Reasons for Leaving Sober Living Houses and their Associations with Substance Use, Psychiatric, and Housing Outcomes

Abstract

Background:

Sober living houses (SLHs) offer abstinence-based housing for individuals with alcohol or drug problems. Research shows residents of SLHs make improvements on measures of substance use and other problems. Length of stay (LOS) in the house is associated with outcomes. However, little is known about the reasons residents leave houses or how reasons are associated with outcomes.

Methods:

Baseline interviews were conducted with individuals entering 48 SLHs in Los Angeles. Residents who left within 12 months completed a 12-month follow-up interview (N=352) assessing reasons they left, abstinence, psychiatric symptoms, and housing status. Multilevel regression models assessed relationships between reasons and outcomes.

Results:

The most commonly cited reason for leaving was wanting to live on one’s own or in a shared living situation with family or friends (45.7%), which was associated with increased odds of abstinence and fewer psychiatric symptoms. Leaving due to not liking something about the house (e.g., rules, or living conditions) was the second most commonly cited reason for leaving (28.1%), but was not associated with any outcomes. Wanting to live on one’s own and did not get along with residents or the house manager were associated with favorable housing (i.e., stable housing or a different SLH).

Conclusion:

Residents who leave SLHs to live more independently in the community may be better prepared than residents who leave for other reasons. Some residents may benefit from seeking stable housing in the community or in another SLH if they are not a good match for their current house.

Introduction

Substance use service providers and researchers increasingly recognize that some persons seeking help for alcohol and drug problems require long-term, stable housing during, after, or in some cases in lieu of professional treatment (Saitz, et al., 2008). Recovery residences, such as SLHs, can be good options for many of these individuals because they offer peer support within an environment that prohibits the use of alcohol and drugs. SLHs use a social model approach to recovery that emphasizes peer support, resident input into decisions affecting the household, and involvement in mutual help recovery programs such as Alcoholics Anonymous. Residents are expected to cover costs for rent and fees, although some receive help from their families or criminal justice institutions. To reduce rent and discourage isolating behaviors most residents share bedrooms for at least some minimal period of time after entering the house (Wittman, Polcin & Sheridan, 2017).

Outcome Studies

SLH outcome studies have documented significant, sustained decreases in substance use and improvements in other areas, such as employment, psychiatric symptoms, and legal issues (Polcin et al., 2010; Witbrodt, Polcin, Korcha & Li, 2019). When stable living situations in the community have been compared with residence in SLHs, there are better substance use outcomes in the SLH group and worse outcomes for persons who are homeless or living in unstable living situations (Polcin & Korcha, 2017). A few studies have examined factors associated with successful transition into the community, such as recovery capital (e.g., Polcin et al., 2023). However, no studies of SLHs have examined how reasons for leaving were associated with outcomes.

Length of Stay

The length of time in treatment for substance use disorders is known to be associated with outcomes. However, many residential treatment programs have strict limits on how long residents can stay. The National Institute in Drug Abuse recommends a period of at least three months of treatment to stop or significantly reduce substance use (Volkow, 2011).

A few studies have looked at the association between length of stay (LOS) and outcomes in peer operated recovery homes, such as Sober Living and Oxford Houses. Importantly, most peer operated recovery homes allow residents to stay at the houses as long as they wish, provided they comply with house rules. Only recently have studies assessed how LOS for persons residing in these houses is associated with outcomes. One study assessing LOS in sober living houses (SLHs) was conducted by Subbaraman, Mahoney, Mericle and Polcin (2023). They found residents who stayed in the house six months or longer had better outcomes. Similarly, Jason et al (2021) found better outcomes for residents who stayed in Oxford Houses for at least six months. Outcomes for the Subbaraman et al (2023) study included dichotomous measures of DSM-5 substance use disorders (yes/no), psychiatric symptoms (psychiatric diagnostic screening questionnaire [PDSQ]), and legal problems (Addiction Severity Index [ASI]).

The studies cited above show LOS is an important factor associated with outcomes and they support the need to examine ways to improve retention to six months or longer. However, the best recommendations for LOS at peer operated recovery homes may not be a one-size-fits-all solution. The circumstances under which residents leave may be critically important in terms of successfully transitioning into the community and maintaining a program of recovery. Subbaraman et al (2023) emphasized that more research on SLHs is needed to inform service providers about risk and protective factors related to leaving the houses. The current study builds on the Subbaraman et al (2023) study by investigating how outcomes might be associated with residents’ reasons for leaving SLHs (i.e., potential risk and protective factors) and the appropriateness of plans for housing and sustained recovery in the community. We suggest the current investigation of reasons for leaving SLHs will add to and complement current LOS studies.

Purpose

While it is important to understand resident readiness to transition out of SLHs in terms of LOS, other factors should be considered as well. The circumstances under which residents leave might be an indication of their readiness to live more independently. The current study aimed to examine whether the reasons residents left SLHs were associated with outcomes after accounting for LOS. We considered reasons that reflected motivation to move toward new, more desirable living situations in the community as well as reasons that reflected a desire to move away from something negative about the house. Examples of negative experiences included conflict with the house manager or other residents, not liking the structure or rules of the house, unable to pay rent and fees, and relapse while living in the house.

We hypothesized that leaving as a desire to live on one’s own or in a shared living situation with family/friends would be associated with favorable outcomes at 12-month follow-up, including abstinence, fewer psychiatric symptoms, and stable living arrangements. In contrast, leaving to escape negative characteristics and experiences at the house was hypothesized to be associated with worse outcomes on measures of substance use, psychiatric symptoms, and housing stability. We assumed outcomes would be worse when residents left the house to escape negative experiences, such as violations of house rules (e.g., relapse), inability to pay rent, or conflicts with the house manager or residents. Under these circumstances, we thought residents might be more likely to leave impulsively and in ways that endangered their recovery.

Methods

Study Sites and Participant Recruitment

Participants (N=352) were recruited from 48 SLHs in economically diverse areas of Los Angeles County. Data collection sites included houses in West Los Angeles (18.8%), Central Los Angeles (20.8%), South Bay/Long Beach (31.3%), and the San Gabriel/San Fernando Valley (35.5%). The distribution of houses by gender included 24 houses for men, 11 for women, and 13 for all genders. In terms of economic status of the neighborhoods, houses were selected to include the lowest quartile (27.1%), the second quartile (20.8), the third quartile (27.1%), and highest quartile (25%). All houses were members of the Sober Living Network (SLN), which is an association of SLHs mostly located in southern California. SLN provides certification, consultation, and advocacy to member houses that comply with standards for health, safety, good neighbor relations, and good business practices.

After SLHs agreed to participate, an interviewer was assigned to the house to monitor new residents as they were admitted and inform them about the study. Recruitment methods included posting flyers, presenting the study at house meetings, making informal visits to the SLH to interact with residents, adding a study information sheet to the SLH’s intake packet, and encouraging prior study participants to share their experiences with new residents. Those who were interested in participating met with the assigned interviewer who screened participants over the phone or in-person and explained the study in greater detail. Prospective study participants were informed that they were expected to attend a 12-month follow-up interview even if they left the house. Those who signed informed consent documents completed the baseline interview.

Inclusion/Exclusion

Inclusion/exclusion criteria included: age 18 or older, entered the house within 30 days of the baseline interview, history of drug or alcohol problems, and were willing to provide informed consent and contact information for the 12-month follow-up interview.

Procedures

Study participants consisted of a subsample of a larger study assessing 12-month outcomes for SLH residents (N=456). The 12-month follow-up rate was 81.9% and 12-month interviews were conducted on average at 383 days (SD=25.9) from entry into the house. There were no significant differences in baseline measures of outcomes for those who completed the 12-month interview versus those who were lost at follow-up. The larger study used an intent-to-treat model that aimed to complete 12-month interviews with all residents, including those still living in the SLH as well as those who had left before the 12-month interview. Because we were studying reasons for leaving SLHs, the analyses for the current study included only residents who had left the houses before the final 12-month interview (N=352).

The Public Health Institute IRB reviewed and approved the study protocol. Participants received $30 for completing the baseline interview and $50 for completing the 12-month interview. It was emphasized that residency in the SLH was not contingent upon participation in the study. Data collection interviews were initially conducted in-person but switched to phone interviews during years three and four of the study to comply with COVID-19 safety protocols.

Measures

1. Demographics: Baseline data from self-reports of age, sex, and race.

2. Reason for leaving: Participants were presented with a list of potential reasons for leaving the SLH. They could respond yes/no for each. Because our interest was to identify the salience of reasons across all residents participating in the study, we allowed participants to endorse as many different reasons as they wanted. This strategy enabled identification of a broad range of reasons that could be quantified based on their frequency across all study participants. Reasons were categorized as:

   1. Wanted to live on one’s own, or a shared living arrangement with friends, or with family. This category captured those who moved out primarily as a way to establish independent living in the community.

   2. Did not like aspects of the SLH – Examples included the rules, structure, neighborhood where the house was located, and the living conditions at the house. The category did not include those who reported having problems with other residents or the sober living house manager.

   3. Relapse – This category was for those who reported they left because had used substances or planned to use after leaving. Some residents left before using because they wanted to stay on good terms with the SLH so they moved out prior to the relapse.

   4. Did not get along with people at SLH – This category was comprised of those who left because they did not get along with the other residents, the house manager, and/or the owner of the SLH.

   5. Could no longer pay/program ended – Payment of house fees is a requirement for living at the SLH, so those who could no longer pay were captured in this category. Some residents had been paying from their own savings that eventually ran out. Others received financial help from organizations or institutions that paid fees for a limited time period. Most often these were criminal justice institutions. Cost was combined with the program ending or closing as a reason for leaving because both reflected external issues largely out of the resident’s control.

   The original list of reasons for leaving consisted of a larger number of categories than those presented here. We combined some reasons for leaving to increase statistical power and we felt the combined categories were conceptually similar.

3. Length of stay (LOS) was calculated as the number of days residing in the house between the date of entry and the date the resident left.

4. Substance use was assessed using the Timeline Follow back (TLFB) (Sobell et al., 1996). At each interview, a calendar was used to help people provide retrospective estimates of their daily alcohol and drug use. The interviewer collected TLFBs for all substances used in the prior six months for each participant at baseline and at 12 months. Because the data distribution was highly skewed, we dichotomized this variable as any use versus none. The assessment of abstinence was also viewed as important because it is the primary goal of SLH residences.

5. The Psychiatric Diagnostic Screening Questionnaire (PDSQ) was used to measure psychiatric symptomology. Zimmerman and Mattia (1999) created this screening instrument, which was comprised of 115 items. Cronbach’s alphas were computed and were greater than 0.80 for 12 of the 13 psychiatric disorders assessed. The instrument also includes an overall score for psychiatric symptoms, which was used for the analyses reported here. Higher scores indicate greater psychological distress. Psychiatric symptoms were used as an outcome variable because it is an important longitudinal aspect of recovery (Booth et al., 2010), and elevated levels may indicate an increased risk for relapse (Erga et al., 2021).

6. Housing status after leaving the SLH used categories described by Polcin and Korcha (2017), which showed housing status was related to substance use and psychiatric outcomes. Categories included: 1) stably housed (owned or rented a residence with or without housemates), 2) living at a SLH, 3) unstable housing (criminal justice incarceration, inpatient treatment, or temporary housing in some other institution) or temporarily staying with friends or family) and 4) homeless (shelter or homeless). Because relatively few participants were homeless, we combined homeless with unstably housed. Housing status before entering the SLH was collected during the baseline interview. At the 12-month interview, we asked about the first housing situation after they left the SLH.

Analysis

Initial analyses identified the percent of residents that indicated each reason as influencing their decision to leave the SLH. Bivariate analyses included chi square procedures to assess relationships between dichotomous outcomes and reasons for leaving. Pearson correlations were used to assess associations between psychiatric symptoms (assessed as a continuous count varisble) and reasons for leaving.

To examine how reasons for leaving were associated with outcomes we then ran multivariable generalized estimating equation (GEE) and multinomial regression models that adjusted for time, LOS, age, sex, and race/ethnicity. Substance use was operationalized as a dichotomous measure, no use versus any use over the past 6 months. Logistic GEE regression models were used to analyze substance use. Psychiatric severity was measured using global score on the PDSQ which was left as a continuous variable. Linear GEE regression models were used to analyze psychiatric severity. Adjusting for baseline outcomes was achieved by using longitudinal generalized estimating equation (GEE) models that included outcomes as time-varying variables and thus inherently adjusted for baseline values.

Housing status was operationalized as a 3-category scale: unstable, stable, and different SLH, with unstable as the referent. For housing outcomes, we used multinomial logistic regression models that controlled housing status at baseline. All regression models controlled intragroup correlations for SLHs and within-subjects effects.

All regression analyses began with each reason entered separately as a predictor for each of the outcomes: abstinence, psychiatric symptoms, and housing status. To assess the relative influence of reasons on each outcome we also ran simultaneous regression models. Reasons that were found to be significant (P<0.05) predictors of each outcome in the individual regression models were entered together in simultaneous models predicting each outcome. The “could not afford to stay/program ended” reason was not included in the simultaneous models because it did not predict any outcomes in the individual models. All analyses were conducted in Stata, version 17.0. The level of significance in all analyses was p<.05.

Results

Demographic characteristics of the sample (N=352) included 37.5% women, 50% nonwhite, and a mean age of 39.8 (SD=12.2). Prior to entering the SLH, 76% indicated their living situation was unstable, which included inpatient substance use treatment (30.8%), staying temporarily with family or friends (16.2), and homeless (15.7%). Sixteen percent of the participants indicated they were living in a different SLH; only 8% indicated they were living in some type of stable living situation, such as owning a house or renting an apartment. Slightly over a quarter (26.8%) indicated they had been abstinent for six months prior to the baseline interview.

Table 1 summarizes the reasons for leaving. Study participants were able to endorse multiple reasons. However, nearly all the study participants indicated one reason for leaving. The mean number of reasons was 1.10 (SD=0.89). The most frequently endorsed reason was wanting to live on their own or in a shared living arrangement with friends or with family (45.7%). The next three most frequently cited reasons were all endorsed by a little over a quarter of the residents who left. These included not liking something about the SLH, relapse, and not getting along with other residents or the house manager. The least frequently cited reason for leaving was that they could no longer pay (16.8%).

Table 1.

Percentages of Sober Living House (SLH) Residents for Reason for Leaving SLH

Reasons for Leaving SLHa (%)	12-Month completers who left SLH
	N=351
Wanted to live on own or with family/friends**	54.3
Did not like something about SLH - structure, rules, neighborhood, living conditions (not people)	28.1
Relapse*	26.7
Did not get along with other residents, house manager, and/or owner**	26.7
Could not afford to stay/program ended/insurance ran out	16.8

Note.

^aAt follow-ups 12 months after baseline, participants who left the SLH were asked “Why did you leave the sober living house?”

Table 2 shows bivariate associations between reasons for leaving and outcome variables. Inability to pay rent and fees as a reason for leaving was not associated with any of the outcomes. In contrast, wanting to leave to live on one’s own or in shared living with family or friends was associated with all of the outcomes in the hypothesized direction (i.e., favorable outcomes). Table 2 also shows that relapse as a reason for leaving was associated with housing outcomes in expected directions (e.g., positive association with homeless/marginal housing). Not liking something about the house and not getting along with other residents or the house manager were both associated with increased housing at a different SLH after they left.

Table 2.

Associations between each reason and each outcome (Abstinence, PDSQ, Post-SLH Housing)

Reason for Leaving	Abstinent for prior 6 months (dichotomized)c	Psychiatric symptoms count P	Stable Housing c	Another SLH c	Marginal/Homeless c
Wanted to live on own or with family/friends	4.24 *	−0.15 **	70.60 ***	16.70 ***	17.44 ***
Did not like something about SLH - structure, rules, neighborhood, living conditions (not people)	0.04	−0.07	4.62 *	11.12 **	0.63
Relapse	4.00 *	0.01	27.81 ***	0.99	30.40 ***
Did not get along with other residents, house manager, and/or owner	2.99	0.10	0.54	9.63 **	3.49
Could not afford to stay/program ended/insurance ran out	0.38	−0.04	1.75	0.64	3.35

^cChi-square test statistic

^pPearson pairwise correlation coefficient

^***p<.001

^**p<.01

^*p<.05

Table 3 shows regression models assessing each reason for leaving as a separate predictor of each outcome. The first column in the table shows odds ratios (ORs) from GEE models with reasons for leaving as predictors of abstinence from alcohol and drugs at 12-month follow-up. Leaving the house to live on one’s own or sharing a living space with family or friends was associated with higher odds of abstinence at 12 months (OR=1.78, 95% CI 1.25 – 2.53), p<.001). Relapse was associated with lower odds of abstinence at 12 months (OR=0.44, 95% CI 0.28 – 0.68, p<.001).

Table 3.

Separate Regression Models Assessing Reasons for Leaving SLHs as Predictors of Abstinence, Psychiatric Symptoms, and Housing

	GEE Models		Multinomial logistic regression
	6-month Abstinence vs. Not	PDSQ psychiatric symptoms count	Stable housing vs. Unstable housing	Another SLH vs. Unstable housing
Reason for leaving	OR (95% CI LL, UL)	IRR (95% CI LL, UL)	RRR (95% CI LL, UL)	RRR (95% CI LL, UL)
Wanted to live on own or with family/friends	1.78 (1.25, 2.53) **	0.99 (0.98, 0.99) ***	8.32 (4.63, 14.95) ***	0.56 (0.30, 1.05)
Did not like something about SLH environment	0.94 (0.65, 1.36)	1.00 (1.00, 1.01)	0.67 (0.34, 1.34)	2.22 (1.15, 4.27) *
Relapse	0.44 (0.28, 0.68) ***	1.00 (1.00, 1.01)	0.11 (0.05, 0.24) ***	0.45 (0.22, 0.91) *
Did not get along with other people at SLH	0.72 (0.49, 1.05)	1.00 (1.00, 1.01)	1.12 (0.66, 1.90)	2.64 (1.47, 4.76) **
Could not afford to stay/program ended/insurance ran out	0.92 (0.56, 1.52)	1.00 (0.99, 1.00)	0.54 (0.26, 1.11)	0.64 (0.27, 1.53)

Note. Separate regression models were used to assess each reason as a predictor of each outcome.

GEE models show odds ratios (ORs, 95% CI) to depict reasons for leaving as predictors of abstinence.

Incident Risk Ratios (IRRs) show results for negative binomial regression models for the counts of psychiatric symptoms on the Psychiatric Diagnostic Screening Questionnaire (PDSQ).

Housing outcomes are reported as relative risk ratios (RRRs) for multinomial logistic regression models with unstable housing as the reference group.

Models controlled for time, length of stay, age, sex, race/ethnicity, and baseline status for each outcome and adjusted for intragroup correlations for SLHs and within-subjects effects.

Bolded values indicate p<.05.

^***p<.001

^**p<.01

^*p<.05

The second column shows Incident Risk Ratios (IRRs) from GEE models assessing with reasons for leaving as predictors of psychiatric symptoms as measured on the total score of the PDSQ. The only significant predictor was wanting to live on own’s or shared living with friends or family, which was associated with fewer psychiatric symptoms (IRR=0.99, 95% CI 0.98 – 0.99, p<.001).

Table 3 also shows results from MLMs assessing how reasons for leaving were associated with housing status immediately after leaving the house. Unstable housing was used as a reference group for stable housing and living in a different SLH. Control variables included time, length of stay, age, sex, race/ethnicity and baseline status for each outcome. Analyses also adjusted for intragroup correlations for SLHs and within subjects effects.

Relative risk ratios showed all reasons except “could not afford to stay” were associated with housing status. Stable housing after leaving was more likely for residents whose reasons for leaving included wanting to live on their own or with friends or family (RRR= 8.32, 95% CI 4.63– 14.95, p<.001). Stable housing was less likely for residents whose reasons for leaving included relapse (RRR=0.11, 95% CI 0.05–0.24, p<.oo1). About 21% of the sample indicated their housing immediately after leaving the SLH environment was a transition to a different SLH. To the extent residents included “did not like something about the house” as a reason for leaving or “did not get along with other people” they were more likely to be living at another SLH after they left than an unstable housing situation. Persons who left because of a relapse were less likely to be living in a different SLH (RRR=0.45, 95% CI 0.22 −0.91, p<.05).

To assess the relative influences of reasons for leaving on outcomes we ran simultaneous GEE and MLM models (Table 4). For each outcome (abstinence, psychiatric symptoms, and housing status), reasons that were shown to be significant predictors in the individual models from Table 3 were entered together. Models included the same control variables as the separate regression models: time, length of stay, age, sex, race/ethnicity and baseline status for each outcome. Analyses also adjusted for intragroup correlations for SLHs and within subjects effects.

Table 4.

Simultaneous Regression Models Assessing Reasons for Leaving SLHs as Predictors of Abstinence, Psychiatric symptoms, and Housing

	GEE Model	Multinomial logistic regression
	6-Month Abstinence vs. Not	Stable housing vs. Unstable housing	Another SLH vs. Unstable housing
Reason for leaving	OR (95% CI LL, UL)	RRR (95% CI LL, UL)	RRR (95% CI LL, L)
Wanted to live on own or with family/friends	1.49 (1.04, 2.14) *	6.46 (3.28, 12.74) ***	0.54 (0.29, 0.99) *
Did not like something about SLH environment	-	0.59 (0.27, 1.29)	1.58 (0.82, 3.07)
Relapse	0.51 (0.33, 0,79) **	0.18 (0.08, 0.41) ***	0.45 (0.20, 1.00)
Did not get along with other people at SLH	-	2.04 (1.08, 3.85) *	2.12 (1.15, 3.93) *

Note. For each outcome, all reasons for leaving that were significant (p<.05) in the separate models depicted above in Table 3 were entered together into one model. In order to test relative influences, simultaneous models required inclusion of two or more reasons that were shown to be significant in the individual models. Therefore, the PDSQ was not assessed as part of the simultaneous models because it lacked two or more predictors. Odds ratios (ORs) show relationships between reasons and 6-month abstinence. Housing outcomes were assessed using relative risk ratios (RRRs) for multinomial logistic regression models with unstable housing as the reference group.

Models controlled for time, length of stay, age, sex, race/ethnicity, and baseline status for each outcome and adjusted for intragroup correlations for SLHs and within-subjects effects.

Bolded values indicate p<.05.

^***p<.001

^**p<.01

^*p<.05

Table 4 shows the reason with the strongest and most consistent relationships with favorable outcomes was wanting to live on own’s own or shared living with family or friends, which was related to higher odds of abstinence (OR=1.49, 95% CI 1.04–2.14, p<.05) and higher likelihood of stable housing (RRR=6.46, 95% CI 3.28–12.74, p<.001). However, there was also a modest association between wanting to live on one’s own and being less likely to live in a different SLH than in an unstable living situation (RRR= 0.54 (0.29, 0.99, p<.05). In contrast, residents indicating relapse as a reason for leaving the house had lower rates of abstinence (OR=0.51, CI 0.33–0.79, p<.01) and lower rates of stable housing (RRR=0.18, 95% CI 0.08–0.41, p<.001).

Two other reasons were also associated with housing outcomes. Residents who indicated they left because they did not get along with other residents or the manager were more likely to report having stable housing after they left (RRR=2.04, 95% CI 1.08–3.85, p<.01) or transition to a different SLH (RRR=2.12, 95% CI 1.15–3.93, p<.05).

DISCUSSION

SLHs offer persons with alcohol and drug problems an abstinence-based living environment that emphasizes a peer-oriented approach to recovery. Because there are no professional services offered, residents do not typically develop formal treatment plans that include strategies for successful transition out of the SLH and into independent living in the community. Moreover, residents and house managers who oversee house operations have few guidelines that can assist them in identifying when a resident is prepared to leave the house.

Recent research shows residents who remain in the house for more than six months have better outcomes than those staying less than six months (Subbaraman, et al., 2023). However, other factors may be important as well. The current study showed when LOS was controlled, significant relationships between reasons for leaving and outcomes remained. Findings from the current study can be used to provide a rationale for additional studies addressing reasons for leaving and other factors influencing transitions out of SLHs and other types of recovery homes.

Leaving to Improve Housing

When residents state they want to leave the house to live on their own or a shared living arrangement with family or friends it implies they want an improved, independent living situation. Nearly half of the sample included this as one of their reasons for leaving and it was significantly related to more favorable outcomes. This finding should provide a measure of reassurance to house managers and residents. In both the individual and simultaneous models this reason was associated with better abstinence, psychiatric, and housing outcomes. Residents who want to live alone or in shared housing with family or friends may have greater access to supportive living environments and other sources of recovery capital that enhance their chances of success. The importance of recovery capital during and after residency in SLHs is supported by research showing residents who relapse have better 12-month outcomes if they have higher levels of recovery capital (Polcin, Mahoney, Witbrodt, Subbaraman, & Mericle, 2023). The effects of recovery capital are strongest after residents leave. However, the ways recovery capital might influence or interact with reasons for leaving are unclear and require additional study. Additionally, understanding these relationships will require a broad approach that examines how different types of recovery capital affect and are affected by different reasons for leaving.

Leaving in Response to Problems

Other reasons for leaving SLHs reflected a desire to avoid negative characteristics of the house, other residents, or the house manager as well as negative experiences while living at the house (e.g., relapse). The finding that there was an association between relapse as a reason for leaving and poor outcomes was expected. There were several considerations supporting this hypothesis. First, relapse is inconsistent with the goal of abstinence, which is central to most SLHs. Second, SLH residents who relapse typically need to leave the house for a specified time period. That can increase the risk of further substance use and related problems. Some residents who relapse leave voluntarily before they are asked to leave. This is particularly common for residents who plan to return to a substance using lifestyle. Separate from reasons for leaving, studies have shown that residents who relapse have worse outcomes than abstainers (Bosso et al., 2021; Polcin, Mahoney et al., 2023).

We hypothesized that residents who left the house because they did not get along with other residents or the house manager would have worse outcomes. We reasoned that transitions out of the houses that were based on conflicts with residents or the house manager might be poorly planned and impulsive. For some of these residents, their focus might plausibly be on leaving and there may be insufficient attention paid to how they plan to continue their recovery and find adequate housing. However, study findings showed residents who left because they did not get along with other residents or the house manager did not have worse abstinence or psychiatric outcomes. Simultaneous regression models showed that these residents were more likely to find stable housing in the community or in a different SLH than in unstable living situations. Although these residents wanted to live separately from their peers and the house manager, they were able to achieve separation in a way that enabled them to transition more often to stable rather than unstable housing.

An important question is how residents who left the SLH due to relationship problems were able to succeed in finding stable living situations in the community. Research is needed to identify factors that facilitated and hindered successful transitions. However, it is possible that some of these residents learned how to avoid impulsive decisions and instead think through challenges by using problem solving skills and other recovery tools they developed while living in the SLH or attending mutual help recovery groups. Another factor could be direct help from SLH managers, senior peers in the house, or professionals in the surrounding community. These individuals might have provided information about housing options, currently available apartments, and housing resources in the community.

We also expected that residents who left because they did not like something about the house itself (e.g., structure, rules, living conditions, neighborhood) would have worse outcomes. We felt this could be an indication that the resident was struggling with their recovery or did not understand how structure, rules, and expectations related to recovery. We felt there was a risk that some could leave prematurely before they were ready. However, the simultaneous model showed disliking aspects of house operations or living conditions as a reason for leaving was unrelated to abstinence and psychiatric symptoms.

In the separate models, persons who left because they did not like aspects of the house were more likely to find housing in a different SLH than in an unstable housing situation. This could reflect flexibility on the part of some house managers in terms of recognizing that some residents may not be a good fit for their house and might be more successful in a house better aligned with their motivations and needs. For example, de Guzman, Korcha, and Polcin (2019) conducted qualitative interviews with SLH residents and found some residents were primarily interested in living in an SLH because they were interested in any type of affordable housing. Less important was the quality of the recovery environment. For others, a focus on recovery and strict enforcement of house rules was essential. Some house managers are skilled at recognizing these differences and they make referrals to other houses when they are deemed to be a better fit for the individual resident.

Inability to Meet Costs

The least endorsed reason for leaving was the resident could not afford the cost to stay in the house (16.8%). For some, funding from the criminal justice system reached a maximum and the resident could not afford to pay on their own. For others, family support decreased and for others they were not able to earn enough through work to meet costs. However, in addition to being a relatively less cited reason for leaving, the frequency of cost as a reason for leaving was not associated with any outcomes. This was a somewhat surprising finding because cost is often cited as a concern by house managers and residents and a recent study cited affordability as a frequent reason for selecting a specific SLH (Mahoney, Subbaraman, Mericle, & Polcin, 2023). .

Preparing Residents to Leave

The finding that some reasons for leaving SLHs are associated with favorable outcomes and others associated with poor outcome raises the question of how SLH providers can use these findings to improve SLH services. Although discharge planning is a routine component of substance use treatment (Reif, et al., 2014) it is not common in SLHs. Most house managers do not have any systematic way to work with residents to plan for their life after leaving the SLH, let alone plan for residents who may be asked to leave for rule violations or for those who may be leaving before they might be ready. This type of planning may be critical to ensuring that gains made within the SLH environment are maintained and that residents are provided with resources to re-engage with treatment and recovery support services if they need them after leaving.

Although SLHs do not have formal mechanisms for developing transition plans, there are a few ways the social model environment in these homes can be leveraged to address transitions. One way of systematically addressing transitions is to make it a standard part of house meetings. For example, during a house meeting the house manager could ask, “how many of you are thinking about leaving over the next month? Are you open to sharing your plans with your peers and receiving feedback?”

Another possibility is to address transitions into new living situations in the community within the context of an alumni group. Persons who left the house and successfully adapted to a new living environment could share their experiences and provide feedback to current residents about their plans to leave. These activities and interactions might also generate new interactions among current residents where they help each other by sharing personal experiences. Examples of potential benefits might include emotional support and encouragement as well as ideas for dealing with challenges and setbacks.

Housing Choice: Peer Operated Recovery Homes and Housing First

Our findings apply to SLHs but not necessarily to other types of recovery residences (e.g., Oxford Houses) or harm reduction housing (e.g., housing first). A few papers (e.g., Polcin, 2016; Wittman et al., 2017) address the strengths and weaknesses of HF and SLH approaches to helping persons with chronic homelessness and persistent psychiatric problems. The Wittman et al (2017) paper suggests that improving the availability and effectiveness of housing and related services (e.g., substance use) will require service providers to consider contrasts in housing models, including definitions of “recovery,” roles of peer support, facility management, roles for professional service, and the architectural designs that support the mission of each type of housing. Rather than advocate for one housing model as the best, the Wittman et al (2017) paper suggests more consideration of consumer needs, fit with housing options, and choices within homeless service systems. The Polcin (2016) paper identifies a variety of limitations in current studies assessing housing services models and makes suggestions for improvements.

Responding to Housing Crises

The need for a flexible, broad based approach to housing that considers individual needs is evident in the recent fires in Los Angeles. Although some residents had access to the financial resources necessary to find emergency housing and plan for longer term housing needs, others needed immediate assistance and the challenges of finding suitable housing increased for persons with longstanding problems such as homelessness, substance abuse, and persistent mental health disorders (Karlinsky et al, 2024).

Los Angeles will continue to grapple with its longstanding housing challenges combined with the recent effects of fires for years. One strategy to address these problems would be to adopt a broad based approach that recognizes potential roles for the expansion of low income housing and subsidized housing assistance. In addition, housing policies should include expansion of housing options that have evidence based support, including harm reduction, recovery housing, and housing services for mental health problems.

Limitations

One limitation of the study was that the measure of reasons for leaving allowed participants to select multiple responses. Therefore, reasons had equal importance and we were not able to analyze primary reasons for leaving. Further research should examine how primary reasons for leaving SLHs are related to outcomes. However, this concern was largely mitigated by the fact that nearly all of the study participants indicated only one reasons for leaving.

Selection bias due to non-random sampling was possible. All participants who were eligible and interested were able to participate, but selection bias could result from having the more motivated and engaged residents as participants. This study was also conducted in Los Angeles County; SLH residents in other areas or other types of recovery residences may have different considerations and experiences.

A final limitation involves the potential effects of a variety of confounding factors that were not accounted for in the current study, as well as the potential effects of mediators and moderators. We believe that the identification of important variables that should be included in new models will evolve sequentially as additional studies are conducted.

Conclusion

Findings from the current study suggest SLH service providers should aim to help residents remain in the house until they express a readiness to live on their own or find suitable housing with friends or family. Residents who reached that point had the best outcomes for stable housing after they left as well as for abstinence and psychiatric symptoms at 12-month follow-up. Consistent with numerous studies assessing alcohol and drug outcomes, we found relapse as a reason for leaving the house was associated with worse housing in the community (i.e., unstable) and lower rates of abstinence.

We hypothesized that leaving to avoid aspects of the house, house manager, or fellow residents would be associated with worse outcomes. However, results showed these reasons were not associated with worse abstinence or psychiatric status at 12-month follow-up. Residents who left because they could not get along with other residents or the house manager were more likely to end up in stable housing or in another SLH than in unstable housing. Research is needed to identify the processes that help residents avoid unstable housing and find successful transitions to more suitable and supportive independent housing or alternate SLHs. Helping residents remain in an SLH until they feel ready and able to transition to more independent living might be achieved by facilitating social model dynamics that build cohesion.