Primary Care and Public Emergency Department Overcrowding

Document Sample
Primary Care and Public Emergency Department Overcrowding Powered By Docstoc
					Primary Care and Public Emergency
Department Overcrowding

Kevin Gnrmbach, MD, Dennis Keane, MPH, and Andrew Bindman, MD
Introduction                                   San Francisco, although there is a rela-
                                               tively extensive "safety net" of hospital-
      Hospital emergency departments           and community-based primary care clin-
play multiple roles in the American health     ics, these facilities have proved insuffi-
care system. Once considered a source of       cient to meet the demand for primary care
care for major injuries and life-threatening   services. For example, appointment wait-
medical conditions, the emergency de-          ing times for patients new to the hospital-
partment has become part primary care          based family practice and general medi-
physician and part social worker to many       cine clinics at San Francisco General
Americans. As early as the 1950s, it was       Hospital average 2 months (San Francisco
noted that the number of emergency de-         General Hospital Outpatient Administra-
partment visits in the United States was       tion audit, unpublished data, July 1990).
rising dramatically and that many of these     Although many of the primary care clinics
visits were for conditions that did not re-    have same-day appointments to accom-
quire emergency treatment.' 2 Analysts         modate the acute care needs of those who
attributed this phenomenon, at least in        are established clinic patients, such ap-
part, to the ascendancy of hospital-based      pointments are often unavailable for pa-
subspecialists and the dwindling founda-       tients without established clinic relation-
tion of community-based general prac-          ships.
titioners in the United States.3 More                The problem of public hospital emer-
recently, overcrowding of hospital emer-       gency department overcrowding invites a
gency departments in the inner city has        number of possible policy responses.
reached desperate proportions.4 We re-         Among these possibilities are augmenting
cently reported the consequences of over-      emergency department resources and/or
crowding at the emergency department at        productivity, expediting transfer to inpa-
San Francisco General Hospital.5 Patients       tient beds for patients requiring hospital-
with noncritical conditions faced waiting       ization, developing urgent care clinics
times as long as 17 hours, and 15% of the       near emergency departments for rapid
patients left without ever seeing a physi-
 cian. When contacted 1 to 2 weeks after
 their emergency department visit, patients    Kevin Grumbach is with the Department of
who left without seeing a physician were       Family and Community Medicine at San Fran-
                                               cisco General Hospital and the Institute for
 twice as likely as patients who did see a     Health Policy Studies at the University of Cal-
 physician to report deterioration of their    ifornia, San Francisco. Andrew Bindman and
 health status.                                Dennis Keane are with the Division of General
       Although use of emergency depart-       Internal Medicine, San Francisco General Hos-
 ments for nonemergency conditions has         pital; Andrew Bindman is also with the Institute
                                               for Health Policy Studies, University of Cali-
 become ubiquitous, this pattern of utiliza-   fornia, San Francisco.
 tion is especially prominent among pa-               Requests for reprints should be sent to
 tients who are poor, non-White, and with-     Kevin Grumbach, MD, Bldg 1, Room 202, San
 out a regular source of primary care.-'"'     Francisco General Hospital, 1001 Potrero Ave,
                                               San Francisco, CA 94110.
 Davidson, in a review published over a               This paper was submitted to the Journal
 decade ago, concluded that "low-income,       December 20, 1991, and accepted with revi-
 inner-city residents tended to use [emer-     sions June 8, 1992.
 gency departments] as substitutes for the            Editor's Note. See related editorial by Gei-
  family doctors they did not have."" In        ger (p 315) in this issue.

                                                                    March 1993, Vol. 83, No. 3
                                                                                                       Emergency Department Overcrwding

treatment of low-acuity problems, institut-      ment survey.5 We surveyed all patients         cause their doctor or clinic told them to go
ing patient cost-sharing, refusing emer-         waiting for emergency care at San Fran-        there or because somebody else chose the
gency department services to patients with       cisco General Hospital during the week of      emergency department for them. Seven-
nonemergency conditions, and allowing            July 9 to July 16, 1990. Patients were eli-    ty-five patients who gave more than one
the emergency department queue itself to         gible for inclusion in the study if theywere   reason for choosing to receive care in an
continue to play a triage role by imposing a     18 years of age or older; spoke English,       emergency department were excluded
high "time price" on patients for use ofthe      Spanish, or Cantonese; were mentally co-       from the analysis of reasons.
emergency department. A different ap-            herent; and were not assigned by a triage           The time of day a patient came for
proach, however, would be to increase ac-        nurse to immediate care. All patients          emergency department care was taken
cess to alternative primary care services        agreeing to participate were provided a        from the triage nurse intake record and was
that offer continuity of care for a full spec-   self-administered survey in the language       categorized as during clinic hours (Monday
trum of acute and chronic care needs-in          of their choice. The survey included ques-     through Friday, 9:00 AM to 5:00 PM) or not
effect, to reverse the trends noted by Da-       tions about the patients' demographics         during clinic hours (all other times).
vidson and substitute family doctors for the     and socioeconomic status, insurance cov-
emergency departments. This strategy             erage, chronic and acute health status,        Appropriateness of Emergency
would require (1) identification of patients     regular source of care, and other factors      Departnment Use
who use the emergency department for             related to use of the emergency depart-              There are no widely accepted vali-
routine health care needs because of bar-        ment. Health status was measured by            dated standards for measuring the clinical
riers to primary care services, (2) timely       means of a chronic disease checklist and       appropriateness of emergency depart-
referral ofthese patients to appointments at     standardized questions about pain and          ment use.12,14,15 Our principal measure of
primary care facilities, and (3) enhance-        general health.13 Survey data were sup-        clinical appropriateness was the acuity
ment of the capacity of the primary care         plemented by information from the triage       score assigned by the emergency depart-
system to accommodate additional indi-           nurse record about the acuity of the pa-       ment triage nurse. As a standard practice
gent patients. Increasing access to primary      tient's condition. Patients were contacted     at the San Francisco General Hospital
care services as an alternative to the emer-     by telephone, by mail, or in person 7 to 14    emergency department, triage nurses use
gency department could potentially reduce        days after their initial emergency depart-     written guidelines to assign each patient
public emergency department overcrowd-           ment visit to complete a follow-up survey      an acuity score. An acuity score of 1 in-
ing, provide indigent patients a less costly     that included questions about health care      dicates a patient with an immediate need
form of care for their immediate needs, and      utilization in the period since the emer-      for care, such as a patient with anterior
establish a regular source of care for those     gency department visit. In addition, we        chest pain consistent with myocardial in-
patients with ongoing health care needs.         reviewed registration or visit logs from all   farction (ineligible for the present study).
      We analyzed data collected in a large      hospitals, every public clinic, and most       An acuity score of 2 indicates a patient
survey of patients waiting for care at the       neighborhood clinics in San Francisco for      with an urgent need for care, such as a
San Francisco General Hospital emer-              study subjects.                               patient with abdominal pain and fever. An
gency department to test whether a policy                                                       acuity score of 3 indicates a patient who
of primary care referral would be clinically     Access                                         needs care within 3 hours, such as a pa-
appropriate for and acceptable to this pop-           We used survey questions about            tient with vaginal bleeding and stable vital
ulation. Our specific objectives were to         health insurance status and income as          signs. An acuity score of 4 indicates a pa-
 determine the extents to which (1) patients     measures of financial access to primary        tient with a nonurgent need for care, such
 rely on the emergency department be-            care. The survey also included questions       as an afebrile patient with a rash. We con-
 cause of lack of alternative sources of         about whether patients had a regular           sidered emergency department use to be
 care; (2) patients using the emergency de-      source of medical care. The questionnaire      clinically appropriate for patients with an
 partment have clinical conditions that do       listed seven reasons why patients might        acuity score of 2, possibly appropriate for
 not require specialized emergency serv-         choose the emergency department as their       patients with a score of 3, and inappropri-
 ices; (3) patients would be willing to use an   source of care. The patients were asked to      ate for patients with a score of 4. We val-
 alternative source of care if one were          select the one reason most important to         idated acuity scores with the outcome of
 available; and (4) patients who already         them, and we subsequently classified pa-       hospitalization (using a x2 contingency ta-
 have a regular source of primary care use       tients into three categories on the basis of   ble) and with patients' subjective ratings
 the emergency department in a more clin-        these reasons. The first category, consist-    of their conditions (using a Spearman's
 ically appropriate manner than do patients      ing of patients who lacked an accessible       rank test). Subjective measures included
 without a regular source of care. Although      alternative, included patients who stated      standardized survey questions about pa-
 a number of studies have investigated one       that they did not know where else to go or     tients' ratings of the seriousness of their
 or two of these elements bearing on emer-       that they did not have insurance to pay for    condition and the amount of pain they
 gency department use and access to pri-         medical care. The second group, consist-       were experiencing,13 as well as a question
 mary care,6-9,12 few have provided a com-       ing of patients who said that they thought     about the duration of their symptoms.
 prehensive and systematic analysis of           the emergency department was the best           Wilingness to Use the Clinic
 these features within a framework that          place to go for their problem, that it was
 could guide policy decisions.                   easy to get to, or that they usually went            When they first registered for emer-
                                                  there for care, was considered to use the     gency department care, patients were
                                                  emergency department because of its at-        asked to answer yes or no to the following
Methods                                           tractive qualities. The third group was        question about a hypothetical care alter-
       We have previously described in de-        made up of patients who were directed to       native: "Sometimes people have to wait
 tail the design of the emergency depart-         the emergency department, either be-           several hours in the emergency room. In-

 March 1993, Vol. 83, No. 3                                                                            American Journal of Public Health   373
Grumbach et al.

                                                                                              male, unemployed, and insured, and to
                                                                                              have more chronic illnesses and worse
                                                                                              general health, than were patients with no
                                                                                              regular source of care.
                                                                                                    When asked why they chose the
                                                                                              emergency department for their care, 45%
                                                                                              of the patients cited access barriers. Un-
                                                                                              insured patients were significantly more
                                                                                              likely to give this reason than were pa-
                                                                                              tients with Medicaid or other insurance
                                                                                              (P < .001 by x2; Table 2). Overall, pa-
                                                                                              tients without a regular source of care
                                                                                              were also more likely to cite access bar-
                                                                                              riers, although when patients were strati-
                                                                                              fied by insurance status this effect was sig-
                                                                                              nificant only among patients with
                                                                                              insurance other than Medicaid.
                                                                                                    Only 13% of the patients surveyed
                                                                                              while waiting for care had conditions that
                                                                                              were definitely clinically appropriate for
                                                                                              emergency department services, as mea-
                                                                                              sured by a nurse-assigned acuity score of
                                                                                              2 (Table 3). One third of the patients were
                                                                                              judged to have nonurgent problems that
                                                                                              were clinically inappropriate for emer-
                                                                                              gency department care (acuity score 4).
                                                                                              Rashes and upper respiratory infections
                                                                                              made up one quarter of these nonurgent
stead ofwaiting now, would you prefer to       emergency department (acuity score 2) vs       problems. (Note that the study excluded
have a doctor's appointment at a definite      questionable or inappropriate use (acuity      patients with an acuity score of 1, who
time in 1 to 3 days?"                          score 3 or 4). Within the model, the inde-     went directly to a treatment area.) Acuity
                                               pendent variables of age, income, number       scores were predictive of hospitalization;
Analysis                                       of chronic illnesses, and general health       likelihood of admission both on the day of
      We compared sociodemographic             score were entered as continuous vari-         the emergency department visit and in the
characteristics, time of emergency depart-     ables, and the variables of sex, race           1- to 2-week follow-up period declined for
ment visit, appropriateness of emergency       (White vs other), employment (part-time        patients with less urgent acuity scores.
department utilization, and patterns of        or full-time vs other), and regular source     Seven patients (3%) with an acuity score
medical care in the 1- to 2-week post-         of care (yes vs no) as dichotomous terms.      of 4 were hospitalized at some time during
emergency department visit period be-          Insurance status was classified as none,       the study period.
tween patients with and without a regular      Medicaid, or other.                                   Patients' own ratings of the severity
source of care and between patients will-                                                      of their condition and their pain indicated
ing and unwilling to trade emergency de-       Result                                          that many were seeking care for relatively
partment care for an appointment. For                                                          routine or chronic problems. One third of
univariate analyses, we used the x2 statis-          Seven hundred patients waiting for        the patients rated their problems as not at
tic for comparisons ofdichotomous or cat-      care (79% of the 882 patients eligible for      all or only a little serious. In addition, 16%
egorical variables, the Mann-Whitney U         the study) agreed to participate in the sur-    stated they had no pain or mild pain. One
test for ordinal variables, and the Stu-       vey. The most common reason for ineli-          third of the patients sought care for prob-
dent's t test for continuous variables. To     gibility was an acuity score of 1, which        lems that had been present for at least a
measure the independent association of a       was assigned to 11% of the patients who         week. Nurse-assigned acuity scores cor-
regular source of care with appropriate-       underwent triage in the emergency depart-       related with patients' pain ratings and du-
ness of emergency department use and           ment. Patients waiting for care in the          ration of symptoms, but not with patients'
reasons for coming to the emergency de-        emergency department were primarily             rating of the seriousness of their problem
partment, we used multivariate logistic re-    poor, unemployed, non-White, young,             (Table 3).
 gression methods and the SPSS pro-            and uninsured (Table 1). Six hundred fifty           Patients assigned an acuity score of 2
 gram,16 controlling for other patient         participants answered the question about        were much more likely to have a regular
 sociodemographic and clinical character-      a regular source of care; two thirds of         source of care than were patients assigned
 istics. In the logistic regression model of   these patients had no regular source. Pa-       scores of 3 or 4. Within the multivariate
                                               tients with a regular source of care were       logistic model, having a regular source of
 appropriateness, we looked for indepen-
 dent predictors of appropriate use of the     significantly more likely to be older, fe-      care remained a significant predictor of

 374 American Journal of Public Health                                                                           March 1993, Vol. 83, No. 3
                                                                                                     Emergency Department Overcwding

more appropriate emergency department
visits, that is, visits for more acute condi-
tions (odds ratio = 2.4; Table 4). No other
variable in the regression model was sig-
nificantly associated with appropriate use
of the emergency department.
Patients' Willingness to Use Clinics
     Overall, 38% of the patients surveyed
expressed a willingness to trade an emer-
gency department visit for a clinic ap-
pointment within 3 days (Table 5). Will-
ingness to accept an appointment at a later
time was not associated with income, em-
ployment status, gender, reason for using
the emergency department, travel time to
the emergency department, or presence of
a regular source of care. Patients without
insurance were slightly more likely to be
willing to trade for an appointment. In
contrast, willingness to trade was strongly
associated with several clinical variables.
Patients who were willing to trade for an
appointment were more likely to have an
acuity score of 4 than were patients un-
willing to trade. Compared with patients
unwilling to trade for an appointment, pa-
tients willing to trade rated their problems
as less serious and had had their problems
for a longer time. Of the patients assigned
an acuity score of 4 who stated a willing-
ness to accept a clinic appointment in lieu
of the emergency department visit, 2 (2%)
were hospitalized at some time in the
study period. One of these patients com-
plained of leg weakness but had no focal
deficits on initial examination by the emer-
gency department physician. A decision
to discharge the patient was reconsidered
after a neurology consultation. The pa-
tient was hospitalized and was found to
have spinal cord impingement from a plas-
macytoma; he underwent a successful op-
eration to remove the tumor. The other
patient had a gastric ulcer, which had
spontaneously stopped bleeding at the
time he was hospitalized; the patient was
discharged the day after admission.
Follow-up Utilization
      Patients with a regular source of care
were significantly more likely than pa-
tients without a regular source of care to      care were equally likely to leave the emer-    for primary care providers. Many patients
visit a clinic in the 1 to 2 weeks after com-   gency room before being seen by a phy-         cited access barriers as their reason for
ing to the emergency department. Of             sician in the face of long waits at the time   seeking care in the emergency depart-
those patients who saw a physician in the       oftheir initial emergency department visit.    ment, had conditions that could be man-
follow-up period, 21% of the patients with-                                                    aged in primary care facilities, and ap-
out a regular source of care had follow-up         iscusion                                    peared willing to accept an alternative
visits consisting exclusively of further                                                       source of care.
emergency department encounters, in                  Our results confirm that many poor              Studies in other emergency depart-
comparison with 8% of patients with a reg-      and uninsured patients in San Francisco        ments have also documented the prob-
ular source of care (P = .04 by x2). Pa-        rely on the San Francisco General Hospi-       lems uninsured and underinsured Ameri-
tients with and without a regular source of     tal emergency department as a substitute       cans face in obtaining routine medical

 March 1993, Vol. 83, No. 3                                                                         American Journal of Public Health 375
Grumbach t ai.

care. A survey conducted at the Univer-        of the emergency department as a default
sity of California Irvine Hospital emer-       choice caused by a lack of altemative
gency department of patients with clinical     sources of care, whereas wealthier pa-
and demographic characteristics similar to     tients may view their medical problems as
those of our study population found that       specifically requiring emergency depart-
33% of the patients had delayed seeking        ment care.
medical care in the previous year and 21%           Hospitalization data validated the
had delayed or been refused care for their     nurse-assigned acuity score and patients'
current medical problem.7 A study of pa-       judgment in being willing to wait for a later
tients with minor illnesses at the George      clinic appointment. Of the 96 patients with
Washington University Hospital emer-           an acuity score of 4 who also stated a will-
gency department in Washington, DC,            ingness to trade for an appointment, only
also found that poor and uninsured pa-         2 required hospitalization during the study
tients were more likely than nonpoor, in-      period. Although both patients had seri-
sured patients to report a lack of a regular   ous medical conditions, it is not clear that
source of care as their reason for using the   the patient with the gastric ulcer was as-
emergency department.8                         signed the correct acuity score or that
      Our results are consistent with those    timely outpatient care rather than care in
of other studies that have found wide-         an emergency department would have al-
spread medically inappropriate use of the      tered either patient's clinical outcome.
emergency department for nonemergency          Treatment of low-acuity patients in pri-
conditions. However, most evaluations of       mary care settings rather than an emer-
the appropriateness of emergency depart-       gency department is also less expensive.17
ment use have not incorporated patients'       It would thus appear to be sound clinical
own views of the severity of their medical     and health policy to redirect patients with     ing longwaits were equally likely to leave
condition. We found that one third of the      low acuity scores and a willingness to use      without being seen and that many pa-
patients themselves considered their           an alternative source of care to more ap-       tients who left had urgent problems that
problem of no or only minor seriousness,       propriate primary care facilities.              subsequently required hospitalization.15
and that more than one third expressed a             How could this redirection be accom-      Emergency department queues do not,
willingness to wait 1 to 3 days for a clinic   plished? One option would be to let the         therefore, appear to be a sufficiently dis-
appointment. In contrast, in a community-      emergency department queue serve a tri-         criminating mechanism for discouraging
wide survey of a more middle-class pop-        age role, as it presently does at San Fran-     inappropriate use.
ulation in Rochester, NY, performed al-        cisco General Hospital. We have previ-               Another option would be to simply
most two decades ago, Stratmann and            ously reported that, faced with long waits      refuse care to patients coming to emer-
Ullman found that 95% of emergency de-         for care, patients with less acute problems     gency departments for clinically inappro-
partment users said that theirproblemwas       were more likely to leave the emergency         priate reasons. The University of Califor-
urgent and required care the same day.2         department without being seen than were        nia Davis Hospital emergency department
Although conditions in urban emergency          patients whowere more seriously ill; how-      implemented such a policy in 1988, deny-
 departments have changed since the             ever, half of the patients who left before     ing services to patients with nonemer-
 1970s, the differences in attitude between     being seen had acuity scores of 2 or 3.5 In    gency conditions and providing these
 the patients at San Francisco General          a similar public hospital emergency de-        patients a list of clinics and private physi-
 Hospital and those in Rochester suggest        partment study, Baker et al. found that        cians accepting new patients.14 Although
 that many indigent patients view their use     patients in different acuity categories fac-   a monitoring program did not detect ob-

376 American Journal of Public Health                                                                            March 1993, Vol. 83, No. 3
                                                                                                          Emergency Department Ovearowding

vious adverse outcomes, such as deaths,           our findings that patients with a regular       ria for refusing care, or implicitly, through
from this triage practice, no follow-up of        source of care used the emergency depart-       long waiting times, without assuring pa-
the patients denied care was performed. It        ment more appropriately and relied less         tients of access to an alternative source of
remains unclear whether patients denied           upon it for follow-up care suggest that es-     care are ethically and clinically unaccept-
care actually received care at other sites        tablishing a regular source of primary care     able. [l
and how the policy affected health out-           may have a sustained effect on altering
comes other than mortality.                       inappropriate patterns of emergency de-
      Our results suggest a slightly different    partment use by indigent patients. Other        Acknowledgments
model of intervention for addressing the          studies, including demonstration projects       Support for this study was provided by the San
demand for public hospital emergency de-          funded by the Robert Wood Johnson               Francisco General Foundation for Medicine.
                                                                                                  Dr. Grumbach was a Pew Health Policy Fellow
partment care by patients who face barri-         Foundation and Medicaid, have also              at the time the study was performed.
ers to primary care. Triage nurse evalua-         found that patients with a regular source of          We wish to thank Sara Levin, Marianna
tion could be used to screen for patients         care use the emergency department more          Caponigro, Alan Jung, Taejoon Ahn, Jesus
whose clinical conditions do not require          appropriately.89,17,22-24 Although direct-      Asenjo, Kristen Wilson, Roberto Diaz, Jose
                                                                                                  Farfan, Lisa Chew, Miguel Gil, Amilcar
emergency department care. Willing pa-            ing some patients to primary care settings      Mayan, Paul Ishimini, Fern Ebeling, RN,
tients could be offered urgent care ap-           should reduce emergency department              Merle Sande, MD, Alan Gelb, MD, Carol Sor-
pointments at primary care clinics in the         waiting times, these gains may be offset if     gen, MA, Ann Stangby, RN, Anita Stewart,
patients' neighborhoods. Because of the           shorter waits attract new patients to the       PhD, Lauren Rauch, Robert Wachter, MD, and
                                                  emergency department.17                         participants of the writing seminar at the Insti-
difficulties patients encounter in schedul-                                                       tute for Health Policy Studies for their contri-
ing timely appointments, it might be nec-               Our policy suggestions are consistent     butions toward this project.
essary for clinics to reserve appointments        with the recommendations of the Ameri-
for emergency department referrals and to         can College of Emergency Physicians for         References
have emergency department staff assign            addressing emergency department over-            1. Shortliffe E, Hamilton T, Nororian E. The
patients a designated appointment slot            crowding.3 There are serious challenges,            emergency room and the changing pattern
                                                                                                      of medical care. N Engl J Med. 1958;258:
rather than simply give patients a clinic         however, to implementing a model pro-               20-25.
telephone number to call. There is increas-       gram of referral from public hospital emer-      2. Stratmann W, Ullman R. A study of con-
ing awareness of the importance of patient        gency departments to primary care pro-              sumer attitudes about health care: the role
preference in medical decision making,18          viders in the current policy context.               of the emergency room. Med Care. 1975;
                                                  Successful referral would require a greater         13:1033-1043.
and our results suggest that patients' pref-                                                       3. Schroeder S. The increasing use of emer-
erences should be considered in the refer-        level of communication and coordination             gency services: why has it occurred? Is it a
ral process. Many patients we surveyed            between public hospital emergency de-               problem? West JMed. 1979:67-69. Edito-
expressed a willingness to receive care at        partments and primary care facilities than          rial.
 a clinic rather than at the emergency de-        currently exists in most communities. Pi-        4. Andrulis D, Kellerman A, Hintz E, Hack-
                                                                                                      man B, Weslowski V. Emergency depart-
partment, and the referral process may            lot programs would need to test whether             ments and crowding in U.S. teaching hos-
prove more successful if the emergency            patients' stated willingness to use alterna-        pitals. Ann Emeig MedL 1991;20:980-986.
 department does not unilaterally make re-        tive sources of care reliably predicts at-       5. Bindman A, Grumbach K, Keane D,
 ferral decisions.                                tendance at scheduled primary care ap-              Rauch L, Luce J. Consequences of queu-
                                                  pointments. Further research could refine           ing for care at a public hospital emergency
      An alternative to referral to primary                                                           department. JAMA. 1991;266:1091-1096.
 care clinics would be development of a            measures for assessing which patients           6. Ullman R, Block J, Stratmann W. An
 hospital-based urgent care clinic that            truly require emergency department serv-           emergency room's patients: their charac-
 could accept new patients with low-acuity         ices. As the long waits for new patient            teristics and utilization of hospital services.
                                                   appointments indicate, public clinics in           Med Care. 1975;13:1011-1021.
 conditions on a drop-in or next-day ap-                                                           7. Pane G, Farner M, Salness K. Health care
 pointment basis. A number of public hos-          cities such as San Francisco may already           access problems of medically indigent
 pitals have such facilities. This type of ser-    be operating at peak capacity and may be           emergency department walk-in patients.
 vice may be simpler to administer and             unable to accommodate new urgent care              Ann Emeg Med. 1991;20:730-733.
 require less patient effort to use. Many          referrals without an infusion of additional     8. ShesserR, KirschT, SmithJ, HirschR. An
                                                                                                      analysis of emergency department use by
 patients may simply want care on an ep-           resources.                                         patients with minor illness. Ann Emerg
 isodic basis and neither need nor desire                Problems of emergency department             Med. 1991;20:743-748.
primary care on a continuity-of-care               use are not limited to uninsured patients in     9. Haddy R, Schmaler M, Epting R. None-
model, although lack of a regular source of        the United States,26 27and the dominant             mergency emergency room use in patients
                                                   factors accounting for use of emergency             with and without primary care physicians.
care is less often voluntary among the                                                                 JlFam Pract. 1987;24:389-392.
poor than among the more affluent.19               departments may differ among settings           10. Schneider K, Dove H. High users of VA
     There are potential advantages, how-          and populations. However, poor and un-              emergency room facilities: are outpatients
ever, to referring emergency departnent            insured patients in the United States are           abusing the system or is the system abusing
                                                   particularly vulnerable in a system that            them? Inquiry. 1983;20:57-64.
patients to facilities offering more compre-                                                       11. Davidson S. Understanding the growth of
hensive primary care services. Studies             does not assure everyone of financial ac-           emergency department utilization. Med
have found that follow-up of diagnostic            cess to basic medical care. Many patients           Care. 1978;16:122-131.
evaluations initiated in emergency depart-          coming to the public hospital emergency        12. Buesching D, Jablonowski A, Vesta E, et
                                                    department may not require emergency               al. Inappropriate emergency department
ments is often inadequate for patients                                                                 visits. Ann Emnerg Med. 1985;14:672-676.
without regular providers,20 and that               services, but almost all have health care      13. Stewart A, Hays R, Ware J. The MOS
many uninsured patients have not re-                needs that deserve medical attention. Pol-         Short-Form general health survey: reliabil-
ceived cancer screening and other health            icies that deny patients emergency depart-         ityandvalidity in a patient population. Med
care maintenance services.2' Moreover,              ment care either explicitly, through crite-        Care. 1988;26:724-732.

 March 1993, Vol. 83, No. 3                                                                              American Journal of Public Health 377
Gnimbach et al.

14. Derlet R, Nishio D. Refusing care to pa-                    19. Hayward R, Bernard A, Freeman H, Co-                                                                                             on emergency room utilization at a com-
    tients who present to an emergency depart-                      rey C. Regular source of ambulatory care                                                                                         munity hospital. Med Care. 1978;16:723-
    ment. Ann Emeng Med 1990;19:262-267.                            and access to health seIvices.AmJPublic                                                                                          729.
15. Baker D, Stevens C, Brook R. Patients                           Health. 1991;81:434-435.                                                                                              24.        Bonham G, Barber G. Use of health care
    who leave a public hospital emergency de-                   20. Brook R, Stevenson RJ. Effectiveness of                                                                                          before and during Citicare. Med Care.
    partment without being seen by a physi-                         patient care in an emergency room. NEngi                                                                                         1987;25:111-119.
    cian. JAMA. 1991;266:1085-1090.                                 JMedL 1970;283:904-907.                                                                                               25.        American College of Emergency Physi-
16. SPSS: Operations Guide. Chicago: SPSS                       21. Woolhandler S, Himmelstein D. Reverse                                                                                            cians. Measures to deal with emergency
    Inc; 1990.                                                      targeting of preventive care due to lack of                                                                                      department overcrowding. Ann Emerg
17. The Munictpal Health Sevwices Prgram                            health insurance. JAMA. 1988;259:2872-                                                                                           Med. 1990;19:944-945.
    Princeton, NJ: The Robert Wood Johnson                          2874.                                                                                                                 26.        Magnusson G. The hospital emergency de-
    Foundation; 1986: special report no. 1.                     22. Hilditch J. Changes in hospital emergency                                                                                        partment as the primary source of medical
18. Barry M, Mulley A, Fowler F, Wennberg                           department use associated with increased                                                                                         care. Scand J Soc Med. 1980;8:149-156.
    J. Watchful waiting vs. immediate tran-                         family physician availability. JlFam Pract.                                                                           27.        Parboosingh E, Larsen D. Factors influ-
    surethral resection for symptomatic                             1980;11:91-96.                                                                                                                   encing frequency and appropriateness of
    prostatism-the importance of patients'                      23. Ullman R, BlockJ, BoatrightN, Stratmann                                                                                          utilization of the emergency room by the
    preferences. JAMA. 1988;259:3010-3017.                          W. Impact of a primary care group practice                                                                                       elderly. Med Care. 1987;25:1139-1147.

                                                                                                                                                                 ...................                             ........

                                                                                                                          .........................                            ...


                                                           ...n                    o
                                                                                1...!       .iJ   A.
                                                                                                                                                                                       Hil"                                                        .... ..............

                  ....                                                                                   ..............                                                                                                                                          ...

                                                          id                                      19:                                                                                         ....


 378 American Joumal of Public Health                                                                                                                                                                                          March 1993, Vol. 83, No. 3