If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Division of Geriatrics, Gerontology, and Palliative Care, University of California, San Diego School of Medicine, 9500 Gilman Dr, San Diego, CA 92161, United States of America
Department of Emergency Medicine, NYU School of Medicine, 550 First Avenue, New York, NY 10016, United States of AmericaDepartment of Population Health, NYU School of Medicine, 550 First Avenue, New York, NY 10016, United States of America
12.5% of those entering treatment for opioid use disorder experienced homelessness.
•
Only a quarter of those received medications to treat opioid use disorder.
•
There were sharp increases in co-use of methamphetamine use in this population.
Abstract
Objective
People experiencing homelessness (PEH) have high rates of substance use, and homelessness may be an important driver of health disparities in the opioid overdose epidemic. However, few studies focus on homelessness among the opioid use disorder (OUD) treatment population. We examine national-level trends in substance use treatment admissions among PEH with OUD.
Methods
This study used data from first-time treatment admissions in the United States from the Treatment Episode Data Set: Admissions (TEDS-A) to examine characteristics and trends of adults experiencing homelessness who entered state-licensed substance use treatment programs for OUD from 2013 to 2017. We used chi-squared analyses to examine changes in characteristics of this population over time and logistic regression to assess characteristics associated with receipt of medications for opioid use disorder (MOUD) among PEH.
Results
Among all adults with OUD entering specialty treatment from 2013 to 2017, 12.5% reported experiencing homelessness. Compared to individuals not experiencing homelessness, PEH were more likely to be male, inject opioids, use cocaine or methamphetamine, and enter into residential detoxification treatment. PEH were less likely to enter outpatient treatment or receive MOUD. From 2013 to 2017, significant increases occurred in the proportion of PEH who had co-occurring psychiatric problems and used methamphetamines. Over time, treatment type shifted significantly from residential detoxification to outpatient treatment. Receipt of MOUD increased among PEH over time (13.7% to 25.2%), but lagged behind increases among individuals not experiencing homelessness. Among PEH, being older was associated with receiving MOUD, while concurrent methamphetamine use [adjusted odds ratio (AOR) 0.63; 95% CI 0.58, 0.69] and living in the southern United States (AOR 0.27; 95% CI 0.25, 0.30) were associated with not receiving MOUD.
Discussion
The proportion of PEH with OUD who receive medications as part of treatment increased over time, but three quarters of PEH entering treatment still do not receive this highest standard in evidence-based care. The sharp increase observed in concomitant methamphetamine use in this population is concerning and has implications for treatment.
Wilson, N., Kariisa, M., Seth, P., Smith, H., 4th, & Davis, N. L. (2020). Drug and opioid-involved overdose deaths - United States, 2017-2018. MMWR. Morbidity and Mortality Weekly Report, 69(11), 290–297. doi:10.15585/mmwr.mm6911a4.
Singh, J. A., & Cleveland, J. D. (2020). National U.S. time-trends in opioid use disorder hospitalizations and associated healthcare utilization and mortality. PLoS One, 15(2), e0229174. doi:https://doi.org/10.1371/journal.pone.0229174.
), as well as the compound comorbidity of homelessness and other risk factors place people experiencing homelessness (PEH) at high risk for both nonfatal and fatal opioid overdose (
). However, many structural barriers and challenges exist for PEH to receive care in substance use treatment settings and to access MOUD. These barriers include stressors and competing priorities related to unstable housing, social isolation, and the dual stigma of substance use and homelessness (
). Additional challenges such as mental illness and criminal justice involvement act as further barriers to MOUD treatment for PEH. Studies of the Veteran population show low rates of MOUD and naloxone receipt among veterans experiencing homelessness or who are unstably housed (
). However, we know little about the characteristics and needs of PEH with OUD nationally who receive care through the specialty substance use treatment system, including their access to MOUD and whether this has shifted over time.
The purpose of this study is to describe the demographic, substance use, and treatment characteristics and trends among patients with OUD who were experiencing homeless at the time of substance use treatment entry and to examine correlates of receiving MOUD among this population. Understanding changing trends nationally is important, as it can help to inform how substance use treatment programs may need to adapt their services based on the current and changing needs of PEH.
2. Material and methods
2.1 Source of data
We used data from the Treatment Episode Data Set: Admission (TEDS-A) from 2013 through 2017. TEDS-A is a federal administrative database managed by the Substance Abuse and Mental Health Services Administration (SAMHSA) that contains information about individual treatment admissions in publicly funded substance use treatment facilities in the United States (including the District of Columbia and Puerto Rico) (
). Most state substance use treatment agencies require facilities that receive any stateor public funding (including federal block grant funds) for the provision of alcohol or drug treatment services to report TEDS data to the state (
Batts, K., Pemberton, M., Bose, J., Weimer, B., Henderson, L., Penne, M., & Strashny, A. (2014). Comparing and evaluating substance use treatment utilization estimates from the National Survey on Drug Use and Health and other data sources. In CBHSQ data review. (pp. 1–120). Substance Abuse and Mental Health Services Administration (US).
). We restricted our study sample to adults age eighteen and older who entered treatment programs primarily for problems related to the use of opioids, including heroin, nonprescription methadone, and other opioids and synthetics. As each record in TEDS-A represents a treatment admission, and no identifier exists to allow identification of a single individual across multiple admissions, analyses used first-time treatment admissions only (N = 762,671). The study defined persons experiencing homelessness as those who indicated their living arrangement to be homeless (no fixed address and includes living in shelters), rather than dependent (such as living with parents, relatives, guardians, or in supervised settings) or independent living (on his or her own) during admission. Analyses excluded 3.6% of records missing information on living arrangements. The final analysis included 735,368 individuals with a first-time admission for OUD during the study period.
2.2 Measures
The sociodemographic characteristics examined in this analysis included gender; age group (age 18–29, 30–38, 40–49, and 50 and older); race/ethnicity; education; employment status; health insurance status (no insurance, private insurance, Medicaid, and Medicare); veteran status; proxy to living in an urban area [dichotomized as urban versus non-urban by Core-based statistical area (CBSA) defined by the U.S. Office of Management and Budget (OMB)]; U.S. region; and arrest in the past 30 days. Also, analyses examined an individual's primary opioid type (heroin versus other opioid); usual route of administration (oral, smoking, nasal, injection, or other); frequency of opioid use in the past month (no use, few to multiple times, or daily or near-daily use); concurrent substance use (alcohol, cannabis, benzodiazepines, cocaine, or methamphetamine); presence of psychiatric problem; referral source (individual, criminal justice, care providers such as a clinic or any health care provider, or school/employer or community); and treatment facility type [detoxification hospital (24-h services inpatient), residential detoxification, residential hospital, residential short term (<30 days) or long-term (≥30 days), ambulatory intensive outpatient (≥2 h per day for ≥3 days a week), ambulatory nonintensive outpatient, or detoxification ambulatory]. Last, analyses examined if the patient received MOUD, defined by TEDS-A as methadone or buprenorphine (or naltrexone, beginning in the year 2016) taken for OUD as part of the treatment plan.
2.3 Data analysis
We first compared sociodemographic, substance use, and treatment characteristics between PEH and non-PEH entering substance use treatment for OUD by using Pearson chi-square tests for aggregated data from 2013 through 2017. To observe if changes occurred in characteristics among PEH entering OUD treatment over time, we examined changes in these characteristics among PEH between 2013 and 2017 by calculating the absolute and relative change over time. We then estimated if a significant change took place for each characteristic between 2013 and 2017 using Pearson chi-squared tests.
Next, to examine correlates of MOUD use among PEH entering OUD treatment, we performed a logistic regression to estimate odds ratios for receiving MOUD based on the presence of each covariate of interest. We then ran a multivariable logistic regression adjusting for all covariates, including admission year, to assess independent correlates of MOUD. As many records had missing data on one or more variables (37%), we conducted a sensitivity analysis by repeating the multivariable analysis using Multiple Imputation using Chained Equations (
). Additionally, health insurance status was missing for 59% of records, with the following states reporting insurance status for less than 50% of episodes: Arizona, California, Connecticut, Florida, Illinois, Michigan, Minnesota, New Mexico, New York, North Carolina, Ohio, Rhode Island, Vermont, Virginia, Washington, and Wisconsin (
). Given the large number of missing data, we add health insurance status only as a supplemental table and do not include it in the main multivariable logistic regression model. We conducted all analyses using Stata/S.E. version 14 (Stata-Corp) with significance indicated at a p-value<0.05. This research was exempt from institutional review board approval because it involved secondary analysis of de-identified existing data.
3. Results
3.1 Comparison of PEH versus non-PEH with OUD 2013–2017
Among the study sample, an estimated 12.5% (81,424) of adults entering substance use treatment for OUD for the first time between 2013 and 2017 reported homelessness. Table 1 presents demographic, opioid use, and treatment characteristics according to homelessness status. PEH were more likely to be male (65.6% vs. 58.4%; p < 0.001), unemployed (94.3% vs. 76.3%; p < 0.001), a veteran (2.5% vs. 2.2%; p < 0.001), arrested in the past 30 days prior to admission (7.9% vs. 6.1%; p < 0.001), live in an urban area (78.3% vs. 73.2%; p < 0.001), and live in the West (43.1% vs. 27.6%; p < 0.001). In terms of opioid use, PEH reported a higher prevalence of heroin as their primary opioid of use (78.3% vs. 60.0%; p < 0.001), injection as their usual route of administration (62.3% vs. 42.6%; p < 0.001), and were more likely to use opioids daily or near-daily (65.4% vs. 61.5%, p < 0.001). They also had higher prevalence of concurrent alcohol use (11.7% vs. 10.4%, p < 0.001), cocaine use (15.6% vs. 11.3%; p < 0.001), and methamphetamine use (19.9% vs. 8.7%, p < 0.001) compared to non-PEH. PEH were more likely to have a comorbid psychiatric problem at admission (33.7% vs. 28.5%, p < 0.001). For treatment, PEH were less likely to receive nonintensive outpatient treatment (26.2% vs. 52.5%; p < 0.001) and more likely to utilize residential detoxification (40.8% vs. 19.3%; p < 0.001) and residential long-term facilities (13.8% vs. 4.6%; p < 0.001). PEH were also less likely to receive MOUD as part of their treatment plan (21.1% vs. 36.4%; p < 0.001) compared to non-PEH.
Table 1Characteristics of adults entering opioid use disorder treatment by people experiencing homelessness vs. people not experiencing homelessness, United States 2013–2017 (N = 735,368).
Persons experiencing homelessness (n = 81,424)
Persons not experiencing homelessness (n = 653,944)
χ2 P-value
Number
%
Number
%
Received MOUD
17,183
21.1
237,827
36.4
<0.001
Female
28,008
34.4
272,106
41.6
<0.001
Age group
18–29
34,179
42.0
291,536
44.6
<0.001
30–39
25,417
31.2
195,792
29.9
40–49
12,137
14.9
88,554
13.5
≥50
9691
11.9
78,062
11.9
Race
White
60,233
74.7
505,714
78.4
<0.001
Black
9178
11.4
69,889
10.8
Other
11,209
13.9
69,279
10.7
Hispanic (any race)
12,187
15.3
78,726
12.3
<0.001
<12 years education
23,494
29.5
157,544
24.9
<0.001
Unemployed
75,917
94.3
485,465
76.3
<0.001
Veteran status
1884
2.5
13,232
2.2
<0.001
Arrest past 30-days
5323
7.9
35,360
6.1
<0.001
Urban
63,754
78.3
478,803
73.2
<0.001
US region
West
35,056
43.1
180,737
27.6
<0.001
Midwest
8476
10.4
84,671
13.0
South
23,535
28.9
226,120
34.6
Northeast
14,256
17.5
161,879
24.8
Opioid type
Heroin
63,723
78.3
392,618
60.0
<0.001
Frequency of opioid use in past month
No use
11,577
14.3
132,348
20.5
<0.001
Few to multiple times
16,471
20.3
116,324
18.0
Daily or near-daily use
53,046
65.4
396,543
61.5
Usual route of administration (opioid)
Oral
9664
11.9
169,891
26.2
<0.001
Smoking
7222
8.9
50,418
7.8
Nasal
12,808
15.8
136,803
21.1
Injection
50,527
62.3
276,096
42.6
Other
933
1.2
15,515
2.4
Concurrent substance use
Alcohol
9549
11.7
67,824
10.4
<0.001
Cannabis
12,039
14.8
113,169
17.3
<0.001
Benzodiazepines
6097
7.5
47,860
7.3
0.08
Cocaine
12,731
15.6
73,760
11.3
<0.001
Methamphetamine
16,233
19.9
56,701
8.7
<0.001
Concurrent psychiatric problem
25,114
33.7
168,054
28.5
<0.001
Treatment facility type
Detoxification hospital
836
1.0
5699
0.9
<0.001
Detoxification residential
33,209
40.8
126,307
19.3
Residential hospital
176
0.2
1804
0.3
Residential short-term
7394
9.1
48,337
7.4
Residential long-term
11,233
13.8
30,095
4.6
Intensive outpatient
4496
5.5
61,116
9.4
Non-intensive outpatient
21,329
26.2
343,468
52.5
Detoxification ambulatory
2751
3.4
37,118
5.6
Referral
Individual (includes self-referral)
49,571
62.3
387,945
60.4
<0.001
Criminal justice system
10,927
13.7
108,515
16.9
Care provider (any program, clinic, or health care provider)
3.2 Changing trends in characteristics of PEH with OUD, 2013–2017
From 2013 to 2017, a significant increase occurred in the number and proportion of people reporting homelessness among first-time admissions for OUD, from n = 14,863 (10.6% of OUD treatment population) in 2013 to n = 17,958 (14.6%) in 2017 (Table A.1). The proportion of PEH receiving MOUD had an absolute increase of 11.5% and a relative increase of 83.9%, from 13.7% in 2013 to 25.2% in 2017 (Fig. 1). A 23.6% relative increase occurred in PEH entering OUD treatment who were aged 50 and older (13.1% in 2017 vs. 10.6% in 2013) and among adults age 30–39 (35.8% in 2017 vs. 27.7% in 2013), a 19.8% relative increase in those reporting heroin as their primary opioid used (83.6% vs. 69.8%), an 81.2% relative increase in reporting concurrent use of methamphetamine (25.0% vs. 13.8%), a 25.7% relative increase in psychiatric comorbidity (38.2% vs. 30.4%), and a 52.3% relative increase in utilizing nonintensive outpatient treatment vs. other treatment types (52.3 vs. 19.3%). Meanwhile, there were decreases in the relative proportion of PEH who lived in urban areas (69.6% in 2017 vs. 88.1% in 2013), and had concurrent use of alcohol (10.0% vs. 14.7%) and cannabis (12.8% vs. 14.8%). We also note a decrease in residential detoxification (23.5% relative decrease), while there were sharp increases in both intensive outpatient (30.0% relative increase) and nonintensive outpatient (52.3% relative increase) treatments.
Fig. 1Trends in prevalence of receiving medications for opioid use disorder (OUD) and substance use patterns among people with OUD experiencing homelessness, 2013–2017.
3.3 Correlates of receiving MOUD among PEH with OUD, 2017
Table 2 shows results from the multivariable logistic regression model with receipt of MOUD as the outcome variable. Results from the adjusted model show that among PEH entering OUD treatment, those aged 50 and older compared to adults aged 18–29 [adjusted odds ratio (AOR) 1.87; 95% CI 1.70, 2.05], PEH identifying as Black (AOR 1.15; 95% CI 1.05, 1.26) or other race (AOR 1.62; 95% CI 1.48, 1.78) compared to White PEH were more likely to receive MOUD. Heroin vs. other types of opioids used (AOR 1.52; 95% CI 1.26, 1.87) and treatment in nonintensive outpatient (AOR 5.95; 95% CI 4.67, 7.57) and ambulatory detoxification (AOR 17.86; 95% CI 13.40, 23.80) were associated with greater MOUD receipt. Conversely, PEH identifying as Hispanic (AOR 0.73; 95% CI 0.66, 0.80), concurrent methamphetamine use (AOR 0.63; 95% CI 0.58, 0.69), living in nonurban areas (AOR 0.58; 95% CI 0.54, 0.62) and the southern United States (AOR 0.27; 95% CI 0.25, 0.30), referral from the criminal justice system (AOR 0.15; 95% CI 0.13, 0.17) or a care provider (AOR 0.59; 95% CI 0.54, 0.64) versus self-referral, residential short-term (AOR 0.31; 95% CI 0.24, 0.40) and residential long-term (AOR 0.43; 95% CI 0.33, 0.55) vs. detoxification hospital were all associated with lower odds of MOUD receipt. Sensitivity analyses with imputed data only slightly qualitatively changed the findings (Table A.2). Adding insurance status as a covariate to the model reduced the sample size to n = 17,771 and insurance status with private insurance (AOR 1.48; 95% CI 1.14, 1.93), Medicaid (AOR 1.86; 95% CI 1.64, 2.11), and Medicare (AOR 1.28; 95% CI 1.02, 1.62) were all associated with receipt of MOUD (Table A.3).
Table 2Adjusted odds ratios of receiving medications for opioid use disorder (MOUD) among people in treatment for opioid use disorder experiencing homelessness, United States 2013–2017 (N = 81,424).
In the United States, the number and proportion of people entering specialty substance use treatment for OUD who reported homelessness increased from 2013 to 2017. While the proportion of this population who received MOUD as part of treatment increased sharply during this period, still in 2017, nearly 75% of patients did not receive buprenorphine, methadone, or naltrexone for OUD upon entry into treatment. The findings of this study highlight the distinctive sociodemographic and substance use characteristics of PEH with OUD and how this population's characteristics and needs are changing with time.
The study emphasizes the rising concern of methamphetamine use, especially among PEH who are in treatment for OUD. Results from our study echo a recent study that showed increases in methamphetamine use among heroin treatment admissions in the United States from 2008 to 2017, with homelessness strongly correlated with the co-use of methamphetamine and heroin (
). Our study examined PEH's admissions in more detail, and found that not only did PEH with OUD have a markedly higher prevalence of methamphetamine use compared to non-PEH, but a sharp increase occurred such that by 2017 a quarter of PEH with OUD entering treatment also used methamphetamine. This increase is occurring as methamphetamine use is escalating nationally among people reporting past-year heroin use (22.5% in 2015 to 37.4% in 2018) (
Palamar, J. J., Han, B. H., & Keyes, K. M. (2020). Trends in characteristics of individuals who use methamphetamine in the United States, 2015-2018. Drug and Alcohol Dependence, 213, 108089. Advance online publication.
Kariisa, M., Scholl, L., Wilson, N., Seth, P., & Hoots, B. (2019). Drug overdose deaths involving cocaine and psychostimulants with abuse potential - United States, 2003-2017. MMWR. Morbidity and Mortality Weekly Report, 68(17), 388–395. doi:10.15585/mmwr.mm6817a3.
). Our results also found that PEH with OUD using alcohol and benzodiazepines were significantly less likely to receive MOUD. While treatment may be challenging for patients engaged with polysubstance use, provider expectations may exist for patients with OUD initiating treatment with MOUD to stop all other substance use. However, it is important to emphasize that MOUD is evidence-based treatment for reducing opioid use and its related harms, and providers should not necessarily expect or demand improvement with other substances. Therefore, marginalized populations such as PEH need interventions to better address co-use of substances; for instance, a combination of MOUD with evidence-based treatment for stimulant use disorder such as contingency management for patients who use methamphetamine (De Crescenzo et al., 2018).
The results of this study also note differences in substance use treatment facility types attended by PEH compared to individuals who were not homeless. PEH had a much higher prevalence of residential detoxification and residential short/long-term care settings versus outpatient settings. One possibility for this finding is that PEH may seek out residential treatment settings to fill gaps in housing, with such settings meeting not only substance use treatment needs but also a subsistence need. Therefore, while nationally a push exists for increasing nonintensive outpatient treatment, which was also seen as a trend in our study among PEH with OUD, reduced access to residential treatment settings could negatively impact PEH in terms of both access to substance use treatment and ability to address a key, urgent subsistence need. Expanding the availability of permanent supportive housing coupled with robust connections to substance use treatment may help to address the concomitant substance use treatment and housing needs of PEH in a potentially more durable way than short-term residential treatment. While permanent supportive housing (including Housing First models) as an intervention itself has shown mixed results on a range of substance use outcomes (
Davidson, C., Neighbors, C., Hall, G., Hogue, A., Cho, R., Kutner, B., & Morgenstern, J. (2014). Association of housing first implementation and key outcomes among homeless persons with problematic substance use. Psychiatric Services (Washington, D.C.), 65(11), 1318–1324. doi:https://doi.org/10.1176/appi.ps.201300195.
). Housing First models have demonstrated undeniable success in meeting their primary outcome goal of housing people and keeping them—including those with substance use disorders—durably housed (
), such an approach may provide long-term benefit for PEH who would traditionally seek residential substance use treatment. More permanent options may work better for this population than state-licensed inpatient or residential detoxification settings, which tend to have relatively short stays and high repeat users, especially among PEH (
This study also found key differences among PEH by race and ethnicity. While PEH with OUD entering treatment identifying as White decreased slightly during the study period, sharp increases occurred among PEH identifying as Black, other race, and Hispanic (any race). The increases are occurring in the setting of significant rises in opioid-involved overdose among Black and Hispanic residents of metropolitan areas in the United States (
Lippold, K. M., Jones, C. M., Olsen, E. O., & Giroir, B. P. (2019). Racial/ethnic and age group differences in opioid and synthetic opioid-involved overdose deaths among adults aged ≥18 years in metropolitan areas - United States, 2015-2017. MMWR. Morbidity and Mortality Weekly Report, 68(43), 967–973. doi:10.15585/mmwr.mm6843a3.
). While in our multivariable analyses, Black race was significantly correlated with receipt of MOUD among PEH with OUD entering treatment (although the effect size was small), people identifying as Hispanic of any race were significantly less likely to receive MOUD. These findings warrant more exploration in future research. In general, the widening racial disparities observed in overdose deaths in the United States in recent years (
Lippold, K. M., Jones, C. M., Olsen, E. O., & Giroir, B. P. (2019). Racial/ethnic and age group differences in opioid and synthetic opioid-involved overdose deaths among adults aged ≥18 years in metropolitan areas - United States, 2015-2017. MMWR. Morbidity and Mortality Weekly Report, 68(43), 967–973. doi:10.15585/mmwr.mm6843a3.
) underscore the necessity of concerted research and programmatic efforts to understand and address structural racism and its effects, including focused attention to PEH who are themselves disproportionately Black and Hispanic (
We found regional differences among PEH entering OUD treatment in the United States. A higher percentage of PEH entering substance use treatment were from the West, which may reflect the high rates of homelessness, particularly unsheltered homelessness in West Coast states, including California, Oregon, and Washington (
) or greater engagement in substance use treatment services among PEH in these regions. Along with a high concentration of methamphetamine use also in the West Coast (
), this finding emphasizes the need for comprehensive approaches to address the overlapping issues of homelessness, OUD, and methamphetamine use among PEH living in this geographical area, and more generally, for recognition of potentially unique regional needs in designing interventions for marginalized populations. We also found a decreasing trend of PEH entering opioid treatment who were from urban areas and that living in a nonurban area was correlated with not receiving MOUD. The latter finding likely reflects that specialty OUD programs that offer MOUD are largely concentrated in urban areas (
). Also, living in the southern U.S. was strongly correlated with not receiving MOUD. The southern U.S., particularly states in the South Atlantic region, has many counties with low availability of MOUD providers and high rates of opioid overdose mortality (
). Strategies to increase access to MOUD in these areas must also focus on PEH. Future studies must examine treatment trends and challenges among PEH across geographical areas.
Despite missing data on health insurance in TEDS-A, our sensitivity analysis of those with insurance data highlights that among PEH with OUD, having health insurance was significantly correlated with receiving MOUD. Expanding health insurance coverage for PEH with OUD should be a priority and may increase receipt of MOUD. Furthermore, while evidence exists that low barrier buprenorphine treatment through street medicine or mobile treatment (
Buzza, C., Elser, A., & Seal, J. (2019). A mobile buprenorphine treatment program for homeless patients with opioid use disorder. Psychiatric Services (Washington, D.C.), 70(7), 635–636. doi:https://doi.org/10.1176/appi.ps.70701.
) may benefit PEH with OUD, the reach of such interventions is currently limited, with many receiving care instead through state-licensed substance use treatment programs not generally tailored to meet the unique needs of PEH. Given the high and rising rates of homelessness among individuals with OUD entering treatment observed in this study, OUD treatment settings should consider not only the unique barriers that PEH experience due to lack of housing but also ways to manage polysubstance use, including the sharp increase in methamphetamine and concurrent psychiatric needs. Engaging PEH already in OUD treatment with MOUD specifically should be a priority.
4.1 Limitations
This study has several limitations. First, we limited our study sample to patients receiving treatment for the first time. While doing so was necessary to account for multiple possible episodes per person, patterns of substance use treatment type may differ from patients with previous treatment episodes, and so our results may not be generalizable to PEH with multiple treatment episodes. Second, TEDS-A collects data from state-certified substance use treatment centers and not office-based (e.g., primary care) settings. Therefore, this study would not be representative for patients who receive MOUD from outside the settings that TEDS-A captures, such as buprenorphine or naltrexone prescribed or delivered in medical office-based settings. In addition, our study sample captures only PEH who are entering treatment for OUD and is not representative of PEH with OUD in general. Future studies should help better understand OUD in the larger population experiencing homelessness, though the limited availability of national data on homelessness presents significant limitations to such research. Third, the TEDS-A definition for MOUD changed in 2016 from including only buprenorphine or methadone to also including naltrexone, potentially expanding the proportion of people in treatment considered to receive MOUD. However, we observed only a tiny increase in the proportion of PEH categorized as receiving MOUD from 2015 to 2016 (from 22.3% to 22.6%), with larger changes in other years, suggesting that the change in TEDS-A definition of MOUD is not primarily responsible for the trend that we observed. Finally, many limitations exist regarding the variables collected as TEDS-A collects data for administrative purposes, not specifically for research.
5. Conclusion
Using national data, we found sharp increases in heroin as the primary opioid used and concomitant use of methamphetamines among PEH with OUD entering substance use treatment. We also found a trend toward decreases in residential treatment in this population and increases in both nonintensive outpatient treatment and MOUD treatment, but still only a quarter of PEH with OUD entering treatment received MOUD in 2017. Therefore, expanding MOUD treatment for OUD among this population and parallel efforts to improve access to safe housing should be a priority for policy-makers and treatment providers.
Funding sources
This research was funded by two grants through the National Institute on Drug Abuse: K23DA039179 (Doran) and K23DA043651 (Han).
CRediT authorship contribution statement
Benjamin H. Han: Conceptualization, Methodology, Writing- Original draft preparation. Kelly M. Doran: Conceptualization, Methodology, Writing- Reviewing and Editing. Noa Krawczyk: Conceptualization, Methodology, Formal analysis, Writing- Reviewing and Editing.
Table A.1Trends in characteristics of people with opioid use disorder experiencing homelessness, 2013–2017.
2013 (n = 14,836)
2014 (n = 15,391)
2015 (n = 16,374)
2016 (n = 16,947)
2017 (n = 17,958)
% change from 2013 to 2017
χ2p-value for difference between 2013 and 2017
Relative % change
Absolute % change
Characteristics
Proportion on agonist therapy (%)
13.7
20.6
22.3
22.6
25.2
83.9
11.5
<0.001
Female
33.9
35.1
34.5
34.8
33.8
−0.3
−0.1
0.94
Age group
18–29
47.2
43.5
42.8
41.7
35.9
−23.9
−11.3
<0.001
30–39
27.7
28.9
30.3
32.4
35.8
29.2
8.1
40–49
14.5
15.4
14.6
14.7
15.3
5.5
0.8
≥50
10.6
12.3
12.2
11.2
13.1
23.6
2.5
Race
White
78.5
75.2
74.3
73.6
72.6
−7.4
−5.8
<0.001
Black
10.4
10.9
11.7
10.9
12.8
23.2
2.4
Other
11.1
13.9
14.0
15.6
14.6
30.7
3.4
Hispanic (any race)
13.9
15.3
15.1
15.8
16.1
16.2
2.2
<0.001
<12 years education
30.4
30.2
30.1
28.1
28.8
−5.3
−1.6
<0.001
Unemployed
95.4
94.2
94.1
93.5
94.3
−1.2
−1.1
<0.001
Veteran status
3.0
2.5
2.3
2.1
2.6
−13.6
−0.4
0.04
Arrest past 30-days
8.3
8.4
7.1
7.5
8.0
−3.6
−0.3
0.36
Urban
88.1
83.4
77.0
75.5
69.6
−21.0
−18.5
<0.001
US region
West
42.2
46.4
42.1
43.0
41.8
−0.9
−0.4
0.002
Midwest
9.4
9.5
12.0
10.7
10.4
10.6
1.0
South
30.5
25.4
28.4
29.3
30.6
0.3
0.1
Northeast
17.8
18.7
17.4
16.8
17.0
−4.5
−0.8
Primary opioid type
Heroin
69.8
76.5
79.4
80.5
83.6
19.8
13.8
<0.001
Frequency of opioid use in past month
No use
14.6
15.2
14.6
13.5
13.7
−6.2
−0.9
0.06
Few to multiple times
20.0
20.2
20.3
21.0
20.0
0.0
0.0
Daily or near-daily use
65.3
64.6
65.1
65.5
66.3
1.5
1.0
Usual route of administration (opioid)
Oral
14.7
12.9
11.2
11.3
10.0
−32.0
−4.7
<0.001
Smoking
7.7
8.4
8.8
9.3
10.1
31.2
2.4
Nasal
14.5
14.3
15.8
16.0
18.0
24.1
3.5
Injection
61.3
63.3
63.3
62.4
61.0
−0.5
−0.3
Other
1.8
1.2
0.9
1.0
1.0
−44.4
−0.8
Concurrent substance use
Alcohol
14.8
12.5
11.6
10.2
10.0
−32.4
−4.8
<0.001
Cannabis
17.3
15.5
14.8
14.0
12.8
−26.0
−4.5
<0.001
Benzodiazepines
8.8
7.3
7.3
7.2
7.1
−19.3
−1.7
<0.001
Cocaine
15.8
14.2
15.1
15.6
17.3
9.5
1.5
<0.001
Methamphetamine
13.8
17.6
19.2
22.8
25.0
81.2
11.2
<0.001
Concurrent psychiatric problem
30.4
31.5
31.2
35.9
38.2
25.7
7.8
<0.001
Referral
Individual (includes self-referral)
60.4
60.4
63.4
63.1
63.8
5.6
3.4
<0.001
Criminal justice system
14.4
14.5
12.8
14.4
12.7
−11.8
−1.7
Care provider (any program, clinic, or health care provider)
Table A.2Adjusted odds ratios of receiving medications for opioid use disorder (MOUD) among people in treatment for opioid use disorder experiencing homelessness, United States 2013–2017, sensitivity analyses with imputed data (N = 81,424).
Table A.3Adjusted odds ratios of receiving medications for opioid use disorder (MOUD) among people in treatment for opioid use disorder experiencing homelessness with available health insurance status, United States 2013–2017 (N = 17,771).
Batts, K., Pemberton, M., Bose, J., Weimer, B., Henderson, L., Penne, M., & Strashny, A. (2014). Comparing and evaluating substance use treatment utilization estimates from the National Survey on Drug Use and Health and other data sources. In CBHSQ data review. (pp. 1–120). Substance Abuse and Mental Health Services Administration (US).
Buzza, C., Elser, A., & Seal, J. (2019). A mobile buprenorphine treatment program for homeless patients with opioid use disorder. Psychiatric Services (Washington, D.C.), 70(7), 635–636. doi:https://doi.org/10.1176/appi.ps.70701.
Davidson, C., Neighbors, C., Hall, G., Hogue, A., Cho, R., Kutner, B., & Morgenstern, J. (2014). Association of housing first implementation and key outcomes among homeless persons with problematic substance use. Psychiatric Services (Washington, D.C.), 65(11), 1318–1324. doi:https://doi.org/10.1176/appi.ps.201300195.
Kariisa, M., Scholl, L., Wilson, N., Seth, P., & Hoots, B. (2019). Drug overdose deaths involving cocaine and psychostimulants with abuse potential - United States, 2003-2017. MMWR. Morbidity and Mortality Weekly Report, 68(17), 388–395. doi:10.15585/mmwr.mm6817a3.
Lippold, K. M., Jones, C. M., Olsen, E. O., & Giroir, B. P. (2019). Racial/ethnic and age group differences in opioid and synthetic opioid-involved overdose deaths among adults aged ≥18 years in metropolitan areas - United States, 2015-2017. MMWR. Morbidity and Mortality Weekly Report, 68(43), 967–973. doi:10.15585/mmwr.mm6843a3.
Palamar, J. J., Han, B. H., & Keyes, K. M. (2020). Trends in characteristics of individuals who use methamphetamine in the United States, 2015-2018. Drug and Alcohol Dependence, 213, 108089. Advance online publication.
Singh, J. A., & Cleveland, J. D. (2020). National U.S. time-trends in opioid use disorder hospitalizations and associated healthcare utilization and mortality. PLoS One, 15(2), e0229174. doi:https://doi.org/10.1371/journal.pone.0229174.
Wilson, N., Kariisa, M., Seth, P., Smith, H., 4th, & Davis, N. L. (2020). Drug and opioid-involved overdose deaths - United States, 2017-2018. MMWR. Morbidity and Mortality Weekly Report, 69(11), 290–297. doi:10.15585/mmwr.mm6911a4.
00230-0&id=gr1.jpg){kind=link}