model based on asset ownership, and the
Peggy Schuber, DSN, RN                                             concentration index was calculated to quantify utilization
                                                                   inequality across income groups. Additionally, a Chi-squared
Presented By: Peggy Schuber, DSN, RN, Assistant Professor          statistic was used to compare the proportions, and the
of Community Health Nursing, School of Nursing, University         association between independent covariates and the binary
of Nevada, Las Vegas, 4505 maryland pkwy, Las Vegas, NV            dependent variable of refraining were obtained from a logistic
89154; Tel: 702-895-5440; Fax: 702-895-4807; Email:                regression analysis.
peggy.schuber@unlv.edu                                             Population Studied: We used the data from the 2002 Iranian
Research Objective: Accurate measurement of attitudes              National Healthcare Cost and Utilization Survey implemented
toward participation in cancer treatment trials (CTs) and          by the Ministry of Health and Medical Education. A nationally
cancer prevention trials (CPTs) across varied groups could         representative sample was obtained by using probabilities
assist health researchers and educators when addressing            proportional to size, resulting in a total of 3479 households
attitudinal barriers to participation in these trials.             comprising 16860 individuals (50.3% male).
Study Design: Two newly developed scales to measure                Principal Findings: With a response rate of 95%,
attitudes toward CTs and CPTs comprise the Attitudes toward        3624(21.6%) out of 16807 respondents expressed a need to
Cancer Trials Scales (ACTS). These scales were used to             outpatient care over the last 2 weeks. From those with a
evaluate the prevalence of positive and negative attitudes         perceived need, 1079 (29.8%) refrained from seeking a care;
toward cancer treatment and prevention trials in a random          from those who sought a care 97.8% received it. The
stratified population of cancer patients.                          concentration index was calculated as         -0.172 (SE: 0.003),
Population Studied: The CT and CPT scales with                     indicating that the poor refrain from visiting healthcare more
sociodemographic and cancer trial history variables were           often than the better-off. Economic barriers and self-
distributed in a mail survey of former patients (N = 1,703) of a   medication were the two most common explanations for
comprehensive cancer center that yielded 925 usable                refraining behavior expressed by the population. 18.9% and
responses (54% response rate).                                     65% of the wealthiest and poorest quintiles mentioned
Principal Findings: Prevalence of favorable attitudes toward       economic barriers as the reason for refraining, respectively
CTs and CPTs was 66% and 69%, respectively. There were no          (p<0.0000). The corresponding figures for self-medication
significant differences in mean scores by cancer site or           were 38.7% and 23.8 %.( p<0.0000). Increased odds of
gender, but African Americans had more positive attitudes          refraining were found for rural populations, males, the lowest
toward cancer trials than European Americans. Multiple             two income quintiles, the uninsured, and those with
regression analysis indicated that older age, lower education      household heads with low education levels.
level, and prior CT participation, but not ethnicity, were         Conclusions: The proportion of the population who sought
associated with more favorable attitudes toward CTs. Prior CT      no care was comparable with that of some other nations, and
participation and prior CPT participation were associated with     results show that the poor are less likely to seek care than the
more favorable attitudes toward CPTs. Ethnicity was                better-off. This implies the need for further research to assess
significantly associated with response rate. Evidence of           the possible impact of any targeted intervention in order to
reliability and construct validity of both scales is presented.    reduce the inequality gap in refraining behavior. Furthermore,
Conclusions: Use of a psychometrically sound instrument to         the problems behind self-medication and drug use remain
measure attitudes may lead to a better understanding of the        unanswered and need to be investigated thoroughly.
reasons for decisions regarding participation in CTs and CPTs.     Implications for Policy, Delivery, or Practice: As the poor
Implications for Policy, Delivery, or Practice: Access to          are more likely to refuse seeking healthcare than the well-off,
cancer trials and cancer prevention trials should be ensured       policies aiming to increase access of poor households to
across demographic groups.                                         healthcare may positively affect the absolute level of the
Primary Funding Source: AHRQ                                       population health and help reduce the health inequalities in
                                                                   the population.
●Non-Utilization of Outpatient Care in Iran:Differntials           Primary Funding Source: No Funding
and Determinants
Saeid Shahraz, M.D., Emmanuela Gakidou, Ph.D., Mohsen              ●The Role of Perceived Racism and Race-Based
Naghavi, M.D. Ph.D., Hamid Reza Jamshidi, Ph.D., Diania            Residential Segregation on Cancer Behavioral Risk Profiles
Lee, Bs, Goran Tomson, M.D. Ph.D.                                  Salma Shariff-Marco, M.P.H.

Presented By: Saeid Shahraz, M.D., Post Doc Fellow, Harvard        Presented By: Salma Shariff-Marco, M.P.H., Ph.D. Candidate,
Initiative for Global Health, 104 Mount Auburn street ,3rd         Health, Behavior and Society, Johns Hopkins Bloomberg
Floor, Cambridge, MA 02138; Tel: (617) 495-8236; Fax: (617)        School of Public Health, 6101 43rd Street, Riverdale, MD
495-8231; Email: saeid_shahraz@harvard.edu                         20737; Tel: (301) 277-4536; Email: sshariff@jhsph.edu
Research Objective: To measure the distribution of refraining      Research Objective: The objectives of this research are to: 1)
behavior across income groups in Iran and illustrate reasons       describe the range of racism experienced by racial/ethnic
for non-utilization.                                               groups in California; 2) describe the relationship between
Study Design: Perceived need for care was measured for a           racism (perceived racism at the individual level and
two week recall period, with an additional 14-item                 segregation at the county level) and cancer prevention
questionnaire detailing reasons for refraining from care           behaviors; and 3) identify potential mediators and moderators
(defined as allopathic or traditional medicine). Income groups     of the relationship including socioeconomic position,
were assigned using a dichotomous hierarchical ordered             immigration status, acculturation (e.g., language, length of US
residence), and neighborhood characteristics between racism        ●Variance in Outcomes and Care Patterns for Diabetes by
and cancer prevention.                                             Insurance Status: What Are the Underlying Causes?
Study Design: Using cross-sectional data from the 2003             Jay Shen, Ph.D., Elmer Washington, M.D., M.P.H.
California Health Interview Survey, the study draws from over
42,000 adult respondents representing the five major               Presented By: Jay Shen, Ph.D., Associate Professor, Health
aggregate racial/ethnic groups in the US. Perceived racism is      Administration, Governors State University, One University
measured with questions about general exposure to racism as        Parkway, University Park, IL 60466; Tel: (708) 235-2131; Fax:
well as exposure within the health care context. Segregation is    (708) 534-8041; Email: j-shen@govst.edu
measured on two dimensions, unevenness and exposure via            Research Objective: To evaluate patterns of hospital based
Index of Dissimilarity and Isolation Index. Cancer prevention      care for patients with Type-2 diabetes of varying insurance
will be measured with a set of primary (e.g., avoiding tobacco     statuses since increase in uninsured population has become a
use, regular exercise) and secondary (e.g., early detection)       serious policy issue in recent years.
prevention behaviors. Exploratory analysis is conducted            Study Design: Data were classified and evaluated according
primarily using univariate and bivariate analyses. For RQ2 &       to insurance status. Accompanying information on
RQ3, ordinal logistic regression and multilevel modeling are       sociodemographic variables were included for comparison as
utilized for the individual level and multilevel analyses,         well. All admissions where Type-2 diabetes was entered as any
respectively.                                                      diagnostic code (primary, secondary, tertiary, etc. were
Population Studied: The study population includes all adult        classified in one of three tiers based upon the primary
respondents from 2003 CHIS who have never been diagnosed           diagnosis. Tier 1 included those diagnoses that were likely to
with cancer, are not pregnant and did not have a proxy             unrelated to the diagnosis of diabetes (e.g. cerebral
interview. In addition, all respondents have provided              hemorrhage, salmonella, etc.). For certain infectious primary
responses to cancer prevention and perceived racism and            diagnoses (e.g. salmonella), the presence of diabetes may
have been successfully geo-coded to a census tract. The 6          have been more likely to cause the infection to be more
major racial/ethnic groups considered in the study include         severe, however, the primary determinant of infection was
Latinos, Non-Hispanic African Americans, Non-Hispanic              more likely to be exposure as opposed to host characteristics.
Asians/Pacific Islanders, Non-Hispanic American                    Tier 2 included those diagnoses for which diabetes was a
Indians/Alaskan Natives, and Non-Hispanic Whites.                  noted risk factor (e.g. myocardial infarction, cellulitis, etc.).
Principal Findings: Approximately 20% of the respondents           Tier 3 included those diagnoses that virtually never occur in
reported experiencing overall discrimination sometimes, often      the absence of diabetes (e.g. Kimmelstein Wilson
or frequently, while 6% of the respondents reported                nephrosclerosis, hyperosmolar coma, etc.). In addition, we
experiencing discrimination within the health care context.        evaluated patterns of admission for principal diagnoses of
These responses varied by race/ethnicity, ranging from 12% to      acute hypoglycemia and acute hyperglycemia across insurance
60% for overall discrimination and 3%-14% within the health        groups.
care context. Overall, secondary cancer prevention is more         Population Studied: 642,610 patients included in the 2001
strongly associated with both measures of racism compared          National Inpatient Sample. Among them, 141,406 were
to primary cancer prevention in unadjusted analyses. These         privately insured, 56,461 were covered by Medicaid, and
associations vary by race/ethnicity.                               20,046 were uninsured. The remaining were Medicare
Conclusions: Preliminary findings of unadjusted analyses           patients.
suggest that the relationship between racism and the two           Principal Findings: Medicaid (6.1%) and uninsured patients
types of cancer prevention (primary and secondary) differ.         (6.8%) were more likely to be admitted with a tier 3 diagnosis
Increased exposure to racism is associated with decreased          than the private insurance counterparts (4.4%) (odds ratios
participation in primary prevention and increased participation    (OR) [95% confidence interval (CI)], 1.25 [1.19, 1.31] for
in secondary prevention. A major limitation of this study is       Medicaid patients and 1.36 [1.27, 1.45] for uninsured). Further,
that the data are cross-sectional and thus, causal inferences      compared to patients covered by privated insurance,
are not possible.                                                  uninsured patients were more likely to be admitted with an
Implications for Policy, Delivery, or Practice: The research       acute, severe hyperglycemic episode (OR [CI], 1.91 [1.57, 2.31])
provides information on the prevalence of perceived racism         while both Medicaid and uninsured patients were more likely
across all major racial/ethnic groups and the effect that it has   to be admitted with an acute hypoglycemic episode (OR [CI],
on preventive behaviors. The research findings may support         1.20 [1.13, 1.28] for Medicaid patients and 1.43 [1.30, 1.57] for
and inform health care setting interventions addressing            uninsured patients). Finally, uninsured patients had the
discrimination and promoting cultural competency.                  youngest average age of admission (52.9 years) followed by
Primary Funding Source: No Funding                                 Medicaid (54.7 years). The average age for the privately
                                                                   insured patients was 57.4 years.
                                                                   Conclusions: This study suggests that uninsured populations
                                                                   are more likely to be admitted to the hospital at a relatively
                                                                   young age as a result of a diabetic complication that could be
                                                                   prevented or delayed. Uninsured patients were also more likely
                                                                   to be admitted from an episode related to acute, uncontrolled
                                                                   hyperglycemia generally suggesting that this group had the
                                                                   most difficulty with access to care. While Medicaid patients
                                                                   were also more likely to be admitted from a diabetic
                                                                   complication that might have been delayed by timely access to
appropriate outpatient care, they had comparable risk of being      Conclusions: Disparities in health care are not yet an area of
admitted for acute, severe hyperglycemia when compared to           operational or quality improvement focus for hospitals,
the privately insured group. Although Medicaid patients had a       despite the fact that “equity” is one of the six domains of
lower risk of being admitted with an acute hypoglycemic             quality identified by the Institute of Medicine. The lack of
episode when compared to the uninsured, the risk was higher         specific disparity reduction projects among the applicant
than for patients with private insurance.                           hospitals is especially surprising given 1) that the applicant
Implications for Policy, Delivery, or Practice: Policies            hospitals may be a sample expected to be biased towards
focused on provider reimbursement directly tied to care for         institutions with an interest in disparities; and 2) almost all
the uninsured are critical to reducing disparities in this          collect race and ethnicity data on patients.
population. With respect to Medicaid populations, funding           Implications for Policy, Delivery, or Practice: Disparities are
for appropriate levels of case management to improve care           not “top-of-mind" for health systems leadership, and are not a
coordination and follow up are likely to improve outcomes.          major focus of hospital quality activities. Changing this may
Primary Funding Source: AHRQ                                        require concerted campaigns to link disparities to the
                                                                    movement to improve quality, and to demonstrate that quality
●Findings from the Planning and Launch of a National                improvement initiatives can indeed reduce disparities in
Disparities Collaborative                                           clinical settings. Current policies around hospital
Bruce Siegel, M.D.,M.P.H., Marcia J. Wilson, M.B.A., Lea            accreditation, data reporting and regulation may also need to
Nolan, MS, Lissette Vaquerano, Khoa Nguyen, M.P.H.,                 be reformed to create incentives for hospitals to view health
Deborah Bear                                                        care disparities as priorities.
                                                                    Primary Funding Source: RWJF
Presented By: Bruce Siegel, M.D.,M.P.H., Research Professor,
Health Policy, George Washington University School of Public        ●A Comparison of Demographics, Social History, Clinical
Health and Health Services, 2021 K St, NW, Ste 800,                 Outcomes, and Resource Utilization by Insurance Payer
Washington, DC 20006; Tel: 2025302399; Fax: 202-296-0025;           Group in Patients Undergoing Total Hip Replacement
Email: siegelmd@gwu.edu                                             Surgery
Research Objective: Expecting Success is a national initiative      Amanda Smith, M.P.H., Adrian Hinman, M.D.
that seeks to use quality improvement techniques to reduce
disparities in heart failure and acute myocardial infarction care   Presented By: Amanda Smith, M.P.H., Research Director,
for African-Americans and Latinos. We sought to understand          Orthopaedic Surgery, UCSF, 3333 California St., #265, San
the visibility of disparities in health care for health system      Francisco, CA 94116; Tel: 415-439-9472; Email:
leaders as part of the planning and design process for this         smitha@ucsf.edu
collaborative. We also attempted to characterize the extent to      Research Objective: Numerous studies have shown that
which hospitals applying to the program collected race and          Medicaid patients have less access to medical care, but few
ethnicity data and engaged in initiatives to reduce disparities.    studies have focused on the consequences of reduced access
Study Design: Key informants were identified through                by the poor to needed specialty care. The purpose of this
discussions with health care philanthropies, professional           study was to determine if Medicaid patients receiving a
associations and researchers. These informants were then            primary total hip replacement differed from other patients
interviewed for one hour using a semi-structured instrument.        with respect to baseline demographic characteristics, social
Information from these interviews was used to construct a 22        history, clinical outcomes, and resource utilization.
item web-based letter of intent (LOI) used by hospitals to          Study Design: A retrospective chart review of subjects who
indicate interest in applying to the collaborative. The LOI         received a primary total hip replacement (THR) was
included questions on hospital characteristics as well as past      conducted and stratified by insurance payer groups (subjects
quality improvement and disparities initiatives.                    reporting more than one source of insurance were excluded).
Population Studied: Forty-two individuals were interviewed,         Baseline demographic information, social history, operative
including healthcare system, association and philanthropy           data, resource utilization and clinical outcomes (using the
leaders, and researchers and clinicians. Subsequently 122           Harris Hip Score, a measure of functional outcomes in total
hospitals completed the LOI. Of these, 29 were community            hip replacement surgery) were compared between Medicaid,
hospitals, 59 were teaching institutions, and 34 identified as      Medicare and managed care participants. Additionally, we
academic medical centers. The largest group (82) was private        measured how far patients lived from the clinic where they
not-for-profit institutions, with 68 serving a mainly urban         were having surgery to measure discrepancies in access.
market.                                                             Population Studied: Consecutive patients at an academic
Principal Findings: In the initial planning phase all of the        medical clinic who received a primary total hip replacement
interviewees reported that disparities in health care are not       who had at least 16 months of post-operative follow-up.
viewed as an important quality of care issue by American            Patients were excluded from analysis if they had a total hip
hospitals. Data from the subsequent LOI revealed that 118 of        replacement (THR) contra-laterally, any revision surgery, or a
the 122 applicant hospitals reported they collect data on           total knee replacement within this 16 month post-operative
patients’ race and ethnicity, and 84 stated they could analyze      period. Records for 78 Medicaid, 54 Medicare and 54
quality data by race and ethnicity. However only six hospitals      managed care participants were analyzed.
indicated that they had planned or implemented formal quality       Principal Findings: In terms of demographics there were
improvement initiatives specifically designed to reduce ethnic      statistically significantly more African-American (p<.05) and
or racial disparities in hospital care.                             Hispanic (p<.05) patients in the Medicaid group and they
                                                                    were significantly more likely to be single that their Medicare
or managed-care counterparts (p<.05). Additionally, though         trust, and access to care influence screening with prostate
Medicaid patients were more likely to smoke tobacco than           specific antigen (PSA) in North Carolina men.
Medicare and managed-care patients (p<.05) there was no            Study Design: Incident prostate cancer cases were identified
statistically significant difference in history of psychiatric     from the North Carolina Central Cancer Registry, and
illness, abuse of non-prescription drugs, use of heavy             standardized telephone questionnaires were conducted
narcotics, or abuse of alcohol. After adjusting for age, we        between November 2001 and July 2004. Physician trust was
found a significant difference in clinical outcomes between        measured as agreement with the statement “Overall, I trust
insurance groups in the pre- (p<.0001) and post-operative (at      my physician.” Access variables included income, health
6 months p<.0001 and at 1 year p< .0001) using the Harris          insurance status, access to a physician for prostate problems,
Hip Scores, although the rate of improvement in hip function       and having had regular checkups in the 12 months prior to
and pain after primary total hip replacement was similar           diagnosis. The outcome variable was self-reported use of PSA
between all three groups (p = .41). Compared to other payer        examinations in the 12 months prior to diagnosis. The
groups, Medicaid patients had to travel twice as far to receive    bivariate relationship between race and PSA usage was
treatment (median of 66.0 miles vs. 38.3 miles). However, the      examined using chi-square analyses. For descriptive purposes,
resource utilization, defined by the amount of appointments        the bivariate relationship between race and both trust and
listed as cancellations or no-shows, and the total number of       access was also examined. Subsequently, multivariate logistic
appointments needed both pre-operatively and post-                 regression examined whether the relationship between race
operatively, was not different among the two groups.               and PSA usage remained when simultaneously considering
Conclusions: Medicaid patients start out at baseline with          access and trust.
similar co-morbidities and demographics as other payer             Population Studied: White (n=347) and black (n=207) men
groups, but present with significantly lower Harris Hip Scores.    between 40 and 75 years diagnosed with prostate cancer
Decreased access to health care and increased time to              within three months prior to the survey (total n = 554).
presentation may account for the decreased functional scores       Principal Findings: Bivariate analyses revealed that compared
seen at the time of presentation. Additionally, although the       with whites, blacks were less likely to have had regular PSAs
number of appointments, cancellations and no-shows was not         (p<.01). Ninety-two percent of whites and 69% of blacks
significantly different between payer groups, there was a trend    reported having regular PSAs in the last 12 months.
for the Medicaid patients to have fewer pre-operative visits,      Compared with whites, blacks also had less access to care, as
which also would also indicate their hip function was more         they were more likely to be uninsured (8% vs. 2%), more to
impaired than patients with other insurance payers.                make less than $20,000 annually (11% vs. 5%), less likely to
Implications for Policy, Delivery, or Practice: As a greater       have seen a doctor’s office for their prostate problem (71% vs.
number of hospitals and clinics are turning away Medi-Cal          85%), and less likely to have had regular checkups in the 12
patients, they have to travel further for their care, present at   months prior to diagnosis (89% vs. 96%) (all p<.05). Blacks
later stages of disease and are in greater need of treatment       were also less trusting of their physician (87% vs. 97%, p<.05).
and therefore undergo treatment sooner after initial               Logistic regression revealed that whites were over four times
presentation than patients in other insurance groups who           as likely as blacks to have had regular PSAs (OR: 4.4, 2.7-7.0).
have more ready access. While decreased reimbursement is           Moreover, race retained predictive power when
implicated as the main cause of the dwindling acceptance of        simultaneously considering trust and access, although trust
Medi-Cal insurance, the complex social situations of Medi-Cal      was not a significant predictor. The multivariate model
patients, who are frequently unemployed and marginally             revealed that men who were most likely to have had regular
housed, may cause some providers to be reluctant to see and        PSAs had regular checkups prior to diagnosis (OR: 7.8, 3.4-
treat Medi-Cal patients. While our results indicate that Medi-     17.7), were white (OR:3.0,1.7-5.1), and saw a doctor’s office for
Cal patients are more likely to be single therefore perhaps        their prostate problems (OR: 2.0,1.1-3.5) (all p<.05).
lacking a support structure, they were not significantly more      Conclusions: Race, access to a doctor’s office, and access to
likely to engage in recreational drug use, alcohol abuse or        regular checkups significantly predict who is likely to undergo
heavy narcotic use than other patients, nor were they more         PSA screenings in North Carolina men with prostate cancer.
likely than other patients to miss or cancel their appointments.   When considered with these other variables, physician trust
Primary Funding Source: No Funding                                 does not significantly impact PSA usage.
                                                                   Implications for Policy, Delivery, or Practice: This work
●Do Racial Disparities in Access to Care and Physician             indicates the need to examine disparities in PSA screening
Trust Influence Prostate Cancer Screening?                         rates from many perspectives. Interventions to improve
Pamela Spain, Ph.D., Paul Godley, M.D.                             access to care as well as education about recommended
                                                                   screening services may reduce health disparities in prostate
Presented By: Pamela Spain, Ph.D., Postdoctoral Research           cancer outcomes.
Fellow, Cecil G. Sheps Center for Health Services Research,        Primary Funding Source: AHRQ
The University of North Carolina at Chapel Hill, 725 Martin
Luther King Jr. Blvd., Chapel Hill, NC 27599; Tel: (919) 966-
7123; Email: pspain@schsr.unc.edu
Research Objective: Screening rates for prostate cancer are
lower in black men than in white men. Studies suggest that
physician trust and access to care, two factors for which there
are also documented racial disparities, may drive cancer
screening rates. This study investigated how race, physician
●Geographic Access to Primary Care Physicians Accepting               ●Location Decisions of Mammography Facilities:
Medicaid: Evidence of a Racial Disparity in Rural South               Racial/Ethnic Disparities Across Zip-Codes
Carolina                                                              Jonathan Sunshine, Ph.D., Mythreyi Bhargavan, Ph.D.
John Stewart, MS, M.P.H., Ana Lopez De Fede, Ph.D.
                                                                      Presented By: Jonathan Sunshine, Ph.D., Senior Director of
Presented By: John Stewart, MS, M.P.H., GIS Manager,                  Research, Research, American College of Radiology, 1891
Institute for Families in Society, University of South Carolina,      Preston White Dr, Reston, VA 20191; Tel: 703-648-8924; Fax:
1600 Hampton St, Columbia, SC 29208; Tel: (803) 777-5516;             703-264-2443; Email: mbhargavan@acr.org
Fax: (803) 777-1120; Email: jstewart@gwm.sc.edu                       Research Objective: Among women age 40 and older,
Research Objective: Serious health disparities exist between          minority women have been found to have mammograms at a
African Americans and whites in South Carolina. Travel                lower rate than white women. One potential cause may be
distance to affordable primary care may be one barrier to good        that minority women live in neighborhoods with no
health, especially for low income, rural residents of the state.      mammography facilities in proximity to them, or because
This study evaluates travel distance to primary care physicians       facilities in their neighborhoods are inadequately committed
accepting Medicaid for residents of African American rural            to patient service. The objective of this study is to measure if
communities versus all other rural communities in South               high percentage of minority population influences facility
Carolina.                                                             location decisions, above and beyond socio-economic factors,
Study Design: Address records for physicians providing                in terms of two sets of outcomes: (a) whether mammography
primary care to South Carolina Medicaid recipients in 2003 (n         facilities locate in or near the zip-code, and (b) commitment
= 3,372) were geocoded (geographically located) using                 to patient service of mammography facilities in it.
MapMarkerPlus, an automated geocoding software package.               Study Design: The unit of analysis is the zip-code, referred to
Rural communities were defined as county subdivisions                 here as the target zip-code. Target zip-codes are categorized
(census civil divisions) in South Carolina with 100% rural            as high-white, high-black, high-Hispanic, and high-Asian based
populations (n = 155; Census 2000, SF3). Rural subdivisions           on the race and ethnicity of the largest sub-population in the
were grouped into two classes, African American (with a               zip-code. Quantitative outcomes of interest for the target zip-
population composition > 75% African American; n = 11) and            code are (i) presence of a mammography facility, (ii) the
all others (n = 144). Using a geographic information system           number of facilities per 100,000 women age 40 and older,
(GIS), actual travel distance (not straight-line distance) was        and (iii) distance to nearest facility if the target zip-code has
measured from the population center of each rural                     none. The explanatory variables are socio-economic
subdivision to the office of the nearest primary care physician       characteristics of the target and surrounding zip-codes.
accepting Medicaid. Differences in travel distance for                Outcome (i) will be estimated using simple logistic regression
residents of African American rural communities versus all            and (ii) and (iii) with linear and log-linear regression. In
other rural communities were evaluated by comparing the               addition, logistic regression is used to compare
mean distance traveled by both groups.                                mammography facilities in high-minority and high-white zip-
Population Studied: Residents of all rural county                     codes in terms qualitative factors, such as whether the facility
subdivisions in South Carolina (2000).                                accepts Medicaid and charity patients, language skills and
Principal Findings: For rural South Carolina residents, travel        minority composition of front office and technologist staff,
distance to the nearest Medicaid primary care provider ranges         types of outreach activities, etc. Data on demographic and
from less than one mile to 28 miles. On average, residents of         socio-economic characteristics of zip-codes are obtained from
African American rural communities must travel farther for            the US Census. The list of mammography facilities and their
Medicaid primary care than residents of all other rural               zip-codes is obtained from the American College of Radiology
communities (11.1 miles versus 7.2 miles; p < .01).                   (ACR) and Food and Drug Administration (FDA) databases of
Conclusions: Particularly in the absence of local pubic               accredited mammography facilities. Patient-service
transportation, travel distance to affordable primary care may        commitment was measured using a pilot phone survey of 105
be a barrier for residents of predominantly African American          mammography facilities.
rural communities.                                                    Population Studied: Women age 40 and older in each zip-
Implications for Policy, Delivery, or Practice: Initiatives to        code
reduce health disparities in the rural South must evaluate            Principal Findings: We find no large differences in the
geographic access to affordable care. Health care policy              minority composition of the population (31% vs. 30% minority
makers must consider supporting the survival of health care           respectively) between zip codes with and without
facilities intrinsic rural areas to increase access to primary care   mammography facilities. However, controlling for zip-code
services. An initiative that sustains and expands the number          socio-economics and population density, high black and
of physicians serving rural African American communities              Hispanic zip-codes are much less likely to have a
continues to be priority.                                             mammography facility in them than predominantly white zip-
Primary Funding Source: Other Government Agency                       codes (for example, 10% increase in the black percentage of
                                                                      the population decreases the odds of a facility in the zip-code
                                                                      by 14%). Facilities in high minority zip-codes seem to be
                                                                      more committed to patient service than those in high-white
                                                                      areas, for example, one-half of the facilities surveyed in high-
                                                                      Hispanic and three-fifths of those in high-black zip-codes use
                                                                      posters and flyers for outreach into the community to educate
women about mammograms, whereas less than one-third of            alcohol use (OR=2.44, p<0.01)], and/or lifetime experience
the facilities in the high-white zip-codes do.                    with IPV (psychological aggression, physical assault, sexual
Conclusions: High-minority zip-codes are less likely to have      coercion) (OR: 3.16, p<0.01). Prevalence and odds of MH/SU
mammography facilities, but facilities in high-minority zip-      service use was also greater for mothers who were not
codes may be more committed to patient care than those in         married, had been unemployed in the past 12 months, lived
high-white zip-codes.                                             below the federal poverty level, had public vs. private
Implications for Policy, Delivery, or Practice: Financial         insurance, and had low levels of social support. There was no
incentives to mammography facilities to locate in high-           significant difference between AR and NAR groups (p=0.10).
minority neighborhoods are likely to be an effective tool in      Racial/ethnic group differences in MH/SU utilization remain
reducing racial/ethnic disparities in access to mammography.      after controlling for these variables.
Primary Funding Source: No Funding                                Conclusions: To our knowledge, little MH/SU utilization
                                                                  research has focused on API (especially NHOPI) populations
●Prevalence & Factors Related To Utilization of Mental            in the U.S. The study results strongly indicate that API
Health/Substance Use Services among Asian Pacific                 women, compared with Whites, are less likely to have ever
Islander Women in the Hawaii Healthy Start Program                received needed MH/SU services.
Van Ta, M.P.H., BS, BA, Hee-soon Juon, Ph.D., Andrea C.           Implications for Policy, Delivery, or Practice: Future
Gielen, ScD, ScM, CHES, Anne Duggan, ScD, Donald                  research should examine the barriers and facilitators to
Steinwachs, Ph.D., MS                                             accessing needed treatments for these disorders. The role of
                                                                  culture should be investigated in influencing an API’s woman
Presented By: Van Ta, M.P.H., BS, BA, Doctoral Candidate,         decision to seek treatment. These results highlight the urgent
Health Policy and Management, Johns Hopkins University,           need to address a priority area – mental health disparities in
5424 W. 2nd St., Santa Ana, CA 92703; Tel: (714) 623-6308;        racial/ethnic minority populations.
Email: vta@jhsph.edu                                              Primary Funding Source: No Funding
Research Objective: The study goal is to elucidate our
understanding of the prevalence of and factors for utilization    ●Diabetic Screening Rates Among Native Hawaiians,
of Mental Health/Substance Use (MH/SU) services among             Filipinos, and Whites in Hawaii
Asian Pacific Islanders (API) women, particularly those with      Deborah Taira, ScD, Komal Warty, BS, Deborah A. Taira, ScD
specific risk factors: depressive symptoms, substance use,
and/or experience with intimate partner violence (IPV).           Presented By: Deborah Taira, ScD, Research Manager, Care
Specific aims for the paper are to: 1) Determine the prevalence   Management, HMSA (BCBS of Hawaii), 818 Keeaumoku
of MH/SU service utilization by race, other socio-demographic     Street, Honolulu, HI 96808-0860; Tel: 8089485337; Fax: 808-
factors, and risk factors; 2) Compare the odds of MH/SU           948-6043; Email: deb_taira@hmsa.com
service utilization by race; 3) Assess the impact of socio-       Research Objective: To examine rates of screenings (retinal
demographic and risk factors on MH/SU service utilization.        eye exam, lipid levels, HA1C) for Native Hawaiians, Filipinos,
Study Design: Data were derived from a randomized trial of        and whites with diabetes. To determine the extent to which
Hawaii’s “Healthy Start Program” (“HSP”), a home visiting         any existing disparities are reduced when we control for
program targeted to families at risk for maltreatment of their    income level and other patient characteristics.
newborns. “At-risk” (AR) families were enrolled in 1994-1995,     Study Design: Retrospective observational study. Self-
randomized to experimental and control groups, and followed       reported ethnicity information from member surveys was
until the children were in Grade 3. When the children were in     merged with administrative (claims) data using unique
Grade 1, an additional comparison group of “not at risk”          identifiers. Multivariable logistic regression analyses were
(NAR) families was recruited. The analyses reported here          used to determine whether there were differences in each of
relate to those who participated in Grade 1 follow-up.            the three screenings related to race or ethnicity. These models
Descriptive statistics were calculated to estimate population     adjusted for age, gender, ACG morbidity level, region,
parameters. Bivariate analysis, exploratory analysis of           household income and type of coverage (HMO, PPO,
potential effect modification and multiple logistic regression    Medicare cost contract).
were performed. The binary outcome is “Ever utilized MH/SU        Population Studied: All diabetic patients enrolled in a large
services” (Yes/No).                                               health plan in Hawaii who had diabetes in 2004 and were
Population Studied: At Grade 1, 709 mothers participated in       Native Hawaiian (n=2866), Filipino (n=2017), or white
the study (491 AR and 218 NAR). It consisted of 97 White, 218     (n=1759).
Asian, 191 Native Hawaiian and Other Pacific Islander             Principal Findings: Native Hawaiians were younger on
(NHOPI) and 203 No Primary Affiliation/Other (NPA).               average [62 years, SD(13)] than Filipinos [65 years, SD(13)] or
Principal Findings: Overall, 151 (21.3%) of the mothers           whites [66 years, SD(13)]. Filipino and Native Hawaiian
reported ever using MH/SU services. MH/SU service use             patients were more likely than whites to be female (56% for
differed among racial/ethnic groups: White (19.2%), Asian         Filipinos and Native Hawaiians vs 45% for whites). Morbidity
(22.5%), NHOPI (23.2%) and NPA (35.1%). Bivariate logistic        levels were highest for whites, with 37% have a morbidity level
regression revealed that compared with White, the odds ratio      of 5 compared to 27% of Native Hawaiians and 25% of
(OR) of MH/SU service utilization for Asian was 0.43              Filipinos. Census tract income level was approximately $3000
(p<0.01), NHOPI was 0.53 (p<0.05) and NPA was 0.83                higher for whites than for Filipinos and Native Hawaiians.
(p=0.49). MH/SU service utilization was significantly             Native Hawaiians were more likely to be enrolled in the PPO
increased by depressive symptoms (OR=2.56, p<0.01),               (70%) than whites (62%) and Filipinos (61%). After
substance use [illicit drug use (OR=3.25, p<0.01) and problem     adjustment, Native Hawaiians were less likely to receive A1C
screening than whites [OR 0.76, 95%CI (0.7,0.9)], while            age, hours worked, income, marital status, physical disability,
Filipino screening rates were similar to those of whites. Both     smoking status, self-reported general health status, and
Filipinos [OR 1.3, 95%CI (1.1,1.5)]and Native Hawaiians [OR        having a chronic health condition all were controlled for.
1.2, 95%CI (1.0,1.3)] were more likely than whites to receive      Population Studied: A population-based, random, telephone-
retinal eye exams and lipid level screening [Native Hawaiian       survey sample of adults in L.A. County who provided accurate
OR 1.2, 9%%CI (1.0,1.3) and Filipino OR 1.5 95%CI (1.3,1.8)].      cross street data so they could be assigned census tracts
Unadjusted results were similar except that differences            (n=12,901).
between whites and Hawaiians for retinal eye exams and lipid       Principal Findings: Racial/ethnic disparities in both types of
level testing did not reach statistical significance.              PA existed. Asian Pacific Islanders (APIs) were half as likely as
Conclusions: After adjustmet, while Native Hawaiians have          Whites to meet the PA guidelines and almost half as likely
significantly lower rates of A1C screening than whites, they       (OR=0.6) to be at least moderately PA. There also was a trend
have higher rates of screenings for lipid levels and retinal eye   for Latinos (p=0.06) to be more likely than Whites to be at
exams. Filipinos screenings rates are similar to whites for        least moderately PA(OR=1.2). The disparities for both PA
A1C, and higher for lipid levels and retinal eye exams.            measures were stronger in some neighborhoods than others.
Implications for Policy, Delivery, or Practice: Native             When residential density and individuals’ perceptions of
Hawaiians are known to have a high prevalence of diabetes. It      neighborhood safety were simultaneiously included in the
may be that outreach efforts to Native Hawaiians and Filipinos     model predicting guideline-meeting PA, the disparities
are improving their quality of care; however, additional           disappeared. Similarly, when percentage of park space was in
targeting interventions may be needed to improve A1C               the model predicting moderate or better PA, the disparities
screening among Native Hawaiians.                                  disappeared. Residents of neighborhoods with higher
Primary Funding Source: Grant number P20MD00173 from               residential densities were more likely to meet the PA
the National Institutes of Health, National Center on Minority     guidelines and those perceiving their neighborhoods to be
Health and Health Disparities                                      unsafe were less likely. Residents of neighborhoods with more
                                                                   park space were more likely to be least moderately PA.
●Neighborhood-Level Correlates of Racial/Ethnic                    Conclusions: Neighborhood characteristics may partially
Disparities in Physical Activity                                   explain racial/ethnic disparities in PA.
Stephanie Taylor, Ph.D., Elizabeth Malcolm, M.D., John             Implications for Policy, Delivery, or Practice: Improving
Romley, Ph.D.                                                      neighborhood safety through increased police, street lighting
                                                                   or community walking groups may increase the likelihood of
Presented By: Stephanie Taylor, Ph.D., Assoc. Social Scientist,    pleasure walking or exercising. Also, further research should
RAND, PO Box 2138, M5S, Santa Monica, CA 90407; Tel: 310-          be conducted to understand what it is about living in densely
393-0411 x7965; Fax: ; Email: staylor@rand.org                     packed housing areas that is linked to increased likelihoods of
Research Objective: Racial/ethnic disparities exist for            being PA. It could be that these urban areas offer more
physical activity (PA), an important disease-preventive            opportunities for utilitarian or social walking.
behavior. Opportunities for PA vary across geographic              Primary Funding Source: National Inst Environmental
regions such as neighborhoods. However, only one study has         Health Sciences
examined a few neighborhood characteristics as correlates of
racial/ethnic disparities in PA. Furthermore, when examining       ●Neighborhood and Patient Level Effects on Response to
PA, most examine only exercise-related PA, not PA that             a Cancer Care Survey
includes all bodily movement. This paper addresses two             Diana Tisnado, Ph.D., Jennifer Malin, M.D., Ph.D., May Tao,
questions: 1) do racial/ethnic disparities vary geographically     M.D., Patti Ganz, M.D., Katherine Kahn, M.D.
and, if so, 2) which of twelve neighborhood characteristics
correlate with that geographic variation, controlling for          Presented By: Diana Tisnado, Ph.D., Adjunct Assistant
individuals’ characteristics? That is, are racial/ethnic           Professor, Division of General Internal Medicine and Health
disparities in PA stronger in some neighborhoods and weaker        Services Research, UCLA, 911 Broxton Plaza, Box 951736, Los
in others, and what might explain that?                            Angeles, CA 90095; Tel: (310)794-0711; Fax: (310)794-0732;
Study Design: PA is defined two ways: 1) meeting Health            Email: dtisnado@mednet.ucla.edu
People 2010 guidelines versus not and 2) being at least            Research Objective: It is understood that patient
moderately active versus not. Respondents reported the             demographic characteristics influence survey response, and
number of minutes and days in a typical week that they             characteristics associated with non-response are often also
vigorously or moderately exercise or performed vigorous            correlated with important health outcomes, raising concerns
activities for at least ten minutes without stopping. We used      about non-response bias. There is also increasing interest in
individual-level data from the most recent (2001 and 2003)         the effects of neighborhood characteristics on health. We
Los Angeles (L.A.) County Health Surveys. Neighborhoods            evaluate the marginal impact of neighborhood characteristics
were operationalized as census tracts (n=1,630). Eleven            on the likelihood of survey response among breast cancer
neighborhood variables were examined: two measures of land         patients in Los Angeles County.
use, two measures of street connectivity, residential density,     Study Design: We enriched existing Los Angeles County
perceived safety, number of arrests, percentage of land space      Cancer Surveillance Program (CSP) data for women
that is park space, racial/ethnic composition, racial/ethnic       diagnosed with breast cancer in 2000 with 2000 Census data,
diversity, mean household income. Cross-sectional, random          and data from the California Office of Statewide Health
effects (multi-level) analyses were conducted of individuals       Planning and Development (OSHPD) on hospital
nested within neighborhoods. Individuals’ gender, education,       characteristics thought to influence quality of care. Patient
demographics (age, race/ethnicity), census tract and primary        ●The Language Spoken at Home and Disparities in
hospital were derived from CSP data, and cancer stage from          Medical and Oral Health, Access to Care, and Use of
registry. Using census data, we identified Poverty Areas            Services in US Children
(>=20% of population below the federal poverty level);              Sandra Tomany-Korman, MS, Glenn Flores, M.D.
Undereducated Areas (>25% of the adult population lacks
high school degree); obtained percentages of nonwhites,             Presented By: Sandra Tomany-Korman, MS, Statistician,
unemployed, foreign-born, and limited English proficiency,          Pediatrics, Medical College of Wisconsin, 8701 Watertown
and linked these to individuals by census tract. Percentages        Plank Rd, Milwaukee, WI 53074; Tel: 414-456-4193; Fax: 414-
were categorized into quartiles. Using OSHPD data, we               456-6385; Email: stomany@mcw.edu
identified hospital characteristics: size, ownership (Public,       Research Objective: 47 million Americans speak a non-
Non-profit, For-profit), teaching (High, Low, or None), and         English primary language at home (NEPL), but little is known
hospital breast cancer volume, using a count of study patients      about health disparities for NEPL children. Our objective was
seen at each hospital. We defined High Volume as >= 75th            to identify disparities for NEPL children in medical and oral
percentile. We also identified American College of Surgeons         health, access to care, and use of services.
(ACOS)-approved cancer programs. In all, 2005 women had             Study Design: The National Survey of Childhood Health was
both complete census tract and hospital information. We             a phone survey in 2003-2004 of a national random sample of
tested the impact on likelihood of CATI response of patient,        parents of 102,353 children 0-17 years old. Disparities in
neighborhood and hospital characteristics using unadjusted          medical and oral health and healthcare were examined for
and adjusted logistic regression, testing for interactions          NEPL children compared to children with English as the
between race/ethnicity and neighborhood characteristics.            primary language at home (EPL). Multivariate analyses were
Analyses were adjusted for patient clustering within hospital.      performed to adjust for race/ethnicity, age, insurance
Population Studied: A population-based sample of women              coverage, income, parental education and employment, and
diagnosed with incident breast cancer in Los Angeles County         number of children and adults in the household.
in 2000.                                                            Population Studied: A national random sample of parents of
Principal Findings: Unadjusted analyses show older age,             102,353 children 0-17 years old interviewed in the National
black, Hispanic, and Other race/ethnicity (versus white), more      Survey of Childhood Health, a phone survey in conducted in
advanced stage (versus ductal carcinoma in situ), low volume        2003-2004.
primary hospital, and all neighborhood characteristics were         Principal Findings: Thirteen percent of US children
significantly, negatively associated with CATI response.            (equivalent to 9.2 million) live in NEPL households; NEPL
Logistic regression results controlling for patient demographic     children are significantly (P<.05) more likely than EPL children
characteristics and cancer stage indicate response was              to be Latino (87%) or Asian/Pacific Islander (7%), poor (42 vs.
significantly, negatively associated with increasing age            13%), and to have unemployed and non-high-school graduate
(O.R.=0.96, p<0.001), black, Hispanic, and Other                    parents. Significantly higher proportions of NEPL children are
race/ethnicity (p<0.001 for all), and stage III cancer (O.R.        uninsured (27 vs. 6%), sporadically insured (20 vs. 10%), not
0.49, p< 0.001). Controlling for hospital characteristics did not   in excellent/very good health (43 vs. 12%), overweight/at risk
change these effects. Low hospital volume was negatively            for overweight (48 vs. 39%), and have teeth in fair/poor
associated with response (O.R.=0.77, p<0.01). The addition of       condition (27 vs. 7%). NEPL children more often have no
neighborhood-level characteristics to the model yielded a           usual source of care (USC; 38 vs. 13%), made no medical (27
significant likelihood ratio test (p<0.05). Poverty area was        vs. 12%) or dental (14 vs. 6%) visits in the past year, and had
significantly, negatively associated with response (O.R.=0.79,      problems getting specialty care (40 vs. 23%), but less often
p<0.001). The p-value of Hispanic ethnicity increased (0.04 to      receive mental healthcare (4 vs. 7%) and are able to get help
0.06), but effects and significance levels of other individual      from USCs via phone (68 vs. 81%). NEPL children more often
and hospital-level characteristics persisted. No interactions       did not receive needed medical care due to cost (43 vs. 26%)
were significant.                                                   or no insurance (59 vs. 39%) and did not receive needed
Conclusions: The potential effects on response rates of             dental care due to health plan problems (20 vs. 8%). Latino
patient and neighborhood characteristics should be                  and Asian NEPL children had several unique disparities
considered in survey design and implementation.                     compared to white NEPL children. Almost all disparities
Implications for Policy, Delivery, or Practice: Individuals of      persisted in multivariate analyses; for example, NEPL children
lower socioeconomic status who may be less likely to                were significantly more likely than EPL children to not be in
participate in survey research are the very individuals at risk     excellent/very good health (OR=2.7 [95% CI, 2.3-3.1]), to have
for experiencing health disparities. Understanding the              teeth in suboptimal condition (OR=1.9 [1.6, 2.1]), to have no
determinants of participation is important to understanding         USC (OR=1.7 [1.4-1.9]), and to have problems getting specialty
the limitations of existing data and to guide improvements in       care (OR=1.7 [1.2-2.3]).
future data collection efforts.                                     Conclusions: Compared to EPL children, NEPL children
Primary Funding Source: NCI, Department of Defense                  experience multiple disparities in medical and oral health,
Breast Cancer Research Program                                      access to care, and use of services.
                                                                    Implications for Policy, Delivery, or Practice: The study
                                                                    findings suggest that disparities for NEPL children in medical
                                                                    and oral health, access to care, and use of services could be
                                                                    reduced or eliminated by decreasing language barriers in the
                                                                    healthcare of NEPL children. One effective policy measure for
                                                                    addressing these disparities and language barriers would be
nationwide third-party reimbursement for medical interpreter       Primary Funding Source: CWF, Robert Wood Johnson
services.                                                          Foundation
Primary Funding Source: No Funding
                                                                   ●Racial/Ethnic (R/E) Disparities in Medical and Oral
●Parents Helping Parents Fight Asthma: A Successful                Health, Access to Care, and Use of Services in US Children
Intervention for Educating Urban Minority Parents of               Sandra Tomany-Korman, MS, Glenn Flores, M.D.
Asthmatic Children on How to Help Other Urban Minority
Parents with Asthmatic Children                                    Presented By: Sandra Tomany-Korman, MS, Statistician,
Sandra Tomany-Korman, MS, Glenn Flores, M.D., Christina            Pediatrics, Medical College of Wisconsin, 8701 Watertown
Snowden-Bridon, Laurie Smrz, Rhonda Duerst                         Plank Rd, Milwaukee, WI 53226; Tel: 414-456-4193; Fax: 414-
                                                                   456-6385; Email: stomany@mcw.edu
Presented By: Sandra Tomany-Korman, MS, Statistician,              Research Objective: Not enough is known about the national
Pediatrics, Medical College of Wisconsin, 8701 Watertown           prevalence of racial and ethnic (R/E) disparities in children's
Plank Rd, Milwaukee, WI 53226; Tel: 414-456-4193; Fax: 414-        medical and dental care. The study objective was to examine
456-6385; Email: stomany@mcw.edu                                   R/E disparities in medical and oral health and healthcare in a
Research Objective: Urban minority children have the highest       national sample.
rates of ED visits, hospitalizations, and mortality due to         Study Design: The National Survey of Childhood Health was
asthma. Inadequate parental education is associated with           a phone survey in 2003-2004 of a national random sample of
preventable pediatric asthma hospitalizations. Effective           parents of 102,353 children 0-17 years old. Disparities in
asthma education programs are needed targeting urban               medical and oral health and healthcare were examined for
minority parents. Our objective was to evaluate whether a          non-Latino white (W), non-Latino African-American (AA),
comprehensive, hands-on curriculum is effective in educating       Latino (L), Asian/Pacific-Islander (API), Native American
urban minority parents about childhood asthma.                     (NA), and multiracial (MR) children. Multivariate analyses
Study Design: As part of a larger study on parent mentors          were performed to adjust for primary language at home, age,
(PMs) and managing childhood asthma, we recruited urban            insurance coverage, income, parental education and
African-American (AA) and Latino (L) mothers of asthmatic          employment, and number of children and adults in the
children from a pediatric asthma clinic to serve as PMs for        household.
other urban minority parents with asthmatic children. Using a      Population Studied: A national random sample of parents of
structured curriculum, PMs participated in two-day training        102,353 children 0-17 years old interviewed in the National
sessions led by an asthma nurse specialist and program             Survey of Childhood Health, a phone survey conducted in
coordinator. Each participant received a training manual;          2003-2004.
curriculum topics included triggers, medications, asthma           Principal Findings: Many significant (P<.05) R/E disparities
pathophysiology, peak-flow meters, written care plans, and         were noted; for example, uninsurance rates were 6% for Ws,
preventing hospitalizations. An emphasis was placed on             21% for Ls, 15% for NAs, 7% for AAs, and 4% for APIs, and
hands-on learning; sessions included identifying triggers in a     the proportions with a usual source of care (USC) were: Ws,
mock apartment. A 34-item test was administered pre- and           90%; NAs, 61%; Ls, 68%, AAs, 77%; and APIs, 87%. The
post-training to assess PM knowledge.                              following disparities persisted for 1 or more minority groups in
Population Studied: Nine parent mentors that completed             multivariate analyses: increased odds of suboptimal health
two days of asthma education training sessions.                    status, overweight, asthma, activity limitations,
Principal Findings: The mean age of the 9 PMs completing           behavioral/speech problems, emotional difficulties,
training was 36 years; all were female and high school             suboptimal dental health, uninsurance, no medical or dental
graduates, 33% were married, 44% were poor, 33% were               visit in the past year, ED visits, unmet medical and dental
unemployed, 67% were AA and 33% were L. The mean (SD)              needs, transportation barriers to care, not receiving
pre-test score was 77.7% (5.3%), and mean post test score          prescription medications, not receiving mental healthcare, no
was 89.9% (5.0%), representing a statistically significant score   USC, USC communication problems, and problems getting
increase (P<.0001). PMs had the greatest score                     specialty care. For example, compared with Ws, the ORs (95%
improvements on questions on controller medications,               CI) of having teeth in less than excellent condition were: APIs,
whether asthma is curable, and preventing hospitalizations,        1.8 (1.3-1.6); AAs, 1.6 (1.4-1.7); Ls, 1.5 (1.4-1.7); and NAs, 1.4
and the least improvement on a question on how albuterol           (1.1-1.8). Certain disparities were particularly marked for
works. PMs were very satisfied with the sessions,                  specific R/E groups: for AAs, asthma, behavior problems, skin
commenting, "I learned more in these couple days of training       disorders, speech problems, and unmet prescription needs;
than I have their whole lives with asthma," and "I learned         for NAs, hearing/vision problems, ED visits, no USC, and
quite a bit: how to breathe in so that all the medicine is taken   unmet medical and dental needs; and for APIs, inferior oral
in, why roaches are a trigger for asthma, etc."                    health, not seeing a doctor in the past year, and problems
Conclusions: A comprehensive, hands-on curriculum is               getting specialty care. MR children also experienced many
effective in educating urban minority parents about childhood      disparities.
asthma, and results in high levels of parental satisfaction.       Conclusions: Minority children experience multiple disparities
Implications for Policy, Delivery, or Practice: Lack of            in medical and oral health, access to care, and use of services.
adequate parental education about childhood asthma is              Certain disparities are particularly marked for specific R/E
known to be associated with preventable pediatric asthma           groups, and MR children experience many disparities.
hospitalizations. This curriculum may prove useful in              Implications for Policy, Delivery, or Practice: The study
educating urban minority parents about childhood asthma.           findings suggest that greater attention needs to be paid to R/E
disparities in US children's medical and dental care.               increasing numbers of elders and fewer formal and informal
Healthcare providers and policymakers particularly should be        providers. Increased hospitalization among low-income
aware of the unique but often overlooked disparities for Native     individuals may be indicative of gaps in health insurance and
American, Asian/Pacific Islander, and multiracial children.         lifelong health disparities. Older women are especially
Primary Funding Source: No Funding                                  vulnerable due to their tendency to live longer, alone and in
                                                                    poverty. Primary Funding Source: T32 predoctoral fellowship
●Relationships Between Rurality, Home Healthcare                    and a grant from the Center for Community Based Care
Service use and Patient Outcomes                                    Research Objective: Determine if degree of rurality and home
Catherine Vanderboom, Ph.D., RN, Elizabeth Madigan, Ph.D.,          healthcare service use influence patient outcomes, specifically
RN                                                                  for hospitalization, use of emergent care, and ability to
                                                                    ambulate. The direct effects of rurality and indirect effects
Presented By: Catherine Vanderboom, Ph.D., RN, Tel: 216-            through home healthcare service use (visits per patient) on
445-5690; Email: cev@cwru.edu                                       patient outcomes were examined in the context of agency and
Research Objective: Determine if degree of rurality and home        community characteristics. Methods: Patient outcomes
healthcare service use influence patient outcomes, specifically     (hospitalization, emergent care, and functional status for
for hospitalization, use of emergent care, and ability to           ambulation) were obtained the Home Health Compare
ambulate. The direct effects of rurality and indirect effects       website, a CMS database of risk-adjusted publicly reported
through home healthcare service use (visits per patient) on         patient outcomes from all Medicare certified home healthcare
patient outcomes were examined in the context of agency and         agencies. Home healthcare service use data (annual number
community characteristics. Methods: Patient outcomes                of home visits and number of patients) were obtained from
(hospitalization, emergent care, and functional status for          Medicare cost reports. Degree of rurality was determined
ambulation) were obtained the Home Health Compare                   using the Rural-Urban Continuum Codes (RUCC) aggregated
website, a CMS database of risk-adjusted publicly reported          into four categories -- urban, suburban, town, and rural.
patient outcomes from all Medicare certified home healthcare        Community income data (mean household income and
agencies. Home healthcare service use data (annual number           poverty level) were obtained from the U.S. Census.
of home visits and number of patients) were obtained from           Population Studied: A stratified random sample of 241 home
Medicare cost reports. Degree of rurality was determined            healthcare agencies was selected from the Home Health
using the Rural-Urban Continuum Codes (RUCC) aggregated             Compare website. Agency data were linked with RUCC, cost
into four categories -- urban, suburban, town, and rural.           report, and census data. Principal Findings: Path analysis
Community income data (mean household income and                    was used to examine the relationships in the model. Initial
poverty level) were obtained from the U.S. Census.                  regression analysis indicated that hospitalization was the only
Population Studied: A stratified random sample of 241 home          outcome associated with rurality, agency characteristics or
healthcare agencies was selected from the Home Health               home healthcare service use. For hospitalization, the overall
Compare website. Agency data were linked with RUCC, cost            equation was significant with 15% explained variance. There
report, and census data. Principal Findings: Path analysis          was no difference in hospitalization based on degree of
was used to examine the relationships in the model. Initial         rurality. Approximately 26-29% of patients were hospitalized
regression analysis indicated that hospitalization was the only     within all degrees of rurality. Increased visits per patient and
outcome associated with rurality, agency characteristics or         low-income community status were associated with increased
home healthcare service use. For hospitalization, the overall       hospitalization across all geographic groups. Rural home
equation was significant with 15% explained variance. There         agencies were smaller, nonprofit, hospital-operated, and
was no difference in hospitalization based on degree of             provided fewer visits per patient than more urban agencies.
rurality. Approximately 26-29% of patients were hospitalized        Four variables (rurality, ownership type, agency size, and
within all degrees of rurality. Increased visits per patient and    hospital operated) influenced visits per patient and, indirectly,
low-income community status were associated with increased          hospitalization. Conclusions: Increased home visits and low-
hospitalization across all geographic groups. Rural home            income status were associated with increased hospitalization
agencies were smaller, nonprofit, hospital-operated, and            across all degrees of rurality. Prior differences in outcomes
provided fewer visits per patient than more urban agencies.         based on rurality are no longer apparent. Although rurality did
Four variables (rurality, ownership type, agency size, and          not have a direct effect on hospitalization, increased rurality
hospital operated) influenced visits per patient and, indirectly,   was associated with decreased home healthcare service use.
hospitalization. Conclusions: Increased home visits and low-        Implications for Research, Policy or Practice: Increased
income status were associated with increased hospitalization        numbers of home visits associated with hospitalization
across all degrees of rurality. Prior differences in outcomes       suggests that higher levels of care may be needed to identify,
based on rurality are no longer apparent. Although rurality did     stabilize, and triage patients to the appropriate care setting.
not have a direct effect on hospitalization, increased rurality     Rural agencies continue to be challenged to maintain patient
was associated with decreased home healthcare service use.          outcomes while providing fewer services. Fewer home visits
Implications for Research, Policy or Practice: Increased            in rural areas constitutes a long term concern associated with
numbers of home visits associated with hospitalization              increasing numbers of elders and fewer formal and informal
suggests that higher levels of care may be needed to identify,      providers. Increased hospitalization among low-income
stabilize, and triage patients to the appropriate care setting.     individuals may be indicative of gaps in health insurance and
Rural agencies continue to be challenged to maintain patient        lifelong health disparities. Older women are especially
outcomes while providing fewer services. Fewer home visits          vulnerable due to their tendency to live longer, alone and in
in rural areas constitutes a long term concern associated with
poverty. Primary Funding Source: T32 predoctoral fellowship         ●Racial Disparities in Quality of Preventive Care at
and a grant from the Center for Community Based Care                VAMCs, Contract CBOC, and VA-staffed CBOC
Research Objective: Determine if degree of rurality and home        Song Wang, M.S., Chuan-Fen Liu, Ph.D., M.P.H., Michael K.
healthcare service use influence patient outcomes, specifically     Chapko, Ph.D., Mark W. Perkins, PharmD, Matthew L.
for hospitalization, use of emergent care, and ability to           Maciejewski, Ph.D.
ambulate. The direct effects of rurality and indirect effects
through home healthcare service use (visits per patient) on         Presented By: Song Wang, M.S., Research Assistant, Health
patient outcomes were examined in the context of agency and         Services Research & Development Center of Excellence,
community characteristics.                                          Northwest Center for Outcomes Research in Older Adults, VA
Study Design: Patient outcomes (hospitalization, emergent           Puget Sound Health Care System HSR&D, 1100 Olive Way,
care, and functional status for ambulation) were obtained the       Suite 1400, Seattle, WA 98101; Tel: (206) 277-4161; Fax: (206)
Home Health Compare website, a CMS database of risk-                764-2935; Email: Song.Wang@va.gov
adjusted publicly reported patient outcomes from all Medicare       Research Objective: Racial disparities in the quality of care
certified home healthcare agencies. Home healthcare service         for patients with chronic conditions have been found in health
use data (annual number of home visits and number of                care systems. This study examines whether there exist racial
patients) were obtained from Medicare cost reports. Degree of       differences in the quality of preventive care for VA primary care
rurality was determined using the Rural-Urban Continuum             patients with diabetes or chronic obstructive pulmonary
Codes (RUCC) aggregated into four categories -- urban,              disease (COPD).
suburban, town, and rural. Community income data (mean              Study Design: Veterans’ race (e.g., African American,
household income and poverty level) were obtained from the          Hispanic, or Caucasian), ICD-9-identified chronic condition
U.S. Census.                                                        (e.g., diabetes or COPD), and CPT-4-identified preventive care
Population Studied: A stratified random sample of 241 home          (e.g., dilated eye exam or influenza vaccination) were obtained
healthcare agencies was selected from the Home Health               from VA administrative data. We stratified the analysis by
Compare website. Agency data were linked with RUCC, cost            patients receiving primary care in different delivery systems,
report, and census data.                                            including VA medical centers (VAMCs) and community-based
Principal Findings: Path analysis was used to examine the           outpatient clinics (CBOCs) whose mission is primary care
relationships in the model. Initial regression analysis indicated   provision. For CBOC patients, we further stratified the analysis
that hospitalization was the only outcome associated with           by Contract and VA-staffed CBOCs. Racial disparities were
rurality, agency characteristics or home healthcare service use.    measured by the relative risk of receiving influenza
For hospitalization, the overall equation was significant with      vaccinations and dilated eye exams in a year, based on
15% explained variance. There was no difference in                  generalized linear models with a Huber-White adjustment for
hospitalization based on degree of rurality. Approximately 26-      heteroskedasticity.
29% of patients were hospitalized within all degrees of rurality.   Population Studied: Our sample included 11,181 veterans
Increased visits per patient and low-income community status        with diabetes and 6,074 veterans with COPD that were seen in
were associated with increased hospitalization across all           72 parent VAMCs, 32 Contract CBOCs, and 76 VA-staffed
geographic groups. Rural home agencies were smaller,                CBOCs from 10/1/99 to 10/1/01.
nonprofit, hospital-operated, and provided fewer visits per         Principal Findings: Overall, 26% of COPD population
patient than more urban agencies. Four variables (rurality,         received at least one influenza vaccination in fiscal year (FY)
ownership type, agency size, and hospital operated)                 2000, and the proportion increased to 32% in FY 2001. 19% of
influenced visits per patient and, indirectly, hospitalization.     diabetes population received at least one dilated eye exam in
                                                                    FY 2000, and the proportion increased to 21% in FY 2001.
Conclusions: Increased home visits and low-income status            Among diabetic VAMC patients, African-Americans were less
were associated with increased hospitalization across all           likely to receive dilated eye exam than Caucasians (p<0.0001).
degrees of rurality. Prior differences in outcomes based on         In VA-staffed CBOCs, African-American veterans with diabetes
rurality are no longer apparent. Although rurality did not have     were more likely to receive dilated eye exam than Caucasians
a direct effect on hospitalization, increased rurality was          (p<0.01). Hispanic veterans with diabetes had comparable
associated with decreased home healthcare service use.              probabilities of receiving dilated eye exam comparing to
Implications for Policy, Delivery, or Practice: Increased           Caucasians in either VAMCs or VA-Staffed CBOCs. Racial
numbers of home visits associated with hospitalization              differences were not significant among veterans with COPD in
suggests that higher levels of care may be needed to identify,      receiving influenza vaccination in VAMCs or CBOCs.
stabilize, and triage patients to the appropriate care setting.     Conclusions: Racial disparities in the provision of diabetes-
Rural agencies continue to be challenged to maintain patient        related preventive care exist in VAMCs, but VA-staffed CBOCs
outcomes while providing fewer services. Fewer home visits          provide dilated eye exam to a higher proportion of African-
in rural areas constitutes a long term concern associated with      American veterans with diabetes. However, preventive care for
increasing numbers of elders and fewer formal and informal          veterans with diabetes or COPD is still provided at lower levels
providers. Increased hospitalization among low-income               than is recommended in VA and ADA guidelines.
individuals may be indicative of gaps in health insurance and       Implications for Policy, Delivery, or Practice: Given the
lifelong health disparities. Older women are especially             inadequate level of preventive care to veterans with diabetes
vulnerable due to their tendency to live longer, alone and in       provided in VA’s delivery systems (e.g., dilated eye exam), VA
poverty.                                                            must examine the reasons for the quality deficit and develop
Primary Funding Source: T32 predoctoral fellowship and a            intervention strategies to improve diabetic preventive care. It
grant from the Center for Community Based Care                      may be useful to understand whether clinical practice style,
staffing, information technology or organizational culture in      Implications for Policy, Delivery, or Practice: When the
VA-staffed CBOCs are responsible for their equitable quality of    rates of changes in health expenditures and self-perceived
care standards, which will provide pointers for improvement        health status do not differ across racial and ethnic groups, it
for VAMCs.                                                         seems that rates of return on health care dollars do not differ
Primary Funding Source: VA                                         across racial and ethnic groups in the way that health care
                                                                   dollars are spent in the years between 1996 and 2001
●Comparison of Changes in Health Care Expenditure and              (inclusive). However, the health status is lower among
Self-Perceived Health Status from 1996 to 2001 across              minority populations. This research suggests that the only
Racial and Ethnic Groups in the United States                      way that the health status of the minority populations will be
Junling Wang, Ph.D., C. Daniel Mullins, Ph.D., Ilene H.            compatible with that of the majority population is if there is a
Zuckerman, Pharm.D., Ph.D., Shelley I. White-Means, Ph.D.,         change in the way that health care dollars are spent. We need
Meghan A. Hufstader, B.A., Genene A. Walker, M.S.                  to either increase the efficiency of way the health care dollars
                                                                   are spent among minority populations or increase the total
Presented By: Junling Wang, Ph.D., Assistant Professor,            amount of health care dollars spent on these populations.
Pharmaceutical Sciences, University of Tennessee, 847              Primary Funding Source: No Funding
Monroe Ave., Room 205R, Memphis, TN 38163; Tel: (901)
448-3601; Fax: (901) 448-4731; Email: jwang26@utmem.edu            ●Why do we need to be Concerned about the Quality of
Research Objective: Previous literature reported that (1)          Rade Code in VA and Medicare Data?
minority populations have lower health status than non-            Iris Wei, Dr.P.H., Beth A. Virnig, Ph.D., M.P.H., Robert O.
Hispanic whites, and (2) minority populations also have spent      Morgan, Ph.D.
fewer health care dollars. The objective of this study is to
explore whether the returns on health care dollars are different   Presented By: Iris Wei, Dr.P.H., Michael E. DeBakey VA
across racial and ethnic groups in the period between 1996         Medical Center, 2002 Holcombe Blvd. (152), Houston, TX
and 2001. If the returns on health care dollars are higher in      77030; Tel: (713) 794-8648; Email: wei.iris@med.va.gov
producing health among minority populations, it would              Research Objective: Medicare administrative and claims data
suggest possibly narrowing disparities in health status.           maintained by the Centers for Medicare and Medicaid Services
Study Design: This is a non-experimental study using               (CMS) provide a rich source of information for examining
secondary database.                                                racial and ethnic disparities in health care use. Similarly, the
Population Studied: The study population was a nationally          National Patient Care Database (NPCD) maintained by the
representative sample of non-institutionalized U.S. civilians in   Veterans’ Health Administration (VA) Office of Information
the Medical Expenditure Panel Survey (1996 and 2001). The          also have been used extensively by researchers and
racial and ethnic disparities were examined separately by          policymakers for assessing equity among VA-users. However,
comparing non-Hispanic whites to non-Hispanic blacks and           there have been concerns that the race code in the Medicare
Hispanic whites, respectively (Hispanic blacks were excluded       data drastically under-estimate Hispanics (less then 40%
from the analysis because of small sample size). The changes       could be identified) and that the race code in the VA NPCD
in self-perceived health status from 1996 to 2001 were             are in poor quality (approximately 40% are missing). Here,
compared across racial and ethnic groups using an ordered          we demonstrate a method of using a Spanish surname match
logistic regression. The changes in health expenditure were        to augment the race codes and evaluate its effectiveness in
compared using classical linear regression (a two-part model       improving the accuracy of race/ethnicity identification and
was also employed to confirm the results of classical linear       reducing bias in interpreting the Medicare and VA data.
model). In the regression models, we controlled for age,
gender, census regions, metropolitan statistical area, poverty     Study Design: Data was obtained by a mail survey from
categories, educational levels, and health insurance status.       Medicare-enrolled male veterans residing in six large U.S.
Principal Findings: The study sample included 31,258 non-          metropolitan areas in 2002. We conducted a Spanish
Hispanic whites, 7,349 non-Hispanic blacks, 11,248 Hispanic        surname match to the Medicare data before drawing our race-
whites. From the descriptive analysis, we found that               and ethnicity-specific survey sample (black, Hispanic, and
compared with non-Hispanic whites, non-Hispanic blacks             white). Using three types of race code (self-report, Medicare,
were 22.98% more likely to report an increase in self-perceived    and surname-match), we compared the racial/ethnic
health status category (p=0.016) from 1996 to 2001; Hispanic       representation in terms of socio-demographic characteristics,
whites are 2.81% more likely to report an increase in self-        health status, health conditions, and health care utilizations
perceived health status (P=0.732); non-Hispanic blacks had         among 2,086 Medicare-enrolled male veterans.
$132 lower increase in health expenditure (P=0.608) and            Population Studied: Medicare-enrolled male veterans aged
Hispanic whites had $556 lower increase in health expenditure      >=65 who resided in six large U.S. metropolitan areas in six
from 1996 to 2001 (P=0.002). However, the rates of changes         states (CA, CO, FL, IL, NY, and TX) in 2002.
in health expenditures across racial and ethnic groups from        Principal Findings: The Spanish surname matching
1996 to 2001 were not significant after controlling for            substantially improved the identification of Hispanics and
confounders. Neither were rates of changes in health status        whites compared to the Medicare race code alone--using self-
across racial and ethnic groups.                                   reported race/ethnicity identification as the gold standard.
Conclusions: Rates of changes from 1996 to 2001 in health          The surname matching did not show as much effect for blacks
expenditures and self-perceived health status did not differ       because the Medicare race code is more accurate for blacks.
across racial and ethnic groups.                                   In general, the surname-matched race code depicted close
                                                                   representation of veterans similar to the self-reported race
codes. By using the race code in Medicare data, we found           • Asian or Pacific Islander • Black or African American (non-
over- or under-estimations of several individual characteristics   Hispanic) • Black or African American (Hispanic) • Hispanic
(e.g., geographic distribution, health conditions)--particularly   or Latino • White
among Hispanics. These discrepancies were largely corrected        Principal Findings: This initiative is still in progress. Baseline
by using the surname-matched race code.                            and follow-up data have been collected from one participant in
Conclusions: Our findings demonstrate that by using the race       the CHCS disparities initiative to test the Disparity Index. The
code in Medicare data, studies could produce biased                results from the test data show that a reduction in disparities
estimates. This is likely to be true for studies using VA race     does not necessarily correspond to an improvement in quality.
codes considering the substantial amount of missing race           Health Plan A has shown a reduction in disparities from
codes in the VA data. Using a Spanish surname match to             baseline as well as a decrease in the overall quality
augment the Medicare or VA race codes, one can generate            component. It is important to note that these results do not
more unbiased study results and interpretations.                   cover the entire project period. Health Plan A: Disparity
Implications for Policy, Delivery, or Practice: The use of a       Index, Baseline=13.7% average difference (across 5 HEDIS
surname match to supplement the Medicare or VA race codes          measures) Disparity Index, Follow-up=4.9% Quality
can greatly enhance researchers’ and policymakers’ ability to      Component, Baseline=43.1% average difference (across 5
examine health care equity.                                        HEDIS measures) Quality Component, Follow-up=37.3%
Primary Funding Source: VA                                         Conclusions: Collection of baseline Disparity Index data from
                                                                   the CHCS disparity initiative is proceeding and expected to
●Health Care Disparity Measurement in a Medicaid                   provide comparable baseline and follow-up measurements
Managed Care Workgroup                                             over several clinical areas.
Karl Weimer, M.P.H., Constance Martin, MHA, David Nerenz,          Implications for Policy, Delivery, or Practice: Health plans
Ph.D., Jane Deane Clark, Ph.D.                                     will be able to use the Disparity Index to target future disparity
                                                                   reduction efforts. A Disparity Index will inform participants
Presented By: Karl Weimer, M.P.H., Senior Research Officer,        about an overall pattern of disparities at the plan level, and, if
Measurement & Evaluation, Center for Health Care Strategies,       used in multiple time periods, whether the disparities
PO Box 3469, Princeton, NJ 08543; Tel: 609-895-8610; Fax:          experienced by racial/ethnic groups are increasing or
609-895-9648; Email: kweimer@chcs.org                              decreasing.
Research Objective: While numerous health plans survey             Primary Funding Source: RWJF
members to assess disparities in care to guide quality
improvement activities, plan efforts are rarely focused on         ●The Cultural Universal of Concern over Status:
reducing disparities across the entire health plan.                Applications to Disparities and other Health Services
Standardizing disparity and related quality of care measures       Issues
will ensure that managed care providers, plans, and states         W Pete Welch, Ph.D., Michael Finch, Ph.D.
have a clear picture of how an organization is reducing
disparities and improving quality. It will also pinpoint where     Presented By: W Pete Welch, Ph.D., 5642 N 20th St.,
initiatives are positively impacting ethnic groups as well as      Arlington, VA 22205; Tel: 703-909-3087; Email:
improving the quality of care within specific clinical areas.      w.pete.welch@gmail.com
Study Design: Health plans participating in initiatives led by     Research Objective: The presumption of social science in the
the Center for Health Care Strategies (CHCS) to reduce ethnic      second half of the twentieth century has been that culture can
and racial health disparities are testing a novel approach to      almost completely modify the behavior of humans. However,
measure and quantify disparities in care and quality across        there are important behavioral patterns that are found in all
several clinical areas and ethnic groups. To date, 12              cultures, such as the concern over status (e.g., disliking
organizations have submitted baseline data on nine HEDIS           humiliation). This paper attempts to ground a number of
measures across the following clinical areas: adolescent           health services phenomena in this cultural universal. In doing
immunization, asthma, diabetes, and prenatal care. The             so, it helps to provide a theoretical foundation to our field.
Disparities Index measures each health plan’s progress             Study Design: Status-seeking behavior appears to be
toward eliminating disparities by the percent difference           universal across human cultures. In support of this
between each racial and ethnic subgroup rate and the most          anthropological evidence, dominant individuals in most
favorable or “best” subgroup rate for each measure. A              mammalian species have priority for mates and food.
reduction in the percent difference from the “best”                Presumably, dominant individuals are more likely to pass on
racial/ethnic group rate is interpreted as a reduction in          their genes to future generations than nondominant
disparity. The Quality component of the Index provides             individuals. Although the concern over status is universal
similar, plan level information in comparison to the 75th          over cultures, the source of status is not necessarily the same
percentile HEDIS Quality of Care rates for Medicaid health         across cultures.
plans. The overall Disparity Index should not be calculated or     Principal Findings: This concern over status plays a key role
reported without the quality component.                            in apparently unrelated health services issues. Prominent
Population Studied: The measures will be produced for the          among these are the (perceived) shame associated with
total population of one target racial/ethnic subgroup and at       seeking treatment for certain conditions (e.g., mental health
least two additional population subgroups. Plans will have a       and AIDS), the stigma associated with enrollment in Medicaid
minimum sample size for each measure within each                   and other social welfare programs, physician anger over the
racial/ethnic subgroup:• American Indian or Alaska Native          shame of malpractice suits, and the need for providers to be
                                                                   respectful toward patients. The prevalence of these issues
across national cultures will be discussed, as will their            exception was that using the need adjustment, NH other
relevance in understanding disparities in health and health          races had a lower OOP expenditure than NH whites. The
services across cultural groups within national systems.             adjusted difference was approximately $331.
Conclusions: Shame, stigma, and related issues fit into a            Insurance status contributed the most to the disparities in the
theoretical framework as defensible as the economic model of         financial burden of prescription drugs. Having insurance in
the desire for material goods. By understanding the                  addition to Medicare lowered the OOP expenditure and
ubiquitousness of status-related issues, health services             proportion in both the market and the egalitarian models. In
researchers would more quickly recognize the role of status in       particular, having Medicaid had the greatest impact.
novel contexts, both in novel categories of issues and when          Controlling for other variables, have Medicaid lowered the
working internationally.                                             OOP expenditure by $1,054 (utilization adjustment) or $917
Primary Funding Source: No Funding                                   (need adjustment). It also lowered the OOP proportion by
                                                                     38.3% (utilization adjustment) or 38.7% (need adjustment).
●Racial and Ethnic Disparities in the Financial Burden of            Conclusions: The disparities in the financial burden of
Prescription Drugs among older Americans                             prescription drugs between non-Hispanic elderly whites and
Tom Xu, Ph.D., Tyrone Borders, Ph.D.,                                Hispanics may be attributable to differences in utilization
                                                                     patterns. However, whether health disparities contribute to
Presented By: Tom Xu, Ph.D., Assistant Professor, Family &           disparities in the financial burden of prescription drugs
Community Medicine, Texas Tech University Health Sciences            requires studies of specific diseases.
Center, 3601 4th St. MS 8161, Lubbock, TX 79430; Tel:                Primary Funding Source: No Funding
(806)743-6983; Fax: (806)743-1292; Email: Ke.Xu@ttuhsc.edu
Research Objective: To examine racial and ethnic disparities
in the financial burden of prescription drugs among older
Americans using a market and an egalitarian models.
Study Design: A nationally representative data set, the
Medical Expenditure Panel Survey 2002, was used. The
financial burden of prescription drugs was measured by the
out-of-pocket expenditure and proportion. The market
orientation defines disparities as different financial burdens
borne by subpopulations with the same utilization level.
Disparities in financial burden are said to be eliminated when
for the same amount of medical services/products received,
all members of the society pay the same amount. The
egalitarian orientation defines disparities as different financial
burdens borne by subpopulations with the same health level.
That is, patients with the same health level should pay the
same, regardless of the variations of the utilization levels that
may be caused by the health care system. It is clear that the
policy implications of the two orientations are different.
To demonstrate whether racial and ethnic disparities in
financial burden among Rx users would persist after
controlling for other characteristics of the respondents,
multivariate analyses were conducted. In the market model,
in which disparities were evaluated against the utilization
level, the total number of Rx, the average refills per Rx, and the
average quantity per Rx were controlled for in addition to the
other independent variables such as demographics. In the
egalitarian model, where disparities were evaluated against
patients’ health needs, disease dummy variables and the two
SF-12 variables were included. To provide nationally
representative estimates for the elderly subpopulation in the
US in 2002, person weights, primary sampling units and
strata were controlled for in all analyses.
Population Studied: 65 years of age or older
Principal Findings: Raw racial and ethnic disparities were
observed in the bivariate analyses between non-Hispanic
whites and Hispanics in the out-of-pocket (OOP) expenditure
and proportion, and between non-Hispanic whites and non-
Hispanic blacks in the out-of-pocket proportion. After
controlling for either utilization or health and the other
independent variables, significant racial and ethnic disparities
in the financial burden of prescription drugs, either measured
by OOP expenditure or proportion, were not found. The only

				
DOCUMENT INFO