Screening for homelessness among individuals initiating medication-assisted treatment for opioid use disorder in the Veterans Health Administration

Abstract

Objective:

To determine the prevalence of homelessness and risk for homelessness among veterans with opioid use disorder initiating treatment.

Setting:

Addiction treatment programs operated by the US Department of Veterans Affairs (VA).

Participants:

All veterans initiating treatment with methadone or buprenorphine for opioid use disorder between October 1, 2013 and September 30, 2014 (n = 2,699) who were administered the VA’s national homelessness screener.

Main outcome measures:

Self-reported homelessness or imminent risk of homelessness.

Results:

The prevalence of homelessness was 10.2 percent and 5.3 percent were at risk for homelessness. Compared to male veterans, women veterans were less likely to report homelessness (8.9 percent vs 10.3 percent) but more likely to be at risk (11.8 percent vs 4.9 percent). By age group, veterans aged 18-34 and 45-54 years most frequently reported homelessness (12.0 and 11.7 percent, respectively) and veterans aged 45-54 and 55-64 years most frequently reported risk, for homelessness (6.5 and 6.8 percent, respectively).

Conclusions:

The prevalence of homelessness in this population is approximately 10 times that of the general veteran population accessing care at VA. Screening identified a substantial number of veterans who could benefit from VA housing assistance and had not received it recently. Programs to address veteran homelessness should engage with veterans seeking addiction treatment. Integration of homelessness services into addiction treatment settings may, in turn, improve outcomes.

INTRODUCTION

Greater awareness of and attention to housing instability stands to facilitate better outcomes among individuals in drug treatment. Aside from poverty, substance abuse is perhaps the most salient individual risk factor for homelessness.^1-5 In the case of opioid use disorder, previous research has found homelessness and unstable housing to be common among individuals entering treatment.⁶ Furthermore, some,^7,8 but not all,⁹ studies have found homeless or unstably housed individuals to have worse treatment outcomes than housed individuals. In this study, we assess the extent of housing instability and homelessness among veterans initiating treatment with methadone or buprenorphine for opioid use disorder in treatment programs administered by the US Department of Veterans Affairs (VA).

The VA presents a unique opportunity to assess housing instability in a clinical context through its use of a universal screening instrument that assesses housing status for veterans receiving VA healthcare services.¹⁰ The implementation of this housing screener reflects the VA’s commitment to two of its priority goals: ending veteran homelessness and improving access to high-quality behavioral healthcare.¹¹

METHODS

Using national VA medical record data, we identified veterans initiating medication-assisted treatment with methadone or buprenorphine between October 1, 2013, and September 30, 2014, by searching for new episodes of “opioid substitution” (stop code 523). This stop code indicates care in a VA opioid treatment program and has previously been used to identify patients enrolled in medication-assisted treatment.¹² In addition, as buprenorphine can be provided outside the opioid treatment program setting (eg, primary care), we identified patients with a prescription for buprenorphine and an International Classification of Diseases-9 Clinical Modification code for “opioid dependence” (304.0x). To identify those initiating treatment, we only included veterans without an “opioid substitution” code or buprenorphine prescription within 6 months of the first visit during the study period.

The Homelessness Screening Clinical Reminder assesses veterans’ current experience of housing instability (“In the past 2 months, have you been living in stable housing that you own, rent, or stay in as part of a household?”) as well as veterans’ imminent risk of homelessness (“Are you worried or concerned that in the next 2 months you may NOT have stable housing that you own, rent, or stay in as part of a household?”). If a veteran answers affirmatively to either question, the screener assesses the veteran’s current living situation (ie, where the veteran lived for most of the previous 2 months): housing (with or without subsidy), with a friend or family, motel or hotel, institution (eg, residential treatment facility), shelter, street, or other. Veterans with multiple screening results were assigned to the highest priority category (ie, a veteran screening negative and then screening at risk for homelessness would be classified as at risk). Veterans who received homeless services through VA Specialized Homeless Programs within the previous 6 months or were receiving long-term or palliative care were excluded from screening. We calculated the prevalence of homelessness and risk for homelessness for the total sample as well as by demographic group. We compared prevalence by demographic group, and living situation among those veterans reporting homelessness or risk, using chi-square tests. This study was deemed exempt by the Philadelphia Veterans Affairs Medical Center Institutional Review Board. All analyses were conducted using SAS version 9.3 (SAS Institute, Cary, NC).

RESULTS

Of 3,216 veterans identified as initiating treatment for opioid use disorder, 83.9 percent (2,699/3,216) were screened; 1.6 percent (52/3,216) were excluded from screening because they reported that they were already receiving homeless assistance elsewhere. The total prevalence of homelessness was 10.2 percent (276/2,699) and 5.3 percent (143/2,699) were at risk (Table 1). Homelessness and risk differed significantly by gender (p < 0.001). Compared to male veterans, women veterans were less likely to screen positive for homelessness (8.9 percent vs 10.3 percent) but more likely to be at risk (11.8 percent vs 4.9 percent). Homelessness and risk also varied significantly by age (p = 0.002), with veterans aged 18-34 and 45-54 most frequently screening positive for homelessness (12.0 and 11.7 percent, respectively) and veterans aged 45-54 and 55-64 years most frequently reporting risk for homelessness (6.5 and 6.8 percent, respectively).

Table 1.

Demographics and homelessness screening results for veterans initiating medication-assisted treatment with methadone or buprenorphine for opioid use disorder

	Positive screen, no. (percent)		Negative screen, no. (percent)
	Homeless	Risk for homelessness
Total	276 (10.2)	143 (5.3)	2,280 (84.5)
Gender
Female	15 (8.9)	20 (11.8)	134 (79.3)
Male	261 (10.3)	123 (4.9)	2,146 (84.8)
Age
18-34	120 (12.0)	44 (4.4)	835 (83.6)
35-44	29 (7.5)	18 (4.6)	341 (87.9)
45-54	56(11.7)	31 (6.5)	391 (81.8)
55-64	66 (9.5)	47 (6.8)	582 (83.7)
65 and older	5 (3.6)	3 (2.2)	131 (94.2)
Living situation*
Housing—with subsidy	1 (0.4)	15 (10.4)	NA
Housing—no subsidy	33 (12.0)	54 (37.8)	NA
With friend or family	104 (37.7)	57 (39.9)	NA
Motel or hotel	19 (6.9)	0 (0)	NA
Institution	24 (8.7)	3 (2.1)	NA
Shelter	25 (9.0)	2 (1.4)	NA
Street	23 (8.3)	1 (0.7)	NA
Other situation	34 (12.3)	7 (4.9)	NA
Not answered	13 (4.7)	4 (2.8)

Note: NA, not applicable. Total sample n = 2,699.

^*Percentages for living situation are based on screening disposition (ie, column percentages); living situation was only recorded for veterans who screened positive for homelessness or risk and refers to their living situation for most of the 2 month period prior to screening.

Among veterans who screened positive for homelessness or risk, living situations (ie, where the veteran lived for most of the previous 2 months) differed significantly between veterans who screened positive for homelessness versus risk (p < 0.001). Compared to those experiencing homelessness, those at risk were more likely to be in housing with a subsidy (10.5 percent vs 0.4 percent) or housing without a subsidy (37.8 percent vs 12.0 percent). A similar proportion of both groups reported living with a friend or family (39.9 percent vs 37.7 percent).

DISCUSSION

We found a rate of homelessness among veterans initiating treatment for opioid use disorder (10.2 percent) that was more than 10 times the rate in the general veteran population receiving care at VA (0.9 percent).¹³ While rates varied among gender and age categories, they were consistently high across all groups. These findings lay the groundwork for research into the impact of housing instability on treatment outcomes, and on how addressing housing concerns as a dimension of treatment can facilitate improved adherence and outcomes. Specifically, further research to evaluate the impact of VA and non-VA homelessness services on treatment outcomes, as well as an evaluation of the impact of addiction treatment on success in securing housing, would allow tailoring and refining these programs to optimize patient outcomes in both domains.

This study has limitations. First, homelessness and risk are based on veterans’ self-report and there may be substantial variation in perceptions of housing instability among individuals. Second, because the Homelessness Screening Clinical Reminder is a national-level screen administered by different types of staff throughout VA, there may be variations in screening practices which influence response.

Outreach efforts targeting veterans with opioid use disorder in medication-assisted treatment is one example of a proactive means to engage veterans at high risk for homelessness. Screening identified a substantial number of veterans who could benefit from VA housing assistance but had not received it recently. As the number of homeless veterans continues to decline, efforts such as these will be critical to the VA’s commitment to eliminate veteran homelessness. Furthermore, integration of homelessness services into addiction treatment settings may improve outcomes.