Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2025 Nov 1.
Published in final edited form as: J Subst Use Addict Treat. 2024 Aug 15;166:209494. doi: 10.1016/j.josat.2024.209494

Housing and Supportive Services for Young Mothers Experiencing Substance Use Disorder and Homelessness: Cost-Effectiveness Analysis of a Randomized Trial

Tansel Yilmazer 1, Jing Zhang 2, Laura Chavez 3, Ruri Famelia 1, Xin Feng 1, Jodi Ford 4, Kelly Kelleher 3, Natasha Slesnick 1
PMCID: PMC11392601  NIHMSID: NIHMS2018659  PMID: 39153732

Abstract

Background:

Mothers experiencing homelessness and caring for young children struggle with high rates of substance use and mental health problems. A comprehensive supportive housing intervention was implemented to assist young mothers experiencing substance use disorder (SUD) and homelessness. The cost-effectiveness of this intensive intervention could inform future dissemination.

Methods:

A cost-effectiveness analysis was conducted alongside a randomized controlled trial that lasted from May 2015 to October 2018. Mothers experiencing homelessness between the ages of 18-24 years with a SUD were randomly assigned to housing+support services (HOU+SS) (n=80), housing-only (HOU) (n=80), or services as usual SAU (n=80). Using incremental cost-effectiveness ratios (ICERs), the study compared the costs of HOU+SS and HOU to SAU for three outcomes: housing stability (percent days of stable housing), substance use (percent days of substance use), and depressive symptoms (Beck Depression Inventory score). Direct intervention costs of HOU+SS and HOU from both payor and societal perspectives were estimated. Cost data were collected from detailed study financial records. Outcomes were taken from 6-month assessments.

Results:

The average societal cost of HOU+SS per participant was $5,114 [CI 95%, $4,949-5,278], while the average societal cost of HOU was $3,248 [CI 95%, $3,248-$3,341] (2019 U.S. dollars). The calculated ICERs show that HOU was more cost-effective than HOU+SS and SAU for housing outcome. For illicit drug use, HOU+SS was more cost-effective than HOU. Finally, for depressive symptoms, neither HOU+SS or HOU were more cost effective than SAU.

Conclusion:

While HOU is more cost-effective for increasing housing, HOU+SS is more cost-effective for reducing illicit drug use. However, housing without improvements in substance use may not be sustainable, and supportive services are likely essential for improved well-being overall beyond the housing outcome alone.

1. Background

On any given day, 650,000 people experience homelessness in the U.S., including 60,000 families with children (Sousa et al., 2023). Families with children are the fastest growing segment of the homeless population, roughly accounting for one-third of people experiencing homelessness. Out of 186,000 people in families experiencing homelessness, 108,190 (58 percent) were under age 18, and 13,289 (7 percent) were between 18 and 24 years of age in 2023. A majority of these 18- to 24-year-olds in families were parents (54 percent). Around 40 percent of young women experiencing homelessness report either being pregnant or a parent (Dworsky et al., 2018). Homelessness is a significant problem that affects not only the mother, but also her children. Mothers experiencing homelessness tend to be young and single, and struggle with significant substance use and mental health problems (Anderson et. al., 2024; Crawford et al., 2011; Bassuk & Beardslee, 2014). Many of their children experience high levels of internalizing and externalizing problems, physical health problems, developmental delays, and challenges meeting educational goals (Clark et al., 2019; Sandel et al., 2015).

As communities confront the crisis of homelessness, they must grapple with inadequate resources and limited information to guide decision making regarding what interventions are effective for improving outcomes among families experiencing homelessness. The effects of independent housing and Housing First interventions on substance use are varied (Xiang, 2013). Evaluations of independent housing interventions that provide housing alone (without integrated therapy or other supportive services) to adults and youth experiencing homelessness are mixed (Kirst et. al, 2015; Baxter et al., 2019; Slesnick et al., 2023; Wang et al., 2019; Guo et al., 2016), suggesting that additional supportive services may be needed for meaningful improvements in outcomes. Although there is some evidence to suggest cost-effectiveness of housing plus services for chronically homeless adults with serious mental illness (Latimer et al., 2019; Rosenheck et al., 2003), these supportive housing interventions have not been tested among young mothers. Young mothers experiencing homelessness and substance use disorder (SUD) need additional support to reduce substance use (Bassuk & Beardslee, 2014) and care for their children. Importantly, we are aware of no prior study that has assessed the cost-effectiveness of supportive housing interventions for mothers with a SUD experiencing homelessness.

Further complicating the development of interventions for young mothers experiencing homelessness, many young mothers are hesitant to seek services from traditional service settings (Reppond & Bullock, 2020; Phipps et. al, 2019). Few young mothers experiencing homelessness report they have ever stayed at a crisis shelter, accessed mental health services, or received treatment for a SUD (Huhn & Dunn, 2020; Azim, MacGillivray & Heise, 2018). Young mothers report being the targets of stereotyping, surveillance, and devaluation in family shelters (Reppond & Bullock, 2020). Most family shelters have governing rules to regulate behaviors of the mothers and children, such as household chores, curfews, visitor limits, child discipline rules, no overnight passes, strict child and adult bedtimes, limited phone use, and required class attendance (Williams, 2016). Most of the time these regulations are constraining and compel young mothers to relinquish autonomy and control over parental practices. Thus, a comprehensive intervention that can be offered outside of the shelter setting may provide greater acceptance among most young mothers and also would be beneficial for communities that do not have shelters available.

The current study examined costs associated with a comprehensive intervention that included housing, supportive services, and substance use treatment. See Slesnick et al. (2023) for the details of the intervention. An intervention directed at the multiple domains impacting these families might be necessary to prevent chronic homelessness, substance use and mental health problems (Slesnick & Erdem, 2013). The main hypothesis of the study was that mothers assigned to housing and supportive services (HOU+SS) would show greater improvements in housing stability, substance use, and mental health than those assigned to the housing-only (HOU) or services as usual (SAU) condition. Consequently, we expected HOU+SS to be cost-effective because it was designed to augment HOU condition by adding strengths-based outreach and advocacy (SBOA), HIV prevention, and substance use and mental health counseling using the Community Reinforcement Approach (CRA, Meyers & Smith, 1995). Likewise, we expected it to improve upon SAU outcomes by providing quicker housing and supportive services for young families.

1.1. Objectives

The objectives of this study were to i) conduct a cost analysis of HOU+SS and HOU conditions; ii) estimate incremental cost-effectiveness ratios (ICERs) for three major outcomes (housing stability, substance use, and depressive symptoms) for HOU+SS and HOU compared to SAU; and iii) compare the costs and benefits of HOU+SS and HOU conditions.

2. Methods

2.1. Participants and Procedure

The study recruited young mothers from a large midwestern city through advertisements at various shelters, a drop-in center, and other agencies serving youth experiencing homelessness. Mothers were eligible to be included if they met the following criteria: i) they were between 18 to 24 years, ii) they were currently experiencing homelessness, as defined by the federal McKinney-Vento Act (2002), iii) they had physical custody of a biological child who was six years old or younger, and (4) they met criteria for a SUD as determined by the Structured Clinical Interview for DSM-5 Disorders (SCID-5) (First, et al., 2015). Two hundred forty mothers were eligible for the study, consented to participate, and were randomly assigned to HOU+SS (n=80), HOU (n=80), or SAU (n=80) conditions. All women were recruited over 42 months from May 2015 to December 2018.

2.2. Description of Intervention

HOU+SS is a comprehensive intervention that integrates housing, SBOA, HIV prevention and CRA. Three months of rental assistance in addition to 6 months of supportive services were provided. The therapist worked with the mother to identify appropriate housing among the available choices and initiated procedures for payments directly to the landlord once housing was identified. The project covered damage deposit, application fees, utilities, and paid rent to landlords at the beginning of each month.

SBOA, HIV prevention and CRA sessions were simultaneously delivered to each mother assigned to HOU+SS by the same therapist. Up to 26 SBOA and 20 CRA sessions were offered over a period of 6 months. The SBOA component focused on assisting clients with meeting their basic needs (i.e., referrals to food pantries), obtaining government entitlements (i.e., SSDI/SSI, cash assistance, food stamps, publicly funded childcare), and connecting to other needed supports (education, childcare, job training). The CRA component explored the use of substances and aimed to reinforce non-substance using, adaptive behaviors through communication skills training, relapse prevention and refusal skills training. Participants had access to their therapist 24/7, for non-session support services. This contact included face-to-face contact, phone calls, text messages, and Facebook messages.

HOU included 3 months of rental assistance and other housing expenses without supportive services. The mothers in HOU and SAU continued to receive usual services offered by the homeless services community including the local drop-in center, which provides food, laundry and shower facilities, counseling, case management, as well as recreational activities. Like other drop-in centers, this center does not provide housing. Upon request, however, standard practice for the drop-in staff is to work with the young adults to link them to housing agencies and other services in the community.

2.3. Cost Analysis

We adhered to the CHEERS 2022 reporting checklist to perform the cost analysis (Husereau et al., 2022). A health analysis plan was developed at the beginning of the study and implemented during the intervention. Cost data were collected from detailed staffing and financial records retained by the project coordinator and project investigators. We calculated cost from two perspectives, payor and societal. In this study, the payor perspective sums up all costs paid by the funding agency, the National Institute on Drug Abuse (NIDA). In this study, NIDA acted as a public housing agency providing supportive housing costs. Societal perspectives add up other cost items including client time, transportation cost, productivity, and education to payor perspective (Neumann et al., 2016). Since most clients were not in school, and transportation costs of clients were also paid by the trial, we only focus on client time and productivity in our calculations of societal perspective. We utilized activity-based costing (ABC) to calculate the costs of each condition (Neumann et al., 2016).

The main components of cost were personnel costs, housing and utilities assistance, materials, and mileage. Personnel costs included personnel time for i) screening and recruiting, ii) SBOA, HIV prevention and CRA sessions, iii) non-session support services, iv) supervision of therapists, and v) administrative work. Four types of personnel were employed in the project: a licensed clinical psychologist supervising the therapists; a project manager maintaining financial records, coordinating scheduling, and supervising the assistants; six therapists providing the treatment to the mothers; and two assistants searching for rental units. Hourly wage rates and fringe benefits (10.5% of the hourly wage for therapists and assistants, and 28.5% of the rate for supervisor and project coordinator) for the personnel were obtained from financial records. Activity times for the psychologist, project manager and assistants were obtained from financial records while activity times for therapists were obtained from individual-level time logs for each study participant. Time included all participant contacts with intervention staff through varied formats (e.g., face-to-face, calls, texts, social media messages). The costs of delivering the program did not include any donated services or volunteer effort as these did not apply.

Monthly rental and utility assistance that each mother received were recorded on financial records. To calculate the societal cost including productivity and client time, we added the mother’s cost of time spent in support services (mother’s time * minimum hourly wage) to the payor’s cost. Average costs for each treatment type were calculated by dividing the total costs over the project period by the number of participants in each group. All costs incurred only for research purposes were excluded. The intervention lasted from May 2015 to October 2018. Discounting for labor cost was not applied due to the study’s short time horizon. However, rental and utility assistance were recorded at their real rate. Costs are reported in 2019 U.S. dollars.

2.4. Cost-effectiveness analysis

Incremental cost-effectiveness ratios (ICERs) were calculated as the difference in mean cost (C) per participant of each condition divided by the difference in mean effectiveness (E) of each condition, (C1-C2)/(E1-E2) (Neumann et al., 2016). Since the trial had three conditions (HOU+SS, HOU, and SAU), three ICERs were calculated: ICERHOU+SS, SAU, ICERHOU, SAU. ICERHOU+SS, HOU. The mean effectiveness (Ei) is calculated as difference in means of study outcomes at 6-months and baseline among the participants in that condition. For example, EHOU+SS for housing is mean of housing outcome of HOU+SS condition at 6-months minus mean of housing outcome of HOU+SS condition at baseline. The mean effectiveness for substance use and depressive symptoms were calculated in a similar way as the difference of the outcome at 6-months and baseline. To address the uncertainty surrounding ICERs which is a point estimate, non-parametric bootstrap methods were utilized to produce 1,000 replicates of mean costs. Each replicate generated a unique ICER, and the standard errors were calculated using 1,000 ICERs.

The main outcomes were housing stability (percent days of stable housing), substance use (percent days of substance use) and depressive symptoms (Beck Depression Inventory score). Housing stability and substance use were reported as the percent days. We calculated the percent days housed (ranging 0 to 100) as the total number of days housed over the total number of days in the assessment period. We calculated the percent days substance (ranging 0 to 100) as the total number of days substance use over the total number of days in the assessment period. The assessment period at baseline was 90 days before the start of the study while the assessment period for the 6-month assessment was the number of days since the 3-month assessment. The substance use measures were the percent days the mother reported i) any substance use (alcohol and/or illicit drugs, excluding tobacco), ii) alcohol, and iii) illicit drugs. The measures of substance use were based on Form 90 semi-structured interview at baseline and the 6-month assessments (Miller, 1996). The depressive symptoms were measured by the Beck Depression Inventory II (BDI-II; Beck et al., 1996), which is the most frequently used self-report instrument for assessment of mood, cognitive and somatic aspects of depression and has shown good psychometric properties (Miner, 1991; Maxwell, 1992). Depressive symptoms were current measures at the time of the assessment and ranged from 0 to 63, where a higher number indicates more depressive symptoms. Little’s MCAR test (Little, 1988) showed that data were missing completely at random.

3. Results

Table 1 shows the bootstrapped average intervention costs for HOU+SS and HOU. It cost on average of 4,893 per participant [CI at 95%, 4,740-5,047] to deliver HOU+SS condition and $3,248 per participant [CI 95%, 3,140-3,354] to deliver HOU condition. The societal total cost of HOU+SS was an additional $217 (s.d. $146) dollars. The per participant number of session and non-session meetings with therapists are also reported on Table 1. The average number of SBOA sessions was 9.8 (with the range of 2-33), the average number of CRA sessions was 7.8 (with the range of 0-18), and the average number of HIV prevention sessions was 1.4 (with the range of 0-2). On average there were 22 non-session face-to-face contacts with the therapist (with the range of 0-325), 300 text messages (with the range of 11-870), 47 social media messages (with the range of 0-598), and 38 non-session phone calls (with the range of 0-166). These non-session interactions are included in the total cost.

Table 1:

Cost of HOU+SS and HOU with Young, Substance Abusing Mothers Experiencing Homelessness

HOU+SS n=76 HOU n=72 SAU n=80
Recruiting to the intervention $3 $3 $3
Screening for eligibility $11 $11 $11
Housing and utilities assistance* $3,033 $2,952 $0
Session support services $463 $0 $0
  Number of sessions [range]
  SBOA 9.8 [2-33] . .
  HIV prevention 1.4 [0-2] . .
  CRA 7.8 [0-18] . .
Non-session support services $520 $0 $0
  Number of sessions [range]
  Non-session face-to-face 22.0 [0-325] . .
  Text messages 300.4 [11-870] . .
  Social media messages 47.0 [0-598] . .
  Non-session phone calls 38.2 [0-166] . .
Supervision of therapists $337 $0 $0
Administrative cost $315 $200 $29
Materials, mileage etc. $197 $82 $5
 
Payor total cost 4,893 (782) $3,248 (547)a $48
   CF at 95% [4,740-5,047] [3,140-3,354]
 
Client time $217 (146) $0 $0
Societal total cost 5,114 (837) $3,248 (546)a $48
   CF at 95% [4,949-5,278] [3,140-3,354]
Open in a new tab

Notes: The table presents averages (per participant). The costs are reported in 2019 dollars. Bootstrapped standard deviations reported for payor and societal total cost in parentheses.

The unit cost for therapist time is $20.5, assistant time is $19, supervisor time is $125, project coordinator time is $45, and mother’s time is $9. The fringe benefits 10.5% for therapists and assistants, and 28.5% for supervisor and project coordinator are added to the unit costs. The mileage rate is $0.565 per mile. SBOA refers to case management, which focuses on assisting mothers to meet basic needs, obtain government entitlement and connect with other supports. CRA refers to community reinforcement approach to substance use which is substance use/mental health counseling,

a

the difference between the cost of HOU+SS and HOU is significantly different (p<0.001).

*

Out of 80, 76 mothers completed the HOU+SS intervention and 72 mothers completed HOU intervention. For HOU+SS, payor total cost, client time and societal total cost are means of 76 observations. For HOU, payor total cost and societal total cost are means of 72 observations.

Among mothers in the study sample, 84.6 percent identified themselves as Black (203 out of 240). The average age was 21.5, with a range from 18 to 24 years. Half of the mothers did not have a high school degree (117 out of 240), while the other half had a high school degree or GED (113 out of 240). The participants had on average 1.6 children ranging from 1 child to 5 children. The follow-up rate at 6-months was 95 percent for HOU+SS (76 out of 80), 94 percent for HOU (75 out of 80) and 86 percent (69 out of 80) for SAU. Little’s MCAR test was not significant suggesting no real differences across conditions in missing data (Slesnick et al., 2023).

Table 2 presents the main outcomes utilized in the cost-effectiveness analysis. The assessment period at 6-months was around 97 days (s.d. 35) since the last assessment. The percent days that the mothers were housed improved for all three conditions from the baseline assessment to 6-months: from 16.3 percent days to 83.7 percent days for HOU+SS condition, from 8.7 percent days to 74.6 percent days for HOU condition, and from 10.2 percent days to 38.1 percent days for SAU condition. At 6-months, the percent days the mothers had housing in HOU+SS and HOU conditions were significantly higher than mothers in SAU condition. While the percent days the mothers had housing was higher in HOU+SS than HOU, the difference was not significant. The percent days that the mothers used any substances (alcohol and/or illicit drugs, except tobacco) decreased for all three conditions from the baseline assessment to 6 months: from 82.6 percent days to 61.2 percent days for HOU+SS, from 81.7 percent days to 75.2 percent days for HOU, and from 90.0 percent days to 68.3 days for SAU. At 6-months, the percent days the mothers used any substance in HOU+SS was significantly lower than the mothers in the HOU condition (p<0.05). The percent days that the participants used any illicit drugs also decreased for all three conditions-- from baseline to 6-months: from 79.7 percent days to 58.7 percent days for HOU+SS, from 80.0 percent days to 72.7 percent days for HOU, and from 87.1 percent days to 65.8 percent days for SAU. At 6-months, the percent days the mothers used any illicit drugs in HOU+SS was significantly lower than the mothers in HOU condition (p<0.05). However, the percent days that the participants had alcohol was not significantly different across three conditions at 6-months (8.1 percent days for HOU+SS, 8.8 for percent days for HOU and 11.9 percent days for SAU). Finally, depressive symptoms decreased in all three conditions from the baseline assessment to 6-months: from 19.5 to 8.3 for HOU+SS, 18.8 to 9.7 for HOU, and 18.7 to 10.4 days for SAU. However, we did not detect any significant differences for depressive symptoms across three conditions at 6-months.

Table 2:

Summary Statistics of Main Outcomes

HOU+SS Mean (SD) n=80 HOU Mean (SD) n=80 SAU Mean (SD) n=80
Percent days housed (%)
Baseline 16.3 (31.5) 8.7 (24.4) 10.2 (29.2)
6-m follow-up 83.7 74.6 38.1 b,c
(37.6) (42.4) (54.8)
n=76 n=75 n=69
Percent days any substance use (except tobacco) (%)
Baseline 82.6 81.7 90.0 b,c
(30.9) (27.3) (21.6)
6-m follow-up 61.2 75.2 a 68.3
(46.4) (43.7) (48.1)
n=76 n=74 n=69
Percent days alcohol use (%)
Baseline 21.0 20.6 27.8
(31.9) (32.9) (37.6)
6-m follow-up 8.1 8.8 11.9
(16.8) (18.1) (27.7)
n=76 n=74 n=69
Percent days illicit drug use (%)
Baseline 79.7 80.0 87.1
(36.4) (30.2) (26.7)
6-m follow-up 58.7 72.7a 65.8
(41.0) (39.6) (42.2)
n=76 n=74 n=69
Depressive symptoms (BDI-II)
Baseline 19.5 18.8 18.7
(13.6) (12.5) (12.0)
6-m follow-up 8.3 9.7 10.4
(10.2) (12.9) (12.1)
n=73 n=72 n=68
Open in a new tab

Notes: The table presents mean values with bootstrapped standard deviations in parenthesis. Percent days are over the assessment period which are on average 97 days. The number of observations is reported under each outcome if it is different than n=80.

a.

The means of HOU+SS and HOU are different (p<0.05)

b.

The means of HOU+SS and SAU are different (p<0.05)

c.

The means of HOU and SAU are different (p<0.05)

Table 3 presents the ICERs for housing, substance use and depressive symptoms. The lower incremental values imply that the cost of achieving a unit improvement in our three main outcomes are smaller. We used societal cost in our calculation of ICERs since societal cost includes both payor cost and young mothers’ opportunity cost for participating in SS. In the calculation of bootstrapped standard deviations of the ICER ratios, only costs of HOU+SS and HOU are bootstrapped. An improvement for housing is one unit increase in percent days housed. For substance use, one unit decrease in percent days substance use is considered an improvement. Finally, for depressive symptoms, an improvement is one unit decrease in BDI-II. First, we compared the ICER ratios for percent days housed. ICERHOU+SS, SAU for one percent (days) increase in housing was $128. This amount was statistically larger than ICERHOU, SAU ($84). Similarly, ICERHOU+SS, HOU ($1,244) was statistically larger than ICERHOU, SAU ($84). Second, ICERHOU+SS, HOU for one percent (days) decrease in substance use was $125 for all substance use (except tobacco), $1,696 for alcohol and $136 for illicit drugs. The values of ICERHOU+SS, SAU and ICERHOU, SAU for substance use are not reported since HOU+SS and HOU were not more effective in reducing substance use compared to SAU. Finally, ICERHOU+SS, SAU for depressive symptoms was $1,746 and ICERHOU+SS, HOU was $888, both of which were statistically lower than ICERHOU, SAU ($3,999). In summary, HOU was more cost-effective than HOU+SS for housing. HOU+SS condition is more cost-effective than HOU for reducing illicit drug use. Based on the alcohol use not being significantly different across three conditions at 6-month follow-up, neither HOU+SS or HOU were more cost effective than SAU for reducing alcohol use. Lastly, based on depressive symptoms not being significantly different across the three conditions at 6-month follow-up, neither HOU+SS or HOU were more cost effective than SAU for reducing depressive symptoms.

Table 3:

Incremental Cost-effectiveness Ratios (ICERs)

HOU+SS vs. SAU Mean (SD) HOU vs. SAU Mean (SD) HOU+SS vs. HOU Mean (SD)
Percent day housed
6-m follow-up $128 (20) $84 (13) a $1,244 (66) b.c
[124–132] [81–86] [1,114-1,374]
Percent days any substance use (except tobacco) (%)
6-m follow-up $125 (68)
[$111-$138]
Percent days alcohol use (%)
6-m follow-up $1,696 (937)
[$1,513-$1,880]
Percent days illicit drug use (%)
6-m follow-up $136 (73)
[$121-$150]
Depressive symptoms (BDI-II)
6-m follow-up $1,746 (286) $3,999 (680) a $888 (470) b.c
[$1,690-$1,803] [$3,865-$4,133] [$796-$981]
Open in a new tab

Notes: The societal cost is used in the calculation of ICERs. Bootstrapped standard deviations are in parenthesis, and 95% are confidence intervals in brackets. Percent day housed is the cost of one percent day increase of being housed. Percent day substance use is the cost of one percent day decrease in substance use. Depressive symptoms is the cost of one-unit decline in BDI-II.

a.

ICER HOU+SS, SAU and ICER HOU, SAU are different (p<0.05)

b.

ICER HOU+SS, SAU and ICER HOU+SS, HOU are different (p<0.05)

c.

ICER HOU, SAU and ICER HOU+SS, HOU are different (p<0.05)

Sensitivity Analysis.

Obtaining activity times for therapists from individual-level time logs for each study participant provided the lower bound for the support services ($463 for session support services and $520 for non-session support services). Each therapist worked with 6 mothers in each month. We calculated the upper bound of cost of services from the financial records: 6 months* $20.5 per hour and 10.8% fringe benefits*80 hours per month divided by 6 mothers equals $1,670. When the upper bound is used in the cost calculations, societal cost for HOU+SS increased by $687 from $5,114 to $5,801. We recalculated ICERs using the upper bound of the societal cost. The main conclusions from Table 3 remained the same: HOU+SS condition was more cost-effective than HOU for substance use. For housing, however, HOU was more cost-effective than HOU+SS.

4. Discussion

The study represents the first formal economic evaluation of HOU+SS among young mothers experiencing SUD and homelessness, and one of few economic evaluations of similar supportive housing programs for populations experiencing homelessness (Latimer et al., 2019; Rosenheck et al., 2003). The societal total cost of HOU was $3,248, and the societal cost of HOU+SS was $1,866 higher. The estimated cost of one percent day increase in housing was lower with HOU ($84) than HOU+SS ($128). At the same time, the cost for reducing illicit drug use with HOU+SS was lower than HOU. Results suggest that while HOU+SS was not a more cost-effective strategy for moving mothers and their children off the streets and into stable housing than HOU, it was a cost-effective strategy for reducing illicit drug use. There are currently no established willingness-to-pay thresholds for our short-term outcomes including stable housing and illicit drug use for this population of young mothers. Therefore, we do not know what decision-makers consider a reasonable relationship between these two outcomes and their costs. However, housing without improvements in substance use may not be sustainable, and supportive services are likely essential for improved well-being overall beyond the housing outcome alone. Finally, results also suggest than HOU and HOU+SS were not more cost-effective than SAU for alcohol use and depressive symptoms. Future research should evaluate other interventions focusing on reducing high alcohol intake and depressive symptoms among these young mothers.

Several limitations should be considered. First, those mothers assigned to SAU continued to seek assistance from the community and drop-in center, providing a more conservative comparison to the active comparison conditions than a wait-list control. Some motivation for receiving assistance might have affected the outcomes. Consequently, the mothers in SAU also experienced improvements in housing and substance use. The service logs show that mothers in SAU were able to secure housing through other sources. In particular, the participants in this study, including those in SAU, were connected to the local drop-in center if they were not already connected. Mothers were provided a referral list to other homeless, substance use, and mental health services available in the community including family shelters. This reduced the differences in outcomes across our three conditions and increased the standard errors of ICERs. Second, the study used convenience sampling engaging participants from a large midwestern community who might not represent samples from other locations around the country. Finally, our cost-effectiveness analysis does not account for short or long term averted medical or societal costs as the results of the intervention. Future research containing data about criminal legal system involvement or emergency medical service use reductions should include those averted costs into the cost-effectiveness analysis. More comprehensive cost analysis will potentially reveal that the supportive interventions produce more averted costs.

However, there were also several strengths. Prior studies engage study participants from local crisis shelters (Krahn et al., 2018), skewing findings to those willing to access shelter services. As this study recruited women from a range of locations, the sample is likely more representative of those who do not access shelters for help. Furthermore, we engaged a relatively large sample of mothers, and had high engagement and retention in service activities, as well as high follow-up rates, enhancing confidence in the internal validity of study findings.

In conclusion, while offering only housing is more cost-effective for increasing housing, offering support services alongside housing is more cost-effective for reducing illicit drug use. Illicit drug use is closely related to being housed. Therefore, decreasing substance use might improve housing among this group of young mothers in the long run, making HOU+SS more cost-effective than HOU for housing as well.

Highlights.

  • HOU condition is more cost-effective than HOU+SS and SAU for increasing housing among young mothers experiencing SUD and homelessness.

  • HOU+SS condition is more cost-effective than HOU for reducing illicit drug use.

  • Neither HOU+SS or HOU are more cost effective than SAU for reducing alcohol use.

  • Neither HOU+SS or HOU are more cost effective than SAU for reducing depressive symptoms.

Funding

The study is funded by the National Institute on Drug Abuse (NIDA) (NIH Grant# 2 R01 DA023908-05A1). The sponsor had no role in the design of the study or writing of the manuscript.

This study was funded by NIDA grant # 2R01DA036589 awarded to the last author.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Ethics Approval and Consent to Participate

The study was reviewed and approved by the Ohio State University Institutional Review Board. The study obtained written informed consent from all research participants.

Consent for Publication

Not applicable.

Availability of Data and Materials

Not applicable.

Competing Interests

The authors declare that they have no competing interests.

References

  1. Anderson J, Trevella C, & Burn AM (2024). Interventions to improve the mental health of women experiencing homelessness: A systematic review of the literature. Plos one, 19(4), e0297865. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Azim KAA, MacGillivray L & Heise D (2019) Mothering in the margin: a narrative inquiry of women with children in a homeless shelter, Journal of Social Distress and Homelessness, 28(1) 34–43. [Google Scholar]
  3. Bassuk EL, & Beardslee WR (2014). Depression in homeless mothers: Addressing an unrecognized public health issue. American Journal of Orthopsychiatry, 84(1), 73–81. [DOI] [PubMed] [Google Scholar]
  4. Baxter AJ, Tweed EJ, Katikireddi SV, & Thomson H (2019). Effects of Housing First approaches on health and well-being of adults who are homeless or at risk of homelessness: systematic review and meta-analysis of randomised controlled trials. J Epidemiol Community Health, 73(5), 379–387. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Beck AT, Ward CH, Mendelson M, Mock J, & Erbaugh J (1996). An inventory for measuring depression. General Psychiatry, 4, 561–571. [DOI] [PubMed] [Google Scholar]
  6. Clark RE, Weinreb L, Flahive JM, & Seifert RW (2019). Infants exposed to homelessness: health, health care use, and health spending from birth to age six. Health Affairs, 38(5), 721–728. [DOI] [PubMed] [Google Scholar]
  7. Crawford DM, Trotter EC, Hartshorn KJS, & Whitbeck LB (2011). Pregnancy and mental health of young homeless women. American Journal of Orthopsychiatry, 81(2), 173. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Dworsky A, Morton MH & Samuels GM (2018). Missed Opportunities: Pregnant and Parenting Youth Experiencing Homelessness in America. ED603004. Accessed on April 10, 2024 at https://eric.ed.gov/?id=ED603004
  9. First MB, Williams JBW, Karg RS, & Spitzer RL (2016). User’s guide for the SCID-5-CV Structured Clinical Interview for DSM-5® disorders: Clinical version. American Psychiatric Publishing, Inc. [Google Scholar]
  10. Huhn AS, & Dunn KE (2020). Challenges for women entering treatment for opioid use disorder. Current Psychiatry Reports 22, 76. [DOI] [PubMed] [Google Scholar]
  11. Husereau D, Drummond M, Augustovski F, de Bekker-Grob E, Briggs AH, Carswell C, Caulley L, Chaiyakunapruk N, Greenberg D, Loder E, Mauskopf J, Mullins CD, Petrou S, Pwu RF, Staniszewska S; CHEERS 2022 ISPOR Good Research Practices Task Force. Consolidated Health Economic Evaluation Reporting Standards 2022 (CHEERS 2022) Statement: Updated Reporting Guidance for Health Economic Evaluations. [DOI] [PubMed]
  12. Guo X, Slesnick N, & Feng X (2016). Housing and support services with homeless mothers: Benefits to the mother and her children. Community Mental Health Journal, 52(1), 73–83. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Kirst M, Zerger S, Misir V, Hwang S, & Stergiopoulos V (2015). The impact of a Housing First randomized controlled trial on substance use problems among homeless individuals with mental illness. Drug and Alcohol Dependence, 146, 24–29. [DOI] [PubMed] [Google Scholar]
  14. Krahn J, Caine V, Chaw-Kant J, & Singh AE (2018). Housing interventions for homeless, pregnant/parenting women with addictions: A systematic review. Journal of Social Distress and the Homeless, 27(1), 75–88. [Google Scholar]
  15. Latimer EA, Rabouin D, Cao Z, Ly A, Powell G, Adair CE, … & At Home/Chez Soi Investigators. (2019). Cost-effectiveness of Housing First intervention with intensive case management compared with treatment as usual for homeless adults with mental illness: secondary analysis of a randomized clinical trial. JAMA network open, 2(8), e199782–e199782. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Maxwell BE (1992). Hostility, depression, and self-esteem among troubled and homeless adolescents in crisis. Journal of Youth and Adolescence, 21(2), 139–150. [DOI] [PubMed] [Google Scholar]
  17. Meyers RJ, & Smith JE (1995). Clinical guide to alcohol treatment: The Community Reinforcement Approach. New York: Guilford Press. [Google Scholar]
  18. McKinney-Vento Homeless Assistance Act, Re-Authorized (2002). 42 U.S.C. 11431 et seq 725
  19. Miller WR (1996). Form 90 a structured assessment interview for drinking and related problem behaviors. Project MATCH Monograph Series (Vol. 5). Bethesda, MD: U.S. Dept. of Health. [Google Scholar]
  20. Miner MH (1991). The self-concept of homeless adolescents. Journal of Youth and Adolescence, 20 (5), 545–560. [DOI] [PubMed] [Google Scholar]
  21. Neumann PJ, Sanders GD, Russell LB, Siegel JE, & Ganiats TG (2016). Cost-Effectiveness in Health and Medicine. 2nd ed. New York: Oxford University Press. [Google Scholar]
  22. Reppond HA, & Bullock HE (2020). Reclaiming “good motherhood”: US mothers’ critical resistance in family homeless shelters. Feminism & Psychology, 30(1), 100–120. [Google Scholar]
  23. Phipps M, Dalton L, Maxwell H, & Cleary M (2019). Women and homelessness, a complex multidimensional issue: Findings from a scoping review. Journal of Social Distress and the Homeless, 28(1), 1–13. [Google Scholar]
  24. Rosenheck R, Kasprow W, Frisman L, & Liu-Mares W (2003). Cost-effectiveness of supported housing for homeless persons with mental illness. Archives of general psychiatry, 60(9), 940–951. [DOI] [PubMed] [Google Scholar]
  25. Sandel M, Sheward R, & Sturtevant L (2015). Compounding Stress. The Timing and Duration Effects of Homelessness on Children’s Health. Center for Housing Policy. Accessed on April 10, 2024 at https://www.childrenshealthwatch.org/wp-content/uploads/Compounding-Stress_2015.pdf [Google Scholar]
  26. Slesnick N, & Erdem G (2013). Efficacy of ecologically-based treatment with substance-abusing homeless mothers: substance use and housing outcomes. Journal of Substance Abuse Treatment, 45(5), 416–425. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Slesnick N & Tonigan JS (2004). Assessment of alcohol and other drug use by runaway Youths. Alcoholism Treatment Quarterly, 22(2), 21–34. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Slesnick N, Feng X, Brakenhoff B, Guo X, Carmona J, Murnan A, et al. (2016). A test of outreach and drop-in linkage versus shelter linkage for connecting homeless youth to services. Prevention Science, 17, 450–460. [DOI] [PubMed] [Google Scholar]
  29. Slesnick N, Zhang J, Feng X, Mallory A, Martin J, Famelia R, Brakenhoff B, Yilmazer T, Wu Q, Ford J, Holowacz E, Jaderlund S, Hatsu I, Chavez L, Luthy E, Walsh L, & Kelleher K (2023). Housing and supportive services for substance use and self-efficacy among young mothers experiencing homelessness: Comparisons among three intervention conditions. Journal of Substance Use Treatment, 144, 108917. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Tanya de Sousa, Andrichik Alyssa, Prestera Ed, Rush Katherine, Tano Colette, and Wheeler Micaiah (2023). The 2023 Annual Homelessness Assessment Report (AHAR) to Congress. Accessed on April 10, 2024 at https://www.huduser.gov/portal/sites/default/files/pdf/2023-AHAR-Part-1.pdf
  31. Wang JZ, Mott S, Magwood O, Mathew C, Mclellan A, Kpade V, Gaba P, Kozloff N, Pottie K and Andermann A (2019). The impact of interventions for youth experiencing homelessness on housing, mental health, substance use, and family cohesion: a systematic review. BMC Public Health, 19(1), 1–22. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Williams JC (2016) A roof over my head: Homeless women and the shelter industry, Boulder, CO: University Press of Colorado. [Google Scholar]
  33. Xiang X. (2013). A review of interventions for substance use among homeless youth. Research on Social Work Practice, 25(1), 34–45. [Google Scholar]

RESOURCES