The effects of a community-based partnership, Project
Access Dallas (PAD), on emergency department
utilization and costs among the uninsured
M. DeHaven1, H. Kitzman-Ulrich1, N. Gimpel2, D. Culica3, L. O’Neil1, A. Marcee2,
B. Foster2, M. Biggs4, J. Walton5
1Texas Prevention Institute, School of Public Health, University of North Texas Health Science Center, Ft. Worth, TX 76107, USA
2Department of Family and Community Medicine, The University of Texas Southwestern Medical Center, Dallas, TX 75235, USA
3Texas/Oklahoma AIDS Education & Training Center, Parkland Health and Hospital System, Dallas, TX 75235, USA
4Department of Psychiatry, The University of Texas Southwestern Medical Center, Dallas, TX 75235, USA
5Baylor Health Care System, Dallas, TX 75246, USA
Address correspondence to Mark DeHaven, E-mail: firstname.lastname@example.org
Background Approximately 19% of non-elderly adults are without health insurance. The uninsured frequently lack a source of primary care
and are more likely to use the emergency department (ED) for routine care. Improving access to primary care for the uninsured is one strategy
to reduce ED overutilization and related costs.
Methods A comparison group quasi-experimental design was used to evaluate a broad-based community partnership that provided access to
care for the uninsured—Project Access Dallas (PAD)—on ED utilization and related costs. Eligible uninsured patients seen in the ED were enrolled
in PAD (n ¼ 265) with similar patients not enrolled in PAD (n ¼ 309) serving as controls. Study patients were aged 18–65 years, ,200% of the
federal poverty level and uninsured. Outcome measures include the number of ED visits, hospital days and direct and indirect costs.
Results PAD program enrollees had significantly fewer ED visits (0.93 vs. 1.44; P, 0.01) and fewer inpatient hospital days (0.37 vs. 1.07; P, 0.05)
than controls. Direct hospital costs were ?60% less ($1188 vs. $446; P, 0.01) and indirect costs were 50% less ($313 vs. $692; P, 0.01).
Conclusions A broad-based community partnership program can significantly reduce ED utilization and related costs among the uninsured.
Keywords health services, primary care, public health
National data indicate that nearly 19% of non-elderly indivi-
duals are without health insurance or nearly 49 million
Americans.1Young adults demonstrate higher rates of being
uninsured with 30% of 19–25 year olds, 28.3% of 26–34
year olds and 22.0% of 35–44 year olds as compared with
14.4% of 55–64 year olds.1These rates increase for indivi-
duals with lower incomes to 34.4% for 19–25 year olds,
36.4% for 26–34 year olds and 31.5% for 35–44 year
olds.1In these age brackets, males have higher rates of being
uninsured as compared with their female counterparts.1
Many of the uninsured are in working families with ?60%
having at least one individual with full-time employment.2
Being uninsured is more common among individuals with
lower incomes and education levels, and among ethnic mi-
norities.1Minority and low-income Americans without in-
surance generally lack a regular source of medical care, and
suffer from medical conditions that are either preventable or
M. DeHaven, Professor
H. Kitzman-Ulrich, Assistant Professor
N. Gimpel, Assistant Professor
D. Culica, Director
L. O’Neil, Associate Professor
A. Marcee, Senior Clinical Analyst
B. Foster, Statistician
M. Biggs, Associate Professor
J. Walton, Vice President of Health Equity and Chief Equity Officer
# The Author 2012, Published by Oxford University Press on behalf of Faculty of Public Health. All rights reserved
Journal of Public Health | Vol. 34, No. 4, pp. 577–583 | doi:10.1093/pubmed/fds027 | Advance Access Publication 31 May 2012
by guest on November 22, 2015
easily treated in the outpatient setting.3Consequently, the
uninsured are four times more likely than the insured to
forgo or postpone needed preventive care, three times more
likely to skip recommended tests or treatments, more likely
to be hospitalized for ambulatory care-sensitive conditions,
more likely to be diagnosed with late-stage colorectal cancer,
melanoma, breast cancer and prostate cancer, more seriously
ill upon hospitalization and three times more likely to die in
the hospital.4–7Although insurance alone is not sufficient
for eliminating health outcome disparities, access to care is
necessary for improving health outcomes.3,8
Emergency departments (ED) have become a vital source
of care for those without insurance who generally lack a
source of primary care, since they are required to evaluate
and treat patients regardless of the ability to provide
payment.9,10Consequently, although care delivered through
the ED is frequently for non-urgent problems, it is substan-
tially more costly than comparable care delivered elsewhere
and can produce significant charges for the uninsured.11–13
Additionally, having a regular source of primary care pro-
vides continuity of care and increases the likelihood of
patients’ receiving preventive screening, both of which are
absent in the ED setting.14,15ED overutilization is becom-
ing a significant financial problem for hospitals as well.16
Improving access to care for the uninsured may be an ef-
fective strategy for reducing ED utilization and related costs,
while also providing better management for patients with
chronic disease.12,17–20Current proposals for addressing
lack of insurance in the USA, focus on providing health care
to all through universal coverage, tax credits or an expanded
Medicaid program, or by expanding existing private insur-
ance through increased availability and affordability.2,21
However, generally absent from these deliberations is a dis-
cussion of the existing approaches which have evolved in
local communities for meeting the needs of the uninsured.
Improving access to care offers the additional possibility of
reducing health disparities and increasing health-related
quality of life.16,22–24
Community-level approaches to caring for the uninsured
tend to reflect local needs and desires, and conclusions
about their effectiveness tend to be based on process and
health-related outcomes.25Consequently, little is known
about whether local faith-based or other community pro-
grams can provide health and/or financial benefits to the
individuals and communities they serve.26,27The experience
of local organizations engaged in meeting the health needs
of the uninsured can provide valuable information on man-
aging and delivering care in a cost-effective way.28Project
Access Dallas (PAD) is a community-wide faith-health
partnership that was developed to provide access to care
and preventive services for low-income working individuals
without health insurance residing in Dallas, TX. PAD’s geo-
graphical focus is ‘Central Dallas,’ an area characterized by
high proportions of groups who are likely to use the ED as
a source of care—racial and ethnic minorities, adults with
less than high school education and those with low socio-
economic status.22,29–33PAD was the result of a long-term
partnership between faith-based organizations, government
agencies and social service organizations, hospitals and the
local medical society, several universities and a medical
school. By creating an organizational and administrative in-
frastructure, PAD provides coordinated access for the unin-
sured to existing faith-based community health clinics,
volunteer primary and specialty care physicians, and local
hospitals and pharmacies. The current study was developed
to examine the effects of PAD on reducing ED utilization,
hospital days and associated direct and indirect costs among
patients enrolled in the program. Our hypothesis was that
the patients enrolled in PAD would have significantly fewer
ED visits and hospital days than comparison patients, and
that their direct and indirect costs related to ED utilization
and hospitalization would be significantly less during the
12-month period following enrollment into the project.
Participants and methods
Patients seeking care at the Baylor University Medical Center
ED (located in Central Dallas) between 10 April 2003 and
30 July 2004 were included in the present study. Eligible
study patients were identified by a trained research assistant
(RA) based in the ED during daytime weekday hours and
by using a patient recording system for those seen after
hours and on weekends. Patients were eligible for enrollment
if they resided in the target area zip code, had an income of
,200% of the federal poverty level; were not eligible for
health insurance through the local public hospital system;
were not receiving Medicaid or Medicare and exceeded the
average ED visit rate (1.5 ED visits during the past 12
months). Patients meeting the PAD eligibility criteria were
approached by the RA during daytime hours, and after
hours patients were contacted within 5 days of their ED
visit, until the monthly enrollment quota of 17 patients was
met. Consenting patients were enrolled by the RA on a con-
tinuous basis until the enrollment quota for the month was
reached (intervention group) over the 16-month study
period. All other eligible patients using the ED during the
study period that were not contacted to participate in the
JOURNAL OF PUBLIC HEALTH
by guest on November 22, 2015
study served as a comparison group (control group).
Patients were followed for 1 year based on the month of en-
rollment. The study protocol was approved by both the
Baylor University Medical Center and the UT Southwestern
Medical Center Institutional Review Boards.
Enrolled patients completed an intake interview and
healthrisk assessment (HRA),
community-clinic or hospital-based coordinator, to assess
their general level of health and for assignment to commu-
nity care coordination (CCC) services. Screening for CCC
eligibility is provided for the four primary components of
the care coordination program: (i) identifying and addressing
social concerns, (ii) identifying patients with or at risk of
developing type 2 diabetes, (iii) identifying patients with de-
pression and (iv) providing education and referrals for
After completing the intake interview and HRA, enrolled
patients were assigned to either CCC or self-care based on
their responses to the CCC screening survey. CCC patients
were assigned to a community health worker (CHW) with
whom they were expected to meet monthly; self-help patients
were not assigned to a CHW, but had access to the patient
telephone help line for medical care questions and could
request CHW services as needed. About 20% of PAD
patients were assigned to CCC. During the study period, the
PAD network included 600 physicians, 15 hospitals, 8 com-
munity charity health clinics, 8 ancillary service support
organizations, 1 national laboratory service organization and
.40 000 nationwide pharmacies. Volunteering physicians,
patients for enrollment into the program.
All PAD patients were assigned to a primary care pro-
vider, could be referred to specialty physicians depending on
their medical needs, received $750 a year in pharmacy bene-
fits and were eligible for laboratory tests, ancillary proce-
dures and inpatient hospital care. PAD’s primary goal was to
improve access for the uninsured to a coordinated system of
care, and to begin better addressing the full range of health
needs that affect the uninsured. Participating physicians,
hospitals and ancillary partners determined their level of
participation by agreeing to donate their services to a speci-
fied number of patients per year.
Direct ED costs, indirect ED costs, number of hospital
days, number of ED visits and hospital costs were collected
from an administrative database at enrollment and retro-
spectively 1 year following enrollment.
Direct ED costs: direct ED costs were the sum of the direct
costs of all ED visits for the year following enrollment
and refers to costs associated with the delivery of care
during an ED visit or hospital admission.
Indirect ED costs: indirect ED costs were the sum of the
indirect costs for all ED visits for the year following en-
rollment and includes the fixed costs related to building,
maintenance, staffing and utilities.
Number of hospital days: this variable was the sum of the
number of inpatient hospital days following an ED visit
for the year after enrollment.
Number of ED visits: this variable was the total number of
ED visits duringthe
Health risk assessment: the HRA was collected at baseline
by an RA and screened for Healthy People 2010 Leading
Health Indicators including: physical activity, weight
control, tobacco use, substance abuse, responsible sexual
behavior, mental health, injury and violence and immuni-
zations. These data were collected for determining study
subjects’ general level of health and to guide CHW coun-
seling for health promotion and disease prevention
among CCC patients.
Severity score: all patients received a severity score for each
ED visit, which was based on a subjective assessment
derived through consultation between the attending phys-
ician and the ED nurse—the scores ranged from 0 to 5
with 0 being not serious and 5 being the most serious.
Although not an objective measure of true disease sever-
ity, the scores allow comparison between the two study
groups to assure comparability of severity upon presenta-
tion to the ED.
Simple comparisons of the outcome measures across the
two groups were completed followed by models that
included the demographic variables and their interactions
with the group. Due to the non-normality of the data,
several approaches were examined. Parametric t-tests, the
non-parametric Mann–Whitney U-tests and categorical ana-
lyses involving chi-square and logistic regressions were used.
The logistic models involved a step-wise procedure which
removed the interaction terms sequentially (highest P-value
removed) until all remaining interactions had a P-value of
0.10 or less; any main effects not involved in the remaining
interactions were sequentially removed (highest P-value
by guest on November 22, 2015
removed) until all remaining main effects had a P-value of
During the 1-year study period, a large number of study
subjects and comparison subjects had no additional ED
visits. Having a large number of zero visits created a floor
effect in the data across all the outcome measures with zero
visits resulting in zero costs and hospital days. Since trans-
formations were unable to resolve the floor effect, the actual
values of the outcome variables and dichotomized values that
take into account the floor effect were reported. Direct and
indirect costs were dichotomized at $50 (versus . $50),
while number of ED visits and number of hospital days
were dichotomized at 0 (versus ? 1).
Project Access program participants (n ¼ 265) and controls
(n ¼ 309) were similar on most demographic variables
(Table 1). The vast majority of patients in both groups were
Black, followed by Hispanic and Whites, respectively.
Similarly, most were single, had no health insurance and had
comparable levels of visit severity. Patients with government
or private insurance were sometimes enrolled into the
program when their insurance was about to expire or were
for some other reason unable to access the insurance.
Although females outnumbered males in both groups, sex
was statistically (P ¼ 0.05) different across the two groups;
significantly more females were enrolled in PAD (62.6%)
than in the control group (54.4%).
Compared with controls, PAD participants had signifi-
cantly fewer ED visits (P, 0.01) and hospital days (P,
0.05) and less direct (P, 0.01) and indirect costs (P,
0.01). Table 2 indicates that the average number of ED
visits for enrollees during the year following enrollment
was 0.93 visits, compared with 1.44 visits for control
patients. PAD patients were also less costly to treat and
used significantly less hospital resources than the control
patients. Related to direct costs, treating control patients
was more than two and one-half times ($1188) more
costly than treating enrollees ($446). An interaction effect
was observed between enrollment and sex (P ¼ 0.10); dis-
proportionately fewer males incurred higher direct costs
among PAD enrollees than among the males in the
control group. The same was true for indirect costs, with
PAD patients accounting for less than half the expend-
iture level ($313) of a typical control patient ($692).
Finally, PAD patients spent on average less than one-half
day (0.37) in the hospital compared with more than one
full day (1.07) for controls. When the analyses were
repeated using the dichotomized outcome variables, the
results were identical, with only the level of significance
changing (Table 3).
Table 1 Sample characteristics
(n ¼ 265)
(n ¼ 309)
Age (mean+SD) 35.7 (+12.0)35.0 (+12.1) 0.49
Black182 (68.7%) 229 (74.1%)0.25
Hispanic 55 (20.8%) 48 (15.5%)
White 28 (10.6%)32 (10.4%)
Females 166 (62.6%)168 (54.4%)0.05a
Males 99 (37.4%)141 (45.6%)
Divorced 7 (2.6%)12 (3.9%)0.33
Married 49 (18.5%)60 (19.4%)
Single203 (76.6%) 235 (76.1%)
Widowed6 (2.3%)2 (0.6%)
Government 20 (7.6%) 21 (6.8%)0.26
Private 19 (7.3%) 13 (4.2%)
Self/unknown223 (85.1%)274 (89.0%)
3.25 (+1.42) 3.13 (+1.63) 0.34
Table 2 PAD enrollment status and hospital utilization
Outcome variableEnrollees (n ¼ 265) Controls (n ¼ 309)(df) t-valueP-value
Independent t-testMann-Whitney U-test
Direct costs, mean (SD) $445.6 (1911)$1188 (4625)(423) 2.580.01,0.01
Indirect costs, mean (SD)$313.3 (1335) $692.1 (2646) (470) 2.210.03,0.01
Number of hospital days, mean (SD)0.37 (2.4) 1.07 (5.6) (427) 1.980.050.02
Number of visits, mean (SD)0.93 (1.7)1.44 (2.6)(542) 2.80,01,0.01
JOURNAL OF PUBLIC HEALTH
by guest on November 22, 2015
Project Access Dallas (PAD) is a broad-based community
partnership, created for the purpose of eliminating health
disparities, by developing a healthcare network for the unin-
sured. This 1-year comparison group trial indicated that
PAD reduced the reliance of uninsured patients on using
the ED for their medical care needs and related costs. PAD
participants had significantly fewer ED visits, hospital days,
and less direct and indirect costs than control patients.
What is already known on this topic
A recent review found that case management, the most
described intervention to reduce non-emergent ED use, has
the potential to reduce ED use and improve social and clin-
ical outcomes in non-elderly individuals described as ‘fre-
quent users’ of the ED (between 4 and 10 visits per year).34
Frequent users of the ED tend to have complex medical
issues and suffer from other social and behavioral disorders
such as homelessness, mental illness and chemical depend-
ency. Unlike the current study, these studies specifically tar-
geted frequent users with concurrent social, behavioral and
medical issues. The goal of the PAD program was to
provide access to health care for uninsured, low-income
individuals or the ‘working poor.’ Therefore, the findings of
the present study extend what is known about the utility of
providing access to care in the context of ED utilization to
a broader population.
Similar to the PAD program, other studies have found
that a collaborative relationship between a community-based
clinic and hospital can benefit both the hospital and
patients.35For example, prior to establishing a community-
based clinic/hospital partnership in a low socioeconomic
county in California, 28% of hospital ED visits were for
non-emergent care producing an annual loss of $1 million
to the hospital. The partnership reduced non-emergent care
visits to 9% and reduced the annual loss by $600 000.
Similarly, another study of primarily African American
females with a mean age of 53 years enrolled indigent
patients with one of five chronic conditions into a program
that provided free primary care and disease management
services and found that program enrollees could reduce
their number of yearly ED visits and costs by more than
half.18Another study assessed ED visits in a medically
underserved urban community of individuals primarily aged
18–45 years before and after a community health center
was established.36Over a 3-year period, individuals who had
visited the ED at least once decreased from 22.1 to 13.8%.
Additionally, there was a decrease in individuals who
reported that the ED was ‘the best place to treat their
problem’ and who cited a lack of physician availability as a
reason for visiting the ED. The PAD program reinforces
these findings that access to primary care resources can
reduce the reliance on the ED for non-emergent health
Much of the research on the uninsured has focused on
health-care financing, and how to affordably expand health
insurance coverage or improve access.37However, little re-
search examines how the current health care system can
improve access to care for the uninsured or improve health
outcomes. The PAD design assumes that by providing
health-care resources upstream before individuals develop
serious disease, it is possible to reduce the downstream costs
associated with expensive hospitalizations and acute care ser-
vices. In Texas, for example, the annual downstream cost for
treating uninsured patients is ?$4.1 billion, with the largest
single share of costs being hospital care delivered for acute
illness.38A significant portion of this cost is unreimbursed
with hospitals assuming responsibility. Other parts of the
country which have experimented with the PAD approach
have witnessed reduced costs associated with providing
access to appropriate preventive and primary care.39Our
study reinforces these findings and demonstrates that exist-
ing but underutilized health system capacity (e.g. unused
hospital space, physician appointments and specialty ser-
vices) can significantly reduce health resource utilization and
expenditures. Indeed, the effectiveness of the PAD approach
and the findings of the present study provide critical infor-
mation needed in federal-level discussions related to resolv-
ing the growing problem of uninsurance.40
Table 3 PAD enrollment status and hospital utilization (dichotomized)
(n ¼ 265)
(n ¼ 309)
Direct costs, freq (%)
169 (63.8) 168 (54.4)0.021.48
96 (36.2)141 (45.6)
Indirect costs, freq (%)
180 (67.9)180 (58.3) 0.021.52
85 (32.1) 129 (41.8)
Number of hospital days, freq (%)
None 251 (94.7)279 (90.3)0.05 1.93
One or more 14 (5.3) 30 (9.7)
Number of visits after contact, freq (%)
None156 (58.9) 147 (47.6),0.01 1.58
One or more109 (41.1)162 (52.4)
by guest on November 22, 2015
There are a number of limitations associated with this study.
First, the study is not a randomized clinical trial. Since we
used a comparison group instead of randomization, it is
possible that our intervention and control groups differ on
important unknown or unmeasured characteristics. For
example, since the study enrolled volunteers in the PAD
program while controls were assigned to the comparison
group, it is possible that the PAD enrollees were more moti-
vated to seek better health based on their own predisposi-
tions rather than the PAD program. Although the study
design provides a measure of control and it is unlikely that
patient characteristics alone could account for the findings
observed. Furthermore, this study was restricted to the in-
formation available from secondary data sources. Due to
the large number of presenting complaints and diagnoses, it
was not possible to create meaningful classifications for the
ED visits at enrollment. Another study limitation is the lack
of data on primary care utilization during the PAD interven-
tion, which was beyond the scope of the present study.
What this study adds
PAD provides a model for improving access to care for un-
insured individuals by utilizing a donated-care program that
delivered charity care for medically indigent patients through
a broad-based community partnership. Donated care pro-
grams have received considerable attention and appear to
have a significant impact on improving access to care among
the uninsured.25,41The present study demonstrated signifi-
cant improvements in several outcomes that could improve
patients’ quality of life and reduce hospital financial burden.
Due to the increasing concern regarding health disparities,
and the financial implications of uninsured individuals, the
PAD program provides relevant scientific information on
the benefits of community-based donated care models that
can be replicated in other urban environments.
This work was supported by the Centers for Disease
Control and Prevention [R06 CCR621627 to M.J.D.].
1The Kaiser Commission on Medicaid and the Uninsured. The un-
insured: a primer. Menlo Park, CA: The Henry J. Kaiser Family
2The Kaiser Commission on Key Facts, Henry J. Kaiser Foundation.
The uninsured and the difference health insurance makes. www.kff.
org/uninsured/1420.cfm (1 July 2011, date last accessed).
3 Lurie N, Dubowitz T. Health disparities and access to health.
4Tolbert J. Approaches to covering the uninsured: a guide. Menlo
Park, CA: The Kaiser Family Foundation, 2008.
5 The Kaiser Commission on Medicaid and the Uninsured.
Uninsured in America, a chart book. Menlo Park, CA: The Henry
J. Kaiser Family Foundation, 2000.
6 Families USA. Getting less care: The uninsured with chronic health
conditions. Washington, DC: Families USA Foundation, 2001.
7 The Kaiser Commission on Medicaid and the Uninsured. Health
insurance coverage in America. www.kff.org/uninsured/4154.cfm
(1 August 2004, date last accessed).
8 Mokdad AH, Marks JS, Stroup DF et al. Actual causes of death in
the United States. JAMA 2004;291:1238–45.
9 Newton M, Keirns C, Cunningham R et al. Uninsured adults presenting
to US emergency departments. JAMA 2008;300(16):1914–24.
10 Weiner SJ, VanGeest J, Abrams RI et al. Avoiding free care at all
costs: a survey of uninsured patients choosing not to seek emer-
gency services at an urban county hospital. J Urban Health
11 Coleman P, Irons R, Nicholl J. Will alternative immediate care ser-
vices reduce demands for non-urgent treatment at accident and
emergency? Emerg Med J 2001;18(6):482–7.
12 Grumbach K, Keane D, Bindman A. Primary care and public
emergency department overcrowding. Am J Public Health 1993;
13 Bamezai A, Melnick G, Nawathe A. The cost of an emergency de-
partment visit and its relationship to emergency department
volume. Ann Emerg Med 2005;45(5):483–90.
14 Zeng F, O’Leary JF, Sloss EM et al. The effect of medicare health
maintenance organizations on hospitalization rates for ambulatory
care-sensitive conditions. Med Care 2006;44(10):900–7.
15 Bindman AB, Grumbach K, Osmond D et al. Preventable
hospitalizations and access
16 Cunningham PJ, Clancy CM, Cohen JW et al. The use of hospital
emergency departments for nonurgent health problems: a national
perspective. Med Care Res Rev 1995;52(4):453–74.
17 Cohen JW . Medicaid policy and the substitution of hospital out-
patient care for physician care. Health Serv Res 1989;24(1):33–66.
18 Davidson RA, Giancola A, Gast A et al. Evaluation of access, a
primary care program for indigent patients: inpatient and emergency
room utilization. J Community Health 2003;28(1):59–64.
19 Hurley RE, Freund DA, Taylor DE. Emergency room use and
primary care case management: evidence from four Medicaid dem-
onstration programs. Am J Public Health 1989;79(7):843–6.
20 Perloff JD, Kletke P, Fossett JW. Which physicians limit their
Medicaid participation, and why. Health Serv Res 1995;30(1):7–26.
21 Feder J, Levitt L, O’Brien E et al. Covering the low-income uninsured:
the case for expanding public programs. Health Aff 2001;20:27–39.
JOURNAL OF PUBLIC HEALTH
by guest on November 22, 2015
22 Guttman N, Zimmerman DR, Nelson MS. The many faces of
access: reasons for medically nonurgent emergency department
visits. J Health Polit Policy Law 2003;28(6):1089–120.
23 Malone RE. Whither the almshouse? Overutilization and the role
ofthe emergency department.
J Health Polit PolicyLaw
24 Evans RG, Stoddard GL. Producing health, consuming health care.
Soc Sci Med 1990;31(12):1247–363.
25 Taylor E, Cunningham P, McKenzie K. Community approaches to
providing care for the uninsured. Health Aff 2006;25:173–82.
26 Baird LC. Spiritually and faith in health care delivery. Community
Health Center Manag1999;33:24–6.
27 DeHaven MJ, Hunter IB, Wilder L et al. Health programs in faith-
based organizations: are they effective? Am J Public Health
28 DeHaven MJ, Gimpel NE. Reaching out to those in need: the case
for community health science. JABFM 2007;20(6):527–32.
29 Malone RE. Heavy users of emergency services: social construction
of a policy problem. Soc Sci Med 1995;40(4):469–77.
30 McNamara P, Witte R, Koning A. Patchwork access. Primary care
in EDs on the rise. Hospitals 1993;67(10):44–6.
31 Padgett DK, Brodsky B. Psychosocial factors influencing non-
urgent use of the emergency room: a review of the literature and
recommendations for research and improved service delivery. Soc
Sci Med 1992;35(9):1189–97.
Washington, D.C: General Accounting Office, Human Resources
33 Brown EM, Goel V. Factors related to emergency department visits
for nonurgent care. Health Aff 1994;13(5):162–71.
34 Althaus F, Paroz S, Hugli O et al. Effectiveness of interventions tar-
geting frequent users of emergency departments: a systematic
review. Ann Emerg Med 2011;58:41–52.
35 Baker LC, Baker LS. Excess cost of emergency department visits
for nonurgent care. Health Aff (Millwood) 1994;13(5):162–71.
36 Hilditch J. Changes in hospital emergency department use asso-
ciated with increased family physician availability. J Fam Pract
37 Burton A, Friedenzohn I, Martinez-Vidal E. State strategies
for policymakers. New York, NY: The Commonwealth Fund,
coverage:trends and lessons
38 Rylander CK. Texas estimated health care spending on the unin-
sured. Window on State Government. www.window.state.tx.us/
uninsure/ (1 July 2010, date last accessed).
39 Rogof DP. Health care safety nets and the art of making crazy
quilts. In: Paper Presented at Community Access Program TA
Conference Call, 2002.
40 Chang DI. Applying lessons learned in communities to programs
and policies at the federal level. Health Aff 2006;25:W192–4.
41 Minyard K, Chllet D, Felland L et al. Lessons from Local Access
Commonwealth Fund, 2007.
Challenges, New York,NY:
by guest on November 22, 2015