Get documents about "Diabetes Treatment Among VA Patients With Comorbid Serious Mental
Document Sample
Diabetes Treatment Among VA Patients
With Comorbid Serious Mental Illness
Sarah L. Krein, Ph.D., R.N.
C. Raymond Bingham, Ph.D.
John F. McCarthy, Ph.D., M.P.H.
Allison Mitchinson, M.P.H.
Jonathan Payes, M.A.
Marcia Valenstein, M.D., M.S.
I
Objective: Patients with serious mental illnesses, such as schizophrenia, ndividuals with serious mental ill-
bipolar disorder, and other psychoses, may be less likely to receive ad- nesses, such as schizophrenia,
equate care for chronic medical conditions than patients without seri- bipolar disorder, and other psy-
ous mental illness. The quality and outcomes of diabetes care were choses, represent a high-risk patient
compared in an observational study among patients with and without subpopulation, with complex, ongoing
serious mental illness. Methods: National data were studied for 36,546 treatment service needs. Many of
individuals receiving care within the U.S. Department of Veterans Af- these individuals also have comorbid
fairs (VA) health care system. Widely used diabetes quality-of-care medical illnesses and are at increased
measures and intermediate outcomes were compared for patients with risk for poor health outcomes. High
diabetes and serious mental illness and age-matched patients with dia- rates of morbidity and mortality have
betes who did not have a serious mental illness. Patients’ use of health been reported among patients with se-
services was also examined. Results: During fiscal year 1998, patients rious mental illness and have some-
with diabetes and serious mental illness were as likely as patients with times been attributed to health system
diabetes alone to have their hemoglobin A1c, low-density lipoproteins characteristics, including inadequate
(LDL), and cholesterol measured; both groups had comparable A1c, integration of the delivery of psychi-
LDL, and cholesterol values. Patients with diabetes and serious mental atric and nonpsychiatric services. Pa-
illness had more outpatient visits, both primary care and specialty vis- tients with serious mental illness also
its, and made more multiclinic visits, including visits to both primary may be at increased risk for poor out-
care and mental health services on the same day. Conclusions: In the comes because of decreased access to
VA, patients with diabetes and serious mental illness appear to receive providers or poorer health habits and
diabetes care that is comparable with the care that other patients with treatment adherence.
diabetes receive, possibly because of increased levels of contact with The U.S. Department of Veterans
the health system and the VA’s integration of medical and mental health Affairs (VA) health care system pro-
services. (Psychiatric Services 57:1016–1021, 2006) vides the ideal context for examining
characteristics of health services treat-
ment among large samples of patients
with serious mental illness. The VA in-
formation system is regarded as one of
the best in the nation, and VA admin-
istrative data include detailed diagno-
sis and treatment measures. Moreover,
Dr. Krein is affiliated with the Center for Practice Management and Outcomes Re- the VA has transformed itself over the
search, Department of Veterans Affairs Ann Arbor Healthcare System, and with the past decade, substantially enhancing
Department of General Medicine, University of Michigan. Dr. Bingham is with the its integration of services. We evaluat-
Transportation Research Institute, University of Michigan. Dr. McCarthy and Dr. ed the hypothesis that patients with se-
Valenstein are with National Serious Mental Illness Treatment Research and Depart-
rious mental illness are less likely to re-
ment of Veterans Affairs, Ann Arbor, and with the Department of Psychiatry, Univer-
sity of Michigan. Ms. Mitchinson is with the Center for Practice Management and Out-
ceive adequate treatment for nonpsy-
comes Research, Department of Veterans Affairs Ann Arbor Healthcare System. Mr. chiatric illnesses. Specifically, we com-
Payes is with the Department of Psychiatry, Case Western Reserve University, Cleve- pared processes of care for diabetes
land, Ohio. Send correspondence to Dr. Krein at HSR&D Center of Excellence, and intermediate disease outcomes
HSR&D/SMITREC, Department of Veterans Affairs, P.O. Box 130170, Ann Arbor, (hemoglobin A1c levels), which influ-
Michigan 48113-0170 (e-mail, skrein@umich.edu). ence later health outcomes (for exam-
1016 PSYCHIATRIC SERVICES ♦ ps.psychiatryonline.org ♦ July 2006 Vol. 57 No. 7
ple, risk of blindness) among VA pa- Methods Measures
tients with diabetes with and without Study design and setting Indicators of diabetes management.
serious mental illness. Approval for this study was obtained Our selection of diabetes process and
As a chronic condition requiring from the institutional review board at outcome measures for the analysis was
ongoing care and often complex treat- the Ann Arbor VA Healthcare System. based on recommended clinical stan-
ment, diabetes can pose substantial Data were obtained from two VA dards for diabetes management.
management challenges for patients health services registries: the National Specifically, we used the set of meas-
and health care providers. Conditions Psychosis Registry (18) and the ures (both technical process and inter-
associated with reduced functioning, Healthcare Analysis and Information mediate outcome) proposed for dia-
especially mental, cognitive, and Group/QUERI-DM (Quality En- betes care monitoring and quality im-
emotional functioning, may particu- hancement Research Initiative for Di- provement through the Diabetes
larly complicate diabetes treatment abetes Mellitus) diabetes registry (19). Quality Improvement Project (21).
and worsen health and other out- We examined data for patients receiv- The indicators used for this study were
comes. For example, several previous ing care in fiscal year (FY) 1998 (Octo- the percentage of patients who had at
studies have reported that depression ber 1, 1997–September 30, 1998). Se- least one hemoglobin A1c level ob-
is associated with poorer adherence rious mental illness was operationally tained in the past year, the hemoglobin
to dietary and self-management rec- defined as having a diagnosis in one of A1c level and the percentage of pa-
ommendations (1,2), poorer glycemic five ICD-9-CM (20) mental health di- tients with a high-risk hemoglobin A1c
control (3), greater social and voca- agnostic categories: all diagnoses of level (A1c ≥9.5 percent) based on the
tional impairments (4), and increased schizophrenia except latent schizo- last value obtained, the percentage of
health care cost and utilization of phrenia, schizoaffective disorder, patients who had a low-density
services (2,5–8). bipolar disorder, other nonorganic lipoprotein (LDL) cholesterol meas-
Diabetes is a prevalent condition psychoses, paranoid states, and affec- ure in the past year, the LDL level and
among patients with schizophrenia tive psychoses. The National Psychosis the percentage of patients with high-
(9–11), and second-generation an- Registry contains utilization data, in- risk LDL values based on the most re-
tipsychotic medications commonly cluding diagnoses and procedures for cent test using a cutoff of 130 mg/dl,
used to treat psychosis may further in- inpatient stays and outpatient visits. and the percentage of patients who
crease patients’ risk for diabetes The FY 1998 diabetes registry con- had a total cholesterol test (because an
(12–15). Among VA patients who re- tains pharmacy, laboratory, and utiliza- LDL level could not be calculated for
ceived a diagnosis of serious mental ill- tion information for individuals with some patients because of elevated
ness in a given year, one in five also re- diabetes who used VA health services triglyceride levels). For quality meas-
ceived a diagnosis of diabetes. Serious during the fiscal year. Patients with di- ures related to having had a test in the
mental illnesses may cause additional abetes were identified as those who re- past year, higher percentages indicat-
difficulties in diabetes management ceived a prescription for insulin, an oral ed higher quality, whereas for meas-
and make positive outcomes harder to hypoglycemic agent, or blood glucose ures based on test values, such as
achieve. However, two recent studies monitoring supplies or who had at least mean A1c value, lower values or per-
(16,17) found little evidence of poorer one inpatient encounter or two outpa- centages indicated better quality.
quality of diabetes care for patients tient encounters with a diabetes-relat- Inpatient and outpatient utilization.
with a co-occurring serious mental ill- ed diagnosis code (ICD-9-CM codes Inpatient utilization measures includ-
ness. These studies focused only on 250.0–250.9, 357.2, 362.0, and 366.41). ed the number and percentage of pa-
the receipt of recommended care The two registries were used to tients with at least one inpatient stay
processes, such as having a foot senso- identify two mutually exclusive groups and the average length of stay per pa-
ry exam, or on a single intermediate of patients. The first group included tient. Outpatient care was measured
outcome (A1c). In addition, they did 18,273 patients with both diabetes and by the total number of outpatient visits
not explore mechanisms that might ac- a comorbid serious mental illness in and the average number of visits per
count for or promote good-quality FY 1998. The second group consisted patient. An outpatient visit was de-
care for these patients. of a randomly selected age-matched fined as all outpatient encounters oc-
In this study, we compared large set of 18,273 patients with diabetes curring on the same day, and each vis-
national samples of patients who had who did not have a diagnosis of serious it was classified according to the types
diabetes and a serious mental illness mental illness. To screen out individu- of clinics visited by the patient: pri-
with patients who had diabetes and als who might have used the VA to re- mary care only; mental health care
no diagnosis of mental illness using ceive medication only, all patients only; specialty care only; or a multi-
several diabetes-specific process needed to have had at least one pri- clinic visit, which was a visit to more
measures and intermediate out- mary care or mental health care visit than one type of clinic on the same
comes. We also examined inpatient during the year to be included in the day—for example, visits to both pri-
and outpatient health care utilization study. Also, because the utilization mary care and mental health care clin-
and patterns of care to better under- part of the study primarily focused on ics. Primary care included general in-
stand how the organization of care ambulatory care, patients with an in- ternal medicine, diabetes clinic, geri-
delivery might influence these proc- patient stay longer than 150 contigu- atrics clinic, and primary care. Mental
esses and outcomes. ous days were excluded. health care included encounters for
PSYCHIATRIC SERVICES ♦ ps.psychiatryonline.org ♦ July 2006 Vol. 57 No. 7 1017
Table 1
Diabetes care processes and intermediate outcomes stratified by treatment regimen for 36,546 Veterans Affairs patients
Diabetes and serious Diabetes only Diabetes and ser-
mental illness (N=18,273) (N=18,273) ious mental illness
versus diabetes only
Treatment and measure N % N % Pooled SE Effect sizea
No hypoglycemic medication 3,605 2,667
HbA1c measured in the past year 2,308 64 1,709 64 .48 .00
HbA1c value (mean±SD)b 6.8±1.68 7.2±1.74 1.72 –.22
HbA1c >9.5 percent 7 10 .28 –.08
LDL measured in past yearc 944 26 728 27 .44 –.03
LDL value (mean±SD mg/dl)d 118±34.58 119±33.53 34.12 –.03
LDL >130 mg/dl 35 37 .48 –.03
Cholesterol measured in the past year 1,694 47 1,253 47 .50 .00
Oral agents only 8,963 9,453
HbA1c measured in the past year 6,213 69 6,123 65 .47 .10
HbA1c value (mean±SD %)b 7.5±1.91 7.8±1.92 1.92 –.15
HbA1c >9.5 percent 13 16 .35 –.07
LDL measured in the past year 3,108 35 3,274 35 .48 .00
LDL value (mean±SD mg/dl)d 118±35.05 118±33.29 34.16 .01
LDL >130 mg/dl 33 32 .47 .02
Cholesterol measured in the past year 5,736 64 5,672 60 .48 .08
Insulin 5,705 6,153
HbA1c measured in the past year 3,960 69 4,005 65 .47 .09
HbA1c value (mean±SD %)b 8.4±2.18 8.5±2.13 2.16 –.09
HbA1c >9.5 percent 25 27 .44 –.03
LDL measured in the past year 1,935 34 1,907 31 .47 .06
LDL value (mean±SD mg/dl)d 118±41.55 117±37.12 39.41 .01
LDL >130 mg/dl 33 32 .47 .03
Cholesterol measured in the past year 3,765 66 3,507 57 .49 .18
a Cohen’s d was used to estimated the effect size for differences in independent means, and Cohen’s h was used as an estimate of the effect size for dif-
ferences in independent proportions (22).
b Values range from 3.0 to 22.3, with higher values indicating poorer HbA1c control.
c Low-density lipoprotein. To convert LDL cholesterol to millimoles per liter, multiply by .02586.
d Values range from 47 to 217, with higher values indicating poorer LDL control.
mental illness treatment (individual timates of effect size were used to aid and serious mental illness group) with
and group), substance abuse treat- in data interpretation. Cohen’s d (22) a mean age of 58±12 years. A higher
ment and rehabilitation, intensive case was used to estimate effect sizes for proportion of those in the diabetes-
management, and intensive individual the ANOVA models, and odds ratios only group were married (29 percent
programs. Specialty care included en- (ORs) (23) represented the effect size compared with 21 percent). Among
counters at clinics offering specialized for chi square tests. Small, medium, patients with an identified race or eth-
care, such as dermatology, ophthal- and large effects were .20, .50, and nicity classification, those in the dia-
mology, and neurology. Outpatient en- ≥.80 for Cohen’s d; 1.2, 1.5, and ≥2.0 betes-only group were slightly more
counters that did not fall into one of for positive ORs; and .83, .67, and ≤.5 likely than patients with comorbid se-
the three classes, such as visits for lab- for negative ORs. All comparisons of rious mental illness to be classified as
oratory tests, were disregarded. diabetes care processes and interme- white (69 percent compared with 64
diate outcomes were stratified by percent) and less likely to be classified
Statistical analysis treatment regimen as a proxy for dis- as Hispanic (8 percent compared with
Group comparisons were initially con- ease severity. Specifically, patients 11 percent) or black (22 percent com-
ducted using one-way analysis of vari- were categorized as to whether they pared with 24 percent). Within the di-
ance (ANOVA) and chi square tests. were treated without any hypo- abetes and serious mental illness
Because of the large sizes of the glycemic agents (indicating milder dis- group, 73 percent were diagnosed as
groups being compared in this study, ease), treated with oral hypoglycemic having schizophrenia, 18 percent as
however, conventional tests of statisti- agents only, or treated with insulin (in- having bipolar disorder, and 9 percent
cal significance were not necessarily dicating more severe disease). as having another type of psychosis.
clinically meaningful, as small differ-
ences were often statistically signifi- Results Quality of diabetes care
cant. Thus, instead of using statistical The study sample was predominantly Table 1 presents the comparison of di-
significance as a primary tool in identi- male (97 percent in the diabetes-only abetes care processes and intermedi-
fying potentially important results, es- group and 96 percent in the diabetes ate outcome measures between the
1018 PSYCHIATRIC SERVICES ♦ ps.psychiatryonline.org ♦ July 2006 Vol. 57 No. 7
Table 2
Utilization of inpatient and outpatient health care services among patients with diabetes only and patients with diabetes and
a comorbid serious mental illness
Diabetes only Diabetes and serious
(N=18,273) mental illness(N=18,273)
Type of utilization N % Mean±SD N % Mean±SD
Inpatient care
Inpatient staya 3,225 17.6 6,953 38.1
Length of stay (mean±SD days)b 8.2±11.4 12.0±15.9
Outpatient visits per patient
Primary care only 18,166 99.4c 4.8±3.9 18,175 99.5c 5.8±5.5
Mental health care only 3,617 19.8 2.1±11.6 17,602 96.3c 17.9±35.3
Specialty care only 13,825 75.7 4.2±5.7 14,873 81.4 5.3±7.9
Multiclinic 6,539 35.8 2.2±2.7 12,533 68.6 3.5±5.9
Total visits per patient (mean±SD)d 16.4±18.2 36.7±40.1
Total outpatient visits 299,530 100 670,070 100
Primary care only 76,143 25.4 77,095 11.5
Mental health care only 34,188 11.4 292,379 43.6
Specialty care only 64,655 21.6 71,461 10.7
Multiclinic 14,186 4.7 43,717 6.5
Othere 110,358 36.8 185,418 27.7
a OR=2.8 for between-group comparison with approximate 95 percent CI based on Woolf’s method of 2.67–2.94.
b Cohen’s d=.26 for between-group comparison.
c High percentages are an artifact of sampling criteria that included a high proportion of patients with these types of care.
d d=.62 for between-group comparison.
e Includes visits for ancillary services such as laboratory tests or physical therapy
two patient groups stratified by dia- ness. Otherwise, within each treat- than patients with solely diabetes and
betes treatment regimen. Only one ment stratum, the two groups of pa- received both more inpatient and
difference in the quality-of-care meas- tients had very similar amounts of dia- outpatient services. Patients with dia-
ures and intermediate outcomes be- betes care and levels of control. betes and serious mental illness were
tween patients with diabetes and those 2.8 times more likely than the dia-
with diabetes and serious mental ill- Utilization of inpatient betes-only group to have had an inpa-
ness reached a small effect size. and outpatient care tient stay and also had a slightly
Among patients who did not receive Differences in inpatient and outpa- longer length of stay when hospital-
hypoglycemic medications (who pre- tient health care utilization between ized. Patients with comorbid serious
sumably had less severe diabetes), pa- the two groups are shown in Table 2. mental illness also made far more vis-
tients with only diabetes had slightly Patients in the diabetes and serious its to outpatient care. On average, pa-
higher A1c levels than their counter- mental illness group had more total tients with diabetes and serious men-
parts who also had a serious mental ill- contact with the health care system tal illness had more than double the
Table 3
Multiclinic visits by type for patients with diabetes with and without comorbid serious mental illness
Diabetes only Diabetes and serious
(N=6,539) mental illness (N=12,533)
Type of visit N % N %
Multiclinic outpatient visits per patient 6,539 12,533
Primary care and mental health care 1,094 16.7 6,137 49.0
Primary care and specialty care 4,840 74.0 2,986 23.8
Mental health care and specialty care 520 8.0 2,927 23.3
Primary care, mental health care, and specialty care 85 1.3 483 3.9
All multiclinic outpatient visits 14,186 43,717
Primary care and mental health care 2,756 19.4 19,060 43.6
Primary care and specialty care 9,274 65.4 8,567 19.6
Mental health care and specialty care 1,812 12.8 14,056 32.2
Primary care, mental health care, and specialty care 344 2.4 2,034 4.6
PSYCHIATRIC SERVICES ♦ ps.psychiatryonline.org ♦ July 2006 Vol. 57 No. 7 1019
number of outpatient visits per per- diabetes care over the past several grated health care system that pro-
son compared with diabetes-only pa- years (24–26). Our overall findings of vides comprehensive access to health
tients (36.7 compared with 16.4; Co- comparable diabetes care processes care across the full spectrum of health
hen’s d=.62, a moderate effect size). and A1c values were somewhat unex- needs. In many areas this integration
pected, given prior research that has includes the actual physical integration
Type of outpatient care demonstrated a clear negative effect of of health care services, which allows
Table 2 also shows the types of outpa- concurrent depression on diabetes out- patients with diabetes and comorbid
tient care received by each patient comes (2,3), even though our findings serious mental illness to visit multiple
group. Notably, patients with diabetes are consistent with two recent studies clinics, receive treatment for both con-
and serious mental illness not only had of individuals with diabetes and comor- ditions, and have their prescriptions
more total contact with the health sys- bid serious mental illness (16,17). Al- filled during a single visit day.
tem and more mental health visits than though the causal processes underlying Integration also may be achieved
patients with diabetes alone, they also these results remain unclear, this study through better communication among
had more primary care, medical spe- offers some possible explanations. providers and referrals across special-
cialty, and multiclinic visits than did Patients with diabetes and serious ties. The VA maintains an electronic
patients with diabetes only. On aver- mental illness had much more contact patient record system that contains
age, patients with diabetes and serious with the health care system than pa- both medical and mental health care
mental illness had 1.0 more visit to pri- tients with diabetes alone. Patients records for all patients treated at that
mary care than patients with diabetes with diabetes and serious mental ill- medical center. Seamless access to pa-
only (mean of 5.8 visits compared with ness used more inpatient care than pa- tients’ records gives providers greater
4.8), 1.1 more specialty medical visits tients with diabetes only and had much insight into their patients’ needs and
(mean of 5.3 compared with 4.2), and higher levels of outpatient use, both of potential conflicts between different
1.3 more multiclinic visits (mean of 3.5 which, even if not focused specifically aspects of their care, which creates op-
compared with 2.2). or solely on diabetes treatment, pro- portunities for improving treatment.
The frequency of multiclinic visits vide an increased opportunity to com- In addition, primary care providers
made by the patients with diabetes plete important diabetes care process- can often refer patients for subspecial-
and serious mental illness is especially es. Patients who also had serious men- ty care and mental health providers
striking, with a total of 43,171 multi- tal illness not only had a greater num- can refer their patients for medical
clinic visits, compared with 14,186 for ber of total outpatient visits, they also care without requiring the patient to
the diabetes-only group. On a per-pa- had more primary care and specialty travel to another facility or burdening
tient basis, 69 percent of those in the medical visits than patients with dia- the patient with the stresses of recon-
group with diabetes and serious men- betes only. Patients with diabetes and ciling conflicts that arise from referrals
tal illness had a multiclinic outpatient serious mental illness made, on aver- outside their health care system.
visit, compared with 36 percent in the age, 36.7 outpatient visits compared We note several study limitations.
diabetes-only group, and nearly half of with 16.4 visits for those with diabetes First, this analysis was based on a pop-
the multiclinic visits for patients with only. They had one additional primary ulation of patients who were using the
diabetes and serious mental illness in- care visit, 1.1 additional specialty med- VA health care system. This popula-
volved both primary care and mental ical care visits, and 1.3 more multiclin- tion was primarily male and likely had
health services (Table 3). ic visits during the year compared with a poorer health status than populations
visits for those with diabetes only. in other health care settings (27). Thus
Discussion The substantial number of visits in- these results may not be generalizable
In general, patients with multiple volving multiple types of health care to all populations. Second, with the ex-
chronic conditions, especially diabetes providers is especially noteworthy. In ception of age and type of mental
and comorbid serious mental illness, particular, 69 percent of patients with health condition, this analysis did not
have challenging health care needs and comorbid serious mental illness had take into account specific sociodemo-
subsequently may be more likely to re- visits involving multiple types of clinic graphic factors, such as race, that may
ceive poorer-quality diabetes care and encounters, compared with 27 percent affect diabetes outcomes or utilization.
have poorer outcomes. However, in of those with diabetes only, and nearly However, in a separate analysis we
this large-sample study, diabetes care 50 percent of multiclinic visits for this compared treatment utilization by the
processes and intermediate outcomes group involved encounters with both groups with diabetes with an age-
of patients with diabetes and serious primary care and mental health servic- matched group of patients with serious
mental illness did not differ apprecia- es. These results indicate that patients mental illness alone for each mental
bly from those of patients with diabetes with diabetes and serious mental ill- health diagnosis category and found
alone. Performance on several of the ness not only had more contact with few effects that appeared to be specif-
measures might be considered poor for health care providers but also may ic to a particular type of mental illness.
both groups, as compared with current have received more intensive and inte- Also, among patients with staff-identi-
performance and quality standards in grated outpatient care. Unlike other fied race and ethnicity information,
the VA (24), although it is important to health care systems that might provide modest differences were observed by
recognize that there have been sub- only medical care and contract out for race and ethnicity, with those in the di-
stantial improvements in VA-delivered mental health care, the VA is an inte- abetes-only group more likely than pa-
1020 PSYCHIATRIC SERVICES ♦ ps.psychiatryonline.org ♦ July 2006 Vol. 57 No. 7
tients (ltr). Lancet 1:495, 1989
tients with comorbid serious mental is paramount to help inform the design
illness to be classified as white (69 per- of health care systems, both inside and 12. Erle G, Basso M, Federspil G, et al: Effect of
chlorpromazine on blood glucose and plasma
cent compared with 64 percent) and outside the VA, that best meet the insulin in man. European Journal of Clinical
less likely to be classified as Hispanic needs of patients with multiple com- Pharmacology 11:15–18, 1977
(8 percent compared with 11 percent) plex chronic conditions. 13. Hagg S, Joelsson L, Mjorndal T, et al: Preva-
or black (22 percent compared with 24 lence of diabetes and impaired glucose toler-
Acknowledgments ance in patients treated with clozapine com-
percent). pared with patients treated with conventional
Also, we were unable to assess the This study was supported by grant DIB-98- depot neuroleptic medications. Journal of
use of health services outside the VA 001 from the Department of Veterans Af- Clinical Psychiatry 59:294–299, 1998
system. We used visit-based criteria to fairs (VA), Health Services Research and 14. Lindenmayer JP, Nathan AM, Smith RC: Hy-
Development Service; by the VA National perglycemia associated with the use of atypi-
restrict our analysis to those who had cal antipsychotics. Journal of Clinical Psychia-
used the VA system, but this restriction Serious Mental Illness Treatment Research try 62(suppl 23):30–38, 2001
could not ensure that patients did not and Evaluation Center; and by grant NIH-
15. Sernyak MJ, Leslie DL, Alarcon RD, et al:
also use services outside the VA. Con- 5060 DK-20572 to Michigan Diabetes Re- Association of diabetes mellitus with use of
search and Training. Dr. McCarthy was atypical neuroleptics in the treatment of
sequently, if outside use was more like-
funded by Career Development Award schizophrenia. American Journal of Psychia-
ly among those with diabetes only, then Merit Review Entry Program MRP-03-320 try 159:561–566, 2002
our results may be somewhat overstat- from the VA Health Services Research and 16. Dixon LB, Kreyenbuhl JA, Dickerson FB, et
ed. Nonetheless, even with more out- Development program. The authors thank al: A comparison of type 2 diabetes outcomes
side use among those with diabetes among persons with and without severe men-
the VA Healthcare Analysis and Informa-
tal illnesses. Psychiatric Services 55:892–900,
only, we feel these results are robust in tion Group for their assistance with data 2004
describing the potential benefits of collection. The views expressed in this arti-
17. Desai MM, Rosenheck RA, Druss BG, et al:
medical and mental health services in- cle are the opinions of the authors and not Mental disorders and quality of diabetes care
tegration. Finally, we examined care necessarily the supporting agencies. in the Veterans Health Administration.
American Journal of Psychiatry 159:1584–
processes and intermediate outcomes 1590, 2002
References
but not costs in this study. Achieving
1. Carney RM, Freedland KE, Eisen SA, et al: 18. Blow FC, McCarthy J F, Valenstein M, et al:
comparable outcomes for vulnerable Care for Veterans With Psychosis in the VHA,
Major depression and medication adherence
diabetes patients with serious mental in elderly patients with coronary artery dis- FY02: 4th Annual Psychosis Registry Report,
illness may require more health care ease. Health Psychology 14:88–90, 1995 2003. Available at www.va.gov/annarbor-
hsrd/smitrecdocs/4thannualpsychosisvha.pdf.
resources and greater funding. 2. Ciechanowski PS, Katon WJ, Russo JE, et al: Accessed Nov 22, 2004
The relationship of depressive symptoms to
19. Krein SL, Hayward RA, Pogach L, et al: De-
Conclusions symptom reporting, self-care and glucose
partment of Veterans Affairs’ Quality En-
control in diabetes. General Hospital Psychi-
This study suggests that VA patients hancement Research Initiative for Diabetes
atry 25:246–252, 2003
Mellitus. Medical Care 38:I38–I48, 2000
with diabetes and serious mental ill- 3. Lustman PJ, Anderson RJ, Freedland KE, et 20. International Classification of Diseases, 9th
ness are receiving diabetes care of al: Depression and poor glycemic control: a Revision, Clinical Modification. Los Angeles,
comparable clinical quality and have meta-analytic review of the literature. Dia- PMIC, 1995
betes Care 23:934–942, 2000
similar intermediate outcomes as VA 21. Fleming BB, Greenfield S, Engelgau MM, et
patients with diabetes only. This find- 4. Wells KB, Stewart A, Hays RD, et al: The al: The Diabetes Quality Improvement Pro-
functioning and well-being of depressed pa- ject: moving science into health policy to gain
ing could be due, in part, to individuals’ tients: results from the Medical Outcomes an edge on the diabetes epidemic. Diabetes
with both conditions using consider- Study. JAMA 262:914–919, 1989 Care 24:1815–1820, 2001
ably more health care services, mean- 5. Katon W: The effect of major depression on 22. Cohen, J: A power primer. Psychological Bul-
ing that there is more opportunity for chronic medical illness. Seminars in Clinical letin 112:155–159, 1992
delivering recommended services. Neuropsychiatry 3:82–86, 1998
23. Wickens TD: Multiway Contingency Tables
Likewise, the integration of medical 6. Katon W, Sullivan MD: Depression and Analysis for the Social Sciences. Hillsdale, NJ,
chronic medical illness. Journal of Clinical Erlbaum, 1989
and mental health services, which is an Psychiatry 51(suppl):3–11, 1990
inherent part of the VA health care de- 24. Kerr EA, Gerzoff RB, Krein SL, et al: Dia-
7. Simon GE, VonKorff M, Barlow W: Health betes care quality in the Veterans Affairs
livery system, may also play an impor- care costs of primary care patients with rec- Healthcare System and commercial managed
tant role in ensuring that patients with ognized depression. Archives of General Psy- care: the TRIAD study. Annals of Internal
diabetes and comorbid serious mental chiatry 52:850–856, 1995 Medicine 141:272–281, 2004
illness receive recommended diabetes 8. Unutzer J, Patrick DL, Simon G, et al: De- 25. Jha AK, Perlin JB, Kizer KW, et al: Effect of
pressive symptoms and the cost of health the transformation of the Veterans Affairs
care. Further research in this area is Health Care System on the quality of care.
services in HMO patients aged 65 years and
warranted, however, including longitu- older: a 4-year prospective study. JAMA New England Journal of Medicine 48:2218–
dinal analyses to examine patterns of 277:1618–1623, 1997 2227, 2003
care over time. Adherence to diabetes 9. Dixon L, Weiden P, Delahanty J, et al: Preva- 26. Asch SM, McGlynn EA, Hogan MM, et al:
Quality of care in the Veteran’s Health Ad-
and antipsychotic medications should lence and correlates of diabetes in national
ministration compared with a national sam-
schizophrenia samples. Schizophrenia Bul-
also be compared to determine how letin 26:903–912, 2000 ple. Annals of Internal Medicine 141:938–
much adherence might contribute to 945, 2004
10. Mukherjee S, Decina P, Bocola V, et al: Dia-
better diabetes outcomes as well as the betes mellitus in schizophrenic patients.
27. Agha Z, Lofgren, RP, VanRuiswyk, JV, et al:
interplay between diabetes treatment Are patients at Veterans Affairs Medical Cen-
Comprehensive Psychiatry 37:68–73, 1996
ters sicker? A comparative analysis of health
and mental health treatment. A more 11. Mukherjee S, Schnur DB, Reddy R: Family status and medical resource use. Archives of
thorough understanding of these issues history of type 2 diabetes in schizophrenic pa- Internal Medicine 160:3252–3257, 2000
PSYCHIATRIC SERVICES ♦ ps.psychiatryonline.org ♦ July 2006 Vol. 57 No. 7 1021
