a b s t r a c t

Despite national concern about homeless veterans, there has been little examination of their use of emergency department (ED) services. This study examines factors related to the use of ED services in the Veterans Affairs healthcare system, where insurance is not a barrier to ambulatory healthcare. National VA administrative data from fiscal year 2010 are used to describe the proportions of ED users among homeless and domiciled VA patients. A case-control design is then used to compare homeless ED and non-ED users on sociodemographic and clinical correlates, as well as use of ambulatory care and Psychotropic medications. Sixteen percent of domiciled VA patients used EDs at least once during the year and 1% were Frequent ED users (N 4 ED visits) compared to 45% of homeless VA patients, 10% who were frequent ED users. Among homeless VA patients, those who used EDs were more likely to have a range of psychiatric and medical conditions, and had more service visits and psychotropic Medication prescriptions than non-ED users. Multivariate analyses suggest their risk for psychiatric and medical conditions increase their likelihood of using ED services. The high rate of ED use among homeless veterans is associated with significant morbidity, but also greater use of ambulatory care and psychotropics suggesting their ED use may reflect unmet psychosocial needs.

Introduction

Homeless adults use emergency department (ED) services at a rate 3 times that of the general population and many express concern that much of this use is driven by factors other than medical necessity [1]. One study defined frequent ED use as more than 4 visits in a year and found that homeless adults had nearly four times the odds of being a frequent ED user than domiciled adults [2].

Several studies have compared health risks between homeless and domiciled ED users finding higher rates of infectious diseases, psychiatric illness, and substance abuse among homeless ED users [3,4]. Fewer studies have compared ED and non-ED users among Homeless patients, which is important to help identify distinctive risk factors for ED use within the homeless population and provide insight into how to meet their needs in more emergent ways. Past studies have found that unstable housing, chronic medical illness, food-insecurity, and victimization were predictors of ED use

? Disclosure: None of the authors report any conflicts of interest. This work was supported by the United States Department of Veterans Affairs, Veterans Health Administration, Office of research and development, but the views presented here are those of the authors, alone, and do not represent the position of the United States Government.

* Corresponding author. Tel.: +1 203 932 5711×2090.

E-mail address: [email protected] (J. Tsai).

among homeless adults [1,5-7], although none have examined homeless veterans.

Homelessness among military veterans is of special national concern, but there have been few studies of their ED use in the Department of Veterans Affairs (VA) healthcare system, where there are no insurance barriers to accessing primary and specialized healthcare, which has been reported in non-VA studies as a reason for frequent ED use [3,5,8]. So presumably, VA ED services may be more likely to be sought for needs other than regular medical needs. The current study examines the proportion of homeless VA service users who use VA ED services at moderate and high levels over a one year period in comparison with domiciled VA service users, and then identifies sociodemographic, clinical, service use, and psychotropic medication variables associated with ED use among homeless VA

service users.

Methods

Data source and measures

National VA administrative data from fiscal year 2010 were obtained from several VA workload databases (Patient Treatment File including discharge abstracts of all inpatient episodes, the Outpatient Encounter File documenting demographic and diagnostic

0735-6757/$ - see front matter. http://dx.doi.org/10.1016/j.ajem.2013.02.046

856 J. Tsai, R.A. Rosenheck / American Journal of Emergency Medicine 31 (2013) 855-858

data for all outpatient service delivery including emergency depart- ment services, and the VA Drug Benefit Management System documenting all filled prescriptions), which were merged together to detail all outpatient and inpatient care and psychotropic medica- tion received from all VA facilities. Outpatient data are based on patient encounters that are documented in the VA’s electronic medical record system by clinicians and were used to obtain demographic characteristics, medical and Psychiatric diagnoses, health service use, and facility-level data.

These administrative data were first used to compare ED use among the population of homeless and domiciled VA service users, categorizing them into three groups: non-ED users, moderate ED users (1-4 ED visits), and frequent ED users (more than 4 ED visits). A further series of analyses, using a cross-sectional case-control design, compared homeless non-ED users to moderate and frequent ED users on sociodemographic characteristics, medical and psychi- atric diagnoses, psychotropic medication prescriptions, and mortality risk indicators.

Table 1

Sociodemographics, medical, psychiatric status, and psychtropic medications of homeless VA patients: ED users and non-users in fiscal year 2010

	Non-ED users	Moderate ED users	Frequent ED users	Moderate ED users vs.	Frequent ED users vs.
	(n = 78604)	(1-4 ED visits) (n = 50485)	(N 4 ED visits) (n = 13606)	non-ED users	non-ED users
Sociodemographics Age	51.88 (10.96)	51.73 (10.46)	52.26 (10.21)	d = 0.01a	d = 0.04
Male	73180 (93.10%)	46800 (92.70%)	12734 (93.59%)	OR = 0.94	OR = 1.08
White	11783 (14.99%)	8567 (16.97%)	2914 (21.42%)	OR = 1.16	OR = 1.55
OEF/OIF/OND status	4583 (5.83%)	2772 (5.49%)	562 (4.13%)	OR = 0.94	OR = 0.70
Service-connection
None	60446 (76.90%)	36743 (72.78%)	9609 (70.62%)	OR = 0.80	OR = 0.72
b50%	10934 (13.91%)	7967 (15.78%)	2172 (15.96%)	OR = 1.16	OR = 1.18
>=50%	7224 (9.19%)	5775 (11.44%)	1826 (13.42%)	OR = 1.28	OR = 1.53

Income medical status

$10429.98 (20485.08)	$10435.33 (16135.60)	$10283.20 (15473.81)	d b 0.001	d = 0.01
8.52 (8.98)	12.15 (11.92)	17.38 (15.82)	d = 0.34	d = 0.69

No. of medical visits^b

Myocardial infarction	511 (0.65%)	676 (1.34%)	297 (2.18%)	OR = 2.09	OR = 3.43
Congestive heart failure	32212 (40.98%)	31775 (62.94%)	10751 (79.02%)	OR = 2.45	OR = 5.43
Peripheral vascular disease	1942 (2.47%)	1964 (3.89%)	748 (5.50%)	OR = 1.60	OR = 2.30
Cerebral vascular accident	2091 (2.66%)	2267 (4.49%)	922 (6.78%)	OR = 1.72	OR = 2.66
Dementia	157 (0.20%)	217 (0.43%)	90 (0.66%)	OR = 2.17	OR = 3.37
Chronic pulmonary disease	8961 (11.40%)	9113 (18.05%)	3842 (28.24%)	OR = 1.71	OR = 3.06
Connective tissue disease	409 (0.52%)	348 (0.69%)	14 (0.10%)	OR = 1.34	OR = 1.98
or rheumatic disease
Peptic ulcer disease	574 (0.73%)	641 (1.27%)	324 (2.38%)	OR = 1.75	OR = 3.31
Liver disease	4575 (5.82%)	4549 (9.01%)	1735 (12.75%)	OR = 1.60	OR = 2.37
Diabetes without complications	10329 (13.14%)	9042 (17.91%)	2987 (21.95%)	OR = 1.44	OR = 1.86
Diabetes with complications	1981 (2.52%)	2045 (4.05%)	848 (6.23%)	OR = 1.63	OR = 2.57
Paraplegia and hemiplegia	338 (0.43%)	369 (0.73%)	146 (1.07%)	OR = 1.69	OR = 2.50
Renal disease	1533 (1.95%)	1817 (3.60%)	766 (5.63%)	OR = 1.88	OR = 3.01
Cancer	3066 (3.90%)	2984 (5.91%)	947 (6.96%)	OR = 1.55	OR = 1.85
Moderate/severe liver disease	244 (0.31%)	404 (0.80%)	249 (1.83%)	OR = 2.59	OR = 5.96
Metastatic cancer	189 (0.24%)	333 (0.66%)	137 (1.01%)	OR = 2.71	OR = 4.21
HIV/AIDS	1140 (1.45%)	1141 (2.26%)	344 (2.53%)	OR = 1.57	OR = 1.76
Charlson index	2.00 (1.89)	2.58 (2.20)	3.20 (2.42)	d = 0.28	d = 0.55
Psychiatric status No. of mental health visits c	22.89 (44.26)	35.70 (54.25)	49.78 (63.86)	d = 0.26	d = 0.49
Alcohol use disorder	24587 (31.28%)	23521 (46.59%)	8279 (60.85%)	OR = 1.92	OR = 3.41
Drug use disorder	25845 (32.88%)	25338 (50.19%)	8894 (65.37%)	OR = 2.06	OR = 3.85
Schizophrenia	5251 (6.68%)	4730 (9.37%)	2025 (14.88%)	OR = 1.44	OR = 2.44
Other psychosis	2602 (3.31%)	3206 (6.35%)	1690 (12.42%)	OR = 1.98	OR = 4.14
bipolar disorder	6005 (7.64%)	6361 (12.60%)	2728 (20.05%)	OR = 1.74	OR = 3.03
Major Depressive disorder	10407 (13.24%)	10491 (20.78%)	3761 (27.64%)	OR = 1.72	OR = 2.50
Dysthymia	24634 (31.34%)	23395 (46.34%)	8285 (60.89%)	OR = 1.89	OR = 3.41
posttraumatic stress disorder	14094 (17.93%)	12576 (24.91%)	4006 (29.44%)	OR = 1.52	OR = 1.91
Other anxiety disorder	10494 (13.35%)	11081 (21.95%)	4855 (35.68%)	OR = 1.83	OR = 3.60
Adjustment disorder	10517 (13.38%)	8779 (17.39%)	2880 (21.17%)	OR = 1.36	OR = 1.74
Personality disorder	3317 (4.22%)	4291 (8.50%)	2411 (17.72%)	OR = 2.11	OR = 4.89
Any psychiatric disorder	57271 (72.86%)	44184 (87.52%)	12968 (95.31%)	OR = 2.61	OR = 7.57
Dual diagnosis	33029 (42.02%)	30498 (60.41%)	10283 (75.58%)	OR = 2.11	OR = 4.27
Psychotropic medications
No. of psychotropic	8.54 (25.74)	18.24 (38.63)	41.75 (67.97)	d = 0.30	d = 0.65
prescriptions
Any antidepressants	32652 (41.54%)	29897 (59.22%)	10058 (73.92%)	OR = 2.04	OR = 3.99
Any antipsychotics	13811 (17.57%)	14610 (28.94%)	6430 (47.26%)	OR = 1.91	OR = 4.20
Any anxiolytics	13048 (16.60%)	14025 (27.78%)	6276 (46.13%)	OR = 1.93	OR = 4.30
Any stimulants	613 (0.78%)	490 (0.97%)	156 (1.15%)	OR = 1.26	OR = 1.48
Any mood stabilizers	12325 (15.68%)	13267 (26.28%)	5661 (41.61%)	OR = 1.92	OR = 3.83
Any lithium	1.66%	1378 (2.73%)	635 (4.67%)	OR = 1.66	OR = 2.90
Any opiates	23.11%	22027 (43.63%)	8610 (63.28%)	OR = 2.57	OR = 5.73

Note: d = Cohen’s d; OR = odds ratio.

^a d = Cohen’s d; OR = Odds Ratio.

^b Medical visits including outpatient medical and surgical visits.

^c Mental health visits include outpatient psychiatric and substance abuse visits.

J. Tsai, R.A. Rosenheck / American Journal of Emergency Medicine 31 (2013) 855-858 857

Use of VA ED services was based on VA administrative records. Homelessness was measured by clinical documentation of use of specialized VA homeless program services and/or a V60.0 International Classification of Diseases, 9th Revision diagnostic code (indicating lack of housing). Medical and psychiatric diagnoses were based on International Classification of Diseases, 9th Revision Diagnostic codes entered by VA clinicians into medical records. Outpatient service use was based on clinic codes in workload files and data on psychotropic prescription fills were obtained from VA administrative pharmacy data. Severity of medical conditions was measured using the Charlson Index [9], which predicts the 10-year mortality for patients.

Results

ED use among homeless and domiciled VA service users

Among the 5388684 domiciled VA service users, 1.25% were identified as frequent ED users (N 4 ED visits), 14.83% as moderate ED users (1-4 ED visits), and 83.92% as non-ED users (no ED visits). Thus, 16.08% of domiciled VA service users were ED users.

In contrast, of the 142695 homeless VA service users, 9.54% were identified as frequent ED users (more than 7 times the rate of domiciled VA service users), 35.38% as moderate ED users, and 55.09% as non-ED users. Thus, 44.91% of homeless VA service users were ED users (nearly three times the rate of domiciled VA service users).

Homeless ED users and homeless non-ED users

Bivariate analyses showed that both moderate and frequent homeless ED users were more likely to have all the medical diagnoses assessed, reported more medical visits, and had a higher risk for mortality than homeless non-ED users (Table 1). Frequent ED users had incrementally higher risk than moderate ED users. A similar pattern was observed for psychiatric variables, where frequent and moderate homeless ED users were more likely to be diagnosed with a range of psychiatric diagnoses, had more mental health visits, and far more psychotropic prescriptions than non-homeless ED users.

Multivariate analyses using ordinal logistic regressions found the largest diagnostic differences between homeless ED and non-ED users were in dementia, moderate/severe liver disease, and congestive heart failure (odds ratio = 1.69-2.05), with the likelihood of these conditions increasing with frequent ED use (Table 2).

Discussion

National VA administrative data indicate that approximately 45% of homeless VA service users used VA ED services at least once in 2010, and nearly 10% were identified as frequent ED users, a rate 7 times higher than that of domiciled VA service users. These findings are consistent with previous studies of homeless adults [1,2], although a lower proportion of frequent ED users was found than among a sample of homeless adults in San Francisco (10% compared to 38%),

[2] perhaps because of the availability of regular healthcare to homeless veterans in the VA system. Nonetheless, these findings demonstrate the distinctively high rate of VA ED use among homeless veterans, in spite of the lack of insurance barriers to obtaining ambulatory medical care, suggesting both medical and non-medical factors play a role in increasing risk for ED use.

The prevalence of psychiatric diagnoses was high among homeless VA service users, and homeless frequent ED users had higher rates of psychiatric diagnoses than homeless non-ED users. This finding coupled with those that have established the higher rate of psychiatric and substance abuse problems among the homeless population in general [10,11] suggests the need for more comprehensive mental health services tailored to this population to not only curb ED use but help them permanently exit homelessness. It is notable though that unlike

Table 2

Ordinal logistic regression of the association between sociodemographics and clinical characteristics, and ED use among homeless veterans

Independent variable	Odds ratio	95% confidence interval
Age	1.00	.99-1.00
White	1.06	1.03-1.10
Service-connection greater than 50% a	1.07	1.03-1.11
# of total medical visits	1.03	1.03-1.03
Myocardial infarction	1.58	1.42-1.76
Congestive heart failure	1.69	1.64-1.73
Cerebral vascular accident	1.21	1.14-1.28
Dementia	2.05	1.68-2.50
Chronic pulmonary disease	1.17	1.13-1.21
peptic ulcer disease	1.47	1.32-1.63
Liver disease	1.08	1.04-1.13
Diabetes without complications	1.13	1.09-1.17
Diabetes with complications	1.12	1.05-1.19
Paraplegia and hemiplegia	1.30	1.13-1.49
Renal disease	1.57	1.47-1.68
Cancer	1.08	1.03-1.14
Moderate/severe liver disease	1.71	1.48-1.98
Metastatic cancer	1.58	1.33-1.87
HIV/AIDS	1.20	1.11-1.30
# of mental health visits	1.00	1.00-1.00
Alcohol use disorder	1.40	1.36-1.43
Drug use disorder	1.51	1.47-1.55
Schizophrenia	1.45	1.38-1.51
Other psychotic disorder	1.68	1.59-1.77
Bipolar disorder	1.37	1.32-1.42
Major depression	1.16	1.13-1.20
Dysthymia	1.31	1.28-1.34
PTSD	.97	.94-1.00
Other anxiety disorder	1.42	1.38-1.46
Adjustment disorder	1.19	1.15-1.23
Personality disorder	1.59	1.52-1.67

Note: All independent variables shown were P b .01. Odds ratio point estimates over

1.50 are bolded.

^a Service-connected greater is compared to not service-connected.

comparisons between non-homeless and homeless adults who use EDs [3,4], the odds ratios for psychiatric and medical diagnoses were comparable suggesting both mental health and medical problems are of similar importance. Homeless frequent ED users were more likely to have a range of medical conditions, including myocardial infarctions, cancer, and liver disease and had a higher risk of mortality (based on the Charlson index) than homeless service users who used EDs less frequently or not at all, extending previous studies that have found similar medical conditions being strong risk factors for death among homeless adults [12,13]. Self-care for these medical conditions can also be negatively impacted by psychiatric and substance abuse problems, for example, noncompliance with treatment [14].

Although the frequency of chronic medical illness is high among homeless adults, housing and Case management programs have been found to successfully reduce hospitalization rates and ED visits [15] and demonstration clinics integrating homeless, primary care, and mental health services have also shown success in reducing ED use [16]. These findings suggest there are interventions available for homeless individuals that not only target housing, but aim to reduce ED use as well. The already high levels of medical and mental health service use found among homeless ED users also suggest that they may be seeking help for unaddressed or unidentified psychosocial needs that deserve further exploration in efforts to prevent overuse of ED services among homeless individuals.

Study limitations

Several methodological limitations require comment. This study relied entirely on VA administrative data, and the validity of available diagnoses is undemonstrated. Second, data on psychosocial factors such as hunger, safety risk, and access to comfortable shelter which

858 J. Tsai, R.A. Rosenheck / American Journal of Emergency Medicine 31 (2013) 855-858

may drive ED use [17] were not available since the VA healthcare system does not survey such factors for administrative purposes. Finally, since these data are from the VA, they predominantly reflect the experiences of older males receiving services in a government operated healthcare system, and their generalizability to other populations and healthcare systems is unknown.

References

1. Kushel MB, Perry S, Bangsberg D, Clark R, Moss AR. Emergency department use among the homeless and marginally housed: results from a community-based study. Am J Public Health 2002;92:778-84.
2. Mandelberg JH, Kuhn RE, Kohn MA. Epidemiologic analysis of an urban, public emergency department’s frequent users. Acad Emerg Med 2000;7:637-46.
3. Ku BS, Scott KC, Kertesz SG, Pitts SR. Factors associated with use of urban emergency departments by the U.S. homeless population. Public Health Rep 2010;125:398-405.
4. D’Amore J, Hung O, Chiang W, Goldfrank L. The epidemiology of the homeless population and its impact on an urban emergency department. Acad Emerg Med 2001;8:1051-5.
5. Kushel MB, Vittinghoff E, Haas JS. Factors associated with the health care utilization of homeless persons. J Am Med Assoc 2001;285:200-6.
6. Padgett DK, Struening EL, Andrews H, Pittman J. Predictors of emergency room use by homeless adults in New York city: the influence of predisposing, enabling and need factors. Soc Sci Med 1995;41:547-56.
7. Kushel MB, Gupta R, Gee L, Haas JS. Housing instability and food insecurity as barriers to health care among low-income Americans. J Gen Intern Med 2006;21: 71-7.
8. Oates G, Tadros A, Davis SM. A comparison of national emergency department use by homeless versus non-homeless people in the United States. J Health Care Poor Underserved 2009;20:840-5.
9. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis 1987;40:373-83.
10. Susser E, Moore R, Link B. Risk factors for homelessness. Am J Epidemiol 1993;15: 546-56.
11. Folsom DP, Hawthorne W, Lindamer LA, et al. prevalence and risk factors for homelessness and utilization of mental health services among 10,340 patients with serious Mental illness in a large public mental health system. Am J Psychiatry 2005;162:370-6.
12. Hwang SW, Lebow JM, Bierer MF, O’Connell JJ, Orav EJ, Brennan TA. Risk factors for death in homeless adults in Boston. Arch Intern Med 1998;158:1454-60.
13. Hibbs JR, Benner L, Klugman L, et al. Mortality in a cohort of homeless adults in Philadelphia. N Engl J Med 1994;331:304-9.
14. DiMatteo MR, Lepper HS, Croghan TW. Depression is a risk factor for noncompliance with medical treatment: meta-analysis of the effects of anxiety and depression. Arch Intern Med 2000;160:2101-7.
15. Sadowski LS, Kee RA, VanderWeele TJ, Buchanan D. Effect of a housing and case management program on emergency department visits and hospitalizations among chronically ill homeless adults: a randomized trial. J Am Med Assoc 2009;301:1771-8.
16. McGuire J, Gelberg L, Blue-Howells J, Rosenheck RA. access to primary care for homeless veterans with serious mental illness or substance abuse: A follow-up evaluation of co-located primary care and homeless social services. Adm Policy Ment Health 2009;36:255-64.
17. Rodriguez RM, Fortman J, Chee C, Ng V, Poon D. Food, shelter and safety needs motivating homeless persons’ visits to an urban emergency department. Ann Emerg Med 2009;53:598-602.