Effectiveness of a supportive housing program for homeless adults with mental illness and substance use: A two-group controlled trial

Abstract

Introduction

The purpose of this study was to assess the effectiveness of a housing transition program for homeless shelter residents with chronic mental illness and substance use.

Method

A two-group controlled study design was used to assess the program, with 10 participants in an intervention group and 10 in a control group. Goal attainment scaling and quality of life ratings were used pre- and post-intervention to determine if a statistically significant difference existed between groups at post-intervention.

Results

A statistically significant difference existed between intervention and control group goal attainment scaling scores (U = 9.50, p < .03, d = 1.34), and on quality of life scores (U = 10.50, p < .04, d = 1.30) at post-intervention. At a 6-month follow-up, 57.14% of intervention group participants had transitioned into supportive housing, while only 25% of control group participants had transitioned.

Conclusion

Findings suggest that intervention participants made greater progress toward desired housing goals and reported higher quality of life ratings than controls at post-intervention. This study provides support for the effectiveness of a housing program for homeless adults with chronic mental illness and substance use histories to achieve housing goals and higher quality of life ratings.

Keywords

Occupational therapy homeless housing transition independent living skills

Introduction

Homelessness is an epidemic international problem. In January 2015, 564,708 people were homeless on any given night in the United States (US) and an estimated 1.6 million used homeless shelters that year (US Department of Housing and Urban Development, 2015). In that same year, 85,460 British adults were reported homeless and applied for government housing assistance (Department for Communities and Local Government, 2016). Similarly, it was estimated that over 235,000 Canadians experienced homelessness in 2014, with approximately 35,000 homeless on any given night (Gaetz et al., 2014).

In the US, approximately one-third of homeless adults have severe chronic mental illness (for example, schizophrenia, bipolar disorder, major depression) (Treatment Advocacy Center, 2014) and two-thirds have a substance use disorder (Office of National Drug Control Policy, n.d.). Homeless adults with mental illness and substance use disorders are more likely to remain homeless and lose housing compared to those without these diagnoses (Substance Abuse and Mental Health Services Administration, 2016). Many self-care and home management skills that are typically learned during the transition from adolescence to adulthood (for example, bathing, hygiene, clothing care, apartment cleaning and organization, meal preparation, and money management) are not acquired by this homeless sub-group, since the onset of mental illness commonly occurs at this life stage and may impact the attainment of daily life skills. Further barriers to the attainment of daily life skills include sporadic periods of childhood and adolescent homelessness and foster care placements (Dworsky et al., 2013). If such skills were learned before or after early adulthood, they have often deteriorated as a result of untreated mental illness and periods of homelessness, hospitalization, or incarceration (Tsemberis et al., 2012).

In the US, homeless adults typically obtain shelter placements lasting several months to 1 year. During this time, case managers help residents to complete applications for federal and state subsidized permanent supportive housing (Tsemberis et al., 2012). Occupational therapists commonly do not provide services in either the shelter or supportive housing environment. Although supportive housing involves the provision of weekly case management and as-needed medical services, residents are expected to live fairly independently.

Many homeless adults with chronic mental illness and substance use histories proceed through the shelter system and receive supportive housing without acquiring the skills needed to manage physical and mental health disabilities, and safely function in and maintain apartment living (Greenwood et al., 2013). Once in supportive housing, the lack of home management and health maintenance skills can lead to problems such as hoarding, unsafe use of stoves and appliances causing fire safety hazards, failure to take prescribed medication resulting in illness events and rehospitalizations, and inability to budget subsidy payments for food and self-care (Khandor et al., 2011; Raphael-Greenfield and Gutman, 2015). Although some occupational therapists have been involved in the development of living skills with this population (Boisvert et al., 2008; Gutman et al., 2004; Helfrich and Fogg, 2007; Helfrich et al., 2006), there remains a critical need for therapists in the US to develop and assess interventions to help homeless adults with mental illness live optimally in supportive housing.

In 2015 we published initial data assessing a housing transition program called the SMART Program (Supporting Many to Achieve Residential Transition) (Gutman et al., 2015). The program was designed for homeless adults with mental illness and substance use histories and was developed by the authors in collaboration with a large urban housing agency. At that time, the SMART Program had three modules addressing the skills needed to (1) prepare for the housing interview and application process, (2) manage an apartment and participate safely and optimally in the community, and (3) understand what it means to be a good tenant and neighbor. The study was a retrospective analysis of program data reporting the feasibility and acceptability of the program’s implementation with 10 male participants in a 200-bed urban shelter. Based on feedback from participants and facilitators, we expanded the original three modules to six: (1) preparing for the housing interview and application process; (2) managing an apartment; (3) being a good tenant and neighbor; (4) living optimally in the community; (5) managing money; and (6) maintaining health and wellness (see Table 1).

Table 1.

Description of SMART (Supporting Many to Achieve Residential Transition) Program modules.

Module 1: housing interview skills	Module 2: apartment living skills
Many shelter residents must participate in a housing interview in which they are asked to demonstrate their readiness for transition to supportive housing. This module addresses the need to:	In this module, participants learn about and practice the skills needed to maintain supportive apartment living either independently or with a roommate:
• avoid substance use before the interview • use appropriate self-care and hygiene skills before the interview • arrive to the interview on time and in appropriate clothing • answer difficult questions about the client’s past (such as incarceration and substance use) • use appropriate eye contact, facial expressions, body language, and posture • advocate for oneself	• apartment cleaning strategies (including using appropriate cleaners and maintaining regularly scheduled cleaning sessions) • apartment organization strategies (for clothing, furniture, important documents, self-care products, and medication) • food storage in the refrigerator, freezer, and pantry • clothing laundering • rent and bill paying • getting along with roommates and negotiating conflict • using safety strategies in an apartment
Module 3: being a good tenant and neighbor	Module 4: c ommunity l iving
In the third module, participants learn about and practice the skills needed to (1) live cooperatively and peacefully with neighbors and (2) interact appropriately with supers and landlords: • understanding what it means to be a respectful neighbor: ○ maintaining the cleanliness of common areas ○ making sure that the building front entrance is always locked ○ refraining from allowing the entrance of unfamiliar people ○ keeping noise levels to a minimum ○ holding elevator doors open for neighbors • being able to appropriately introduce oneself to neighbors • understanding how to negotiate conflict with neighbors • understanding when it is appropriate to contact supers or landlords for help and how to interact with them • understanding that rent and utilities must be paid on time	Many clients feel isolated once they transition from the shelter to supportive housing. Because clients often experience post-traumatic stress disorder from homelessness and frequently have a past of conflicted relationships, many do not reintegrate into the community and remain isolated in their apartments. Module 4 helps clients: • explore their new neighborhood • identify free community activities of interest • learn safety skills to negotiate urban streets and subways
Module 5: m anaging m oney	Module 6: h ealth and w ellness
Many clients never learned money management skills and struggle with bill payments once they transition to supportive housing. Module 5 helps clients learn about and practice the basic skills of money management:	Many clients have not had adequate medical care throughout their lives and, as a result of homelessness, have prematurely aged and have chronic medical conditions that impact daily life. This module helps clients to understand:
• examining spending habits • adhering to a budget • modifying budgets when financial needs change • paying bills on time • learning saving strategies: ○ using bank saving and checking accounts ○ using store coupons and bonus saving cards ○ using thrift and second hand stores ○ using debit instead of credit cards • Learning strategies to protect oneself from identity theft and scams	• the components of health and wellbeing (physical, emotional, mental, spiritual) • how to maintain physical health through diet and exercise • how to facilitate restful sleep habits • the need for regular medical care and screenings to prevent and manage illness • managing chronic health conditions • medication management • the importance of smoking cessation • stress management • health literacy skills (understanding medication labels and accessing health information) • how to enact a wellness recovery action plan if and when mental health symptoms return

Each of the six SMART modules is based on hybrid education theories and multimodal learning. Hybrid education involves the use of both digital materials and instructor delivery of educational content (Biddix et al., 2015). We selected hybrid education because it is paced at the learner’s processing speed and allows the learner to integrate audiovisual learning, hands-on practice, and consultation with a face-to-face facilitator. Because many chronically homeless adults with mental health and substance use histories have cognitive problems and limited education (Hyman et al., 2011), we believed that hybrid education would best facilitate learning. Multimodal learning involves the use of visual, aural, and kinesthetic materials to enhance the educational process (Canazza and Foresti, 2013). Each SMART module uses digital screen images overlaid with printed words and verbal narration to support comprehension. Screen text was designed to be easily readable, with accompanying images that reflect similar demographics to those of program participants. Each 1-hour module can be paused so that a facilitator can engage participants in hands-on functional activities that allow the practice of skills presented in the modules; modules can be used with groups or individuals. The SMART Program has been manualized, and facilitator scripts, interventionist training procedures, fidelity checklists, and intervention guidelines were developed for all six modules. Although occupational therapists have been involved in the development and assessment of home management skills for homeless adults (Boisvert et al., 2008; Gutman et al., 2004; Helfrich and Fogg, 2007; Helfrich et al., 2006), we are unaware of the creation of digitally based programs combining multimodal and hybrid learning formats with hands-on practice for this population.

The purpose of this study was to assess program effectiveness with an intervention and control group selected from a convenience sample. Our research question asked whether participants who attended the SMART Program would have higher transition rates to supportive housing, attain greater progress toward apartment living goals, and report higher quality of life compared to controls who did not attend the program.

Method

Research design

This study assessed program effectiveness using a two-group, controlled trial in which one group attended the SMART Program and received treatment-as-usual, while the control group received only treatment-as-usual. The program was implemented over 6 weeks, using twice-weekly 1-hour sessions. Approval for program assessment was obtained by the Columbia University Medical Center (CUMC) institutional review board and shelter agency, and all participants provided written consent.

Participants

A convenience sample of 20 homeless adult men was recruited from a large New York City shelter over a 1-week period. Participant mental health diagnosis, ethnicity, and level of education were obtained from self-report and clinical chart audit by a case manager. Inclusion criteria were English speaking, interest, and availability to attend the program. The first 10 participants who met the inclusion criteria were enrolled. Control participants (n = 10) were enrolled based on interest in program participation but were unable to attend due to conflicting schedules. Both control and intervention arms received treatment-as-usual.

Instruments

Goal attainment scale (GAS) scores were used at pre- and post-intervention to both establish participant-selected goals and evaluate progress toward those goals at program end. In goal attainment scaling, participants work with a therapist to set individualized goals that are meaningful to the client (a list of example goal topics is shown in Table 2). Each goal is broken down into five possible outcomes on a scale ranging from −2 to +2, where 0 is equal to the participant’s expected functional level at post-intervention. Scores of +1 or +2 indicate that the participant made gains above the expected level, scores of −1 indicate no change in performance, and scores of −2 indicate performance regression (Kiresuk and Sherman, 1968; Ottenbacher and Cusick, 1990, 1993) (see Table 3). GAS scores allow large multiple-step performance skills to be broken down into smaller increments that can be evaluated. Because this population requires lengthy periods to make small changes, we believed that GAS scores would best detect and evaluate progress. GASs have been found to have high interrater reliability among trained raters (r = 0.88), and convergent validity with the Health–Sickness Rating Scale (r = 0.70, p < 0.001) and the Target Complaints Scale (r = .50, p < 0.01) (Shefler et al., 2001).

Table 2.

Examples of participant SMART (Supporting Many to Achieve Residential Transition) Program goal topics.

• I will show that I can prepare for the housing interview by showering, dressing appropriately, and avoiding any substance use on three consecutive days of the SMART Program.

• I will show that I can answer three difficult questions about my past in a role-played housing interview session.

• I will show that I understand how to keep my apartment clean by demonstrating three household cleaning strategies in the occupational therapy simulated apartment.

• I will show that I understand what to do if I have a conflict with my roommate or another resident by using three appropriate negotiation strategies in a role-played session.

• I will show that I can create a monthly budget that includes rent, food, bills, self-care items, and recreation.

• I will demonstrate that I am able to manage my medications by correctly placing pills in a weekly pillbox and using a check-off system to indicate that I have taken my daily medications.

Table 3.

Example of a participant goal attainment scale where 0 is the expected outcome level, −2 is the most unfavorable outcome, and +2 is the most favorable outcome.

+2 In a role-played housing interview, I was able to consistently maintain an upright posture and eye contact with the interviewer, and answer two questions about the housing skills I learned in the shelter. [Greater than expected outcome]

+1 In a role-played housing interview, I was able to consistently maintain an upright posture and eye contact with the interviewer, and answer one question about the housing skills I learned in the shelter. [Somewhat greater than expected outcome]

0 In a role-played housing interview, I was able to consistently maintain an upright posture and eye contact with the interviewer. [Expected outcome]

−1 In a role-played housing interview, I was able to maintain an upright posture and eye contact for approximately 50% of the interview. [Baseline performance]

−2 In a role-played housing interview, I was able to maintain an upright posture and eye contact for 25% or less of the interview. [Regression from baseline performance]

The Manchester Short Assessment of Quality of Life Scale (MANSA) was designed for community-based mental health populations and is a 16-item, seven-point Likert self-report scale (1 = could not be worse, 7 = could not be better) that can be completed in 10 minutes (Priebe et al., 1999). The scale has convergent validity with the Lancashire Quality of Life Scale (r = 0.83, p < 0.05) and moderate internal consistency with a Cronbach’s alpha of 0.74, p < 0.001. In this study the MANSA was completed by participants at pre- and post-intervention.

Intervention

Intervention description

The SMART Program was implemented as a 6-week, twice-weekly program; each of the six modules was delivered in 1-hour sessions over 2 days. We limited program participation to a convenience sample of 10 shelter residents, who were the first to consent to participation. This sample size was based on educational literature suggesting that small group sizes best facilitate learning, particularly when students have learning difficulties (Berkeley et al., 2011). Sessions occurred in the shelter’s occupational therapy room, which was equipped with a small simulated apartment. Program materials were delivered through a DVD format that could be paused at specific intervals to allow for hands-on practice of apartment living skills. Each module was manualized to enhance intervention fidelity and consisted of a facilitator instruction manual, appendix of practice skills and activities, facilitator training procedures, and fidelity checklists.

Facilitator

The facilitator, who was external to the research team, was an occupational therapist with 25 years of experience in mental health practice and was assisted by two graduate occupational therapy students carrying out Level II fieldwork. The facilitator received two hours of training in program implementation by the first author, along with weekly 1-hour consultations before and after each session. During the study, the facilitator provided intervention to SMART Program participants only and was blinded to group assignment.

Intervention fidelity

To assess intervention fidelity, a nine-item fidelity checklist was used to ensure that the facilitator implemented modules as intended. The checklist was completed by the first author once per week over the 6-week program to rate the facilitator’s congruence with manualized intervention guidelines. The facilitator averaged a 95% agreement with manualized instructions over the 6-week intervention.

Treatment-as-usual

Both control and intervention group participants received treatment-as-usual. Treatment-as-usual was provided by shelter staff (employees and volunteers of the shelter) and consisted of weekly case management, psychiatry, occupational therapy, recreational therapy, and medical and dental care. Occupational therapy consisted of group and individual sessions addressing resident housing goals (such as meal preparation, resume writing, self-care); cooking, exercise, current events, and music groups were run weekly. Treatment-as-usual occupational therapy services were provided by CUMC doctoral occupational therapy students and faculty (who were shelter volunteers and external to the research team).

Data collection

In the week before intervention, the second author (an occupational therapist) met individually with control and intervention group participants to conjointly develop and rank three personalized goals addressing the shelter transition and apartment maintenance process. At this time, the MANSA was also completed by each participant. One week post-intervention, the second author again met with each participant to determine progress made for each GAS. Participants completed post-intervention MANSA scales at this time. The second author was blinded to participant group assignment. At 6-month follow-up, the second author contacted each participant’s case manager to determine whether and when participants transitioned from the shelter to supportive apartment living.

Data analysis

T scores were used to determine if a difference existed between control and intervention groups on GAS scores at post-intervention. Using the formula below, pre- and post-intervention GAS scores were converted to T scores:

T = 50 + \frac{(10 \sum WiXi)}{\sqrt{(1 - r) \sum {Wi}^{2} + r (\sum Wi) 2}}

Wi represents each goal’s weighting, Xi represents each goal’s outcome score, and r represents the estimated average inter-correlation for each outcome score. An r of .30 is used as a constant in the formula (Ottenbacher and Cusick, 1990, 1993).

The T score is a standardized score with a mean of 50 and a standard deviation of 10. The standardization of T scores allows their comparison to a larger population. When T scores equal 50, they indicate a raw score of 0 on the original GAS (the participant’s expected level of performance). When T scores are greater than 50, they indicate that the participant performed above his or her expected level. When T scores are below 50 they indicate that the participant regressed in performance, below his or her expected level. The conversion of raw scores to T scores allows clearer interpretation of the participant’s overall score when multiple, separate goals are involved. T scores also allow the participant’s weighting of each goal to be factored into the final outcome score (Kiresuk and Sherman, 1968; Ottenbacher and Cusick, 1990, 1993). Once T scores were calculated for all participants, a Mann Whitney U test was used to determine if a statistically significant difference existed between control and intervention groups at post-intervention (Portney and Watkins, 2015).

A Mann Whitney U test was also used to determine if a statistically significant difference existed between control and intervention group MANSA scores at post-intervention (Portney and Watkins, 2015).

Data were analyzed using SPSS version 23 and significance level was set at α < 0.05. A power analysis suggested that with a sample size of 10 per group with a standard deviation of 5, we were powered to detect between-group differences of 6.6 points on GAS T scores with a power of 0.8 at α < 0.05.

Results

Twenty male participants were enrolled in the study, 10 in each group. Three participants were withdrawn from the intervention group due to rehospitalization (n = 2) and failure to attend program sessions (n = 1). Two participants were withdrawn from the control group due to rehospitalization. Fifteen participants completed the study (7 in the intervention group and 8 in the control group). Intervention group participants who attended nine out of 12 sessions (75%) were designated as study completers. Intervention group participants ranged in age from 29–59 (M = 43.4, SD = 10.16); control group participants ranged in age from 30–60 (M = 45.3, SD = 9.79). Both groups shared similar races/ethnicities, educational levels, and diagnoses (see Table 4).

Table 4.

Intervention and control group participant breakdown of demographics.

	Intervention group n = 7	Control group n = 8
Race/ethnicity
White	1 (14.3%)	1 (12.5%)
African American	4 (57.1%)	4 (50%)
Hispanic	2 (28.6%)	3 (37.5%)
Diagnosis
Schizophrenia	4 (57.1%)	4 (50%)
Depression	2 (28.6%)	2 (25%)
Bipolar disorder	1 (14.3%)	2 (25%)
Education
GED	1 (14.3%)	3 (37.5%)
Some high school	1 (14.3%)	0
High school	1 (14.3%)	2 (25%)
Some college	3 (42.9%)	1 (12.5%)
2-yr college degree	1 (14.3%)	1 (12.5%)
4-yr college degree	0	1 (12.5%)

GED: General Educational Development exam.

At pre-intervention, all participant-selected GAS scores were set at −1 (the participant’s present performance level). A Mann Whitney U test found no statistically significant difference between pre-intervention MANSA scores for both the intervention (M = 39.14, SD = 9.45) and control groups (M = 38.62, SD = 7.06), suggesting that there was no difference in their quality of life ratings.

At post-intervention, a Mann Whitney U test found a statistically significant difference between intervention (M = 65.73, SD = 4.66) and control group (M = 58.25, SD = 6.31) T scores (converted from GAS scores) (U = 9.50, p < 0.03) with a large effect size (d = 1.34). This suggests that at post-intervention, the intervention group participants reached higher performance levels on GAS scores than did the control group (see Table 5).

Table 5.

Participant goal attainment scale (GAS) and T scores.

Intervention participant #	Pre-intervention GAS	Post-intervention GAS	Weight	T score
1	−1	1	1	63.22
	−1	1	2
	−1	1	3
2	−1	2	3	74.23
	−1	2	2
	−1	1	1
3	−1	1	2	65.42
	−1	2	1
	−1	1	3
4	−1	1	3	69.82
	−1	2	1
	−1	2	2
5	−1	1	1	63.22
	−1	1	3
	−1	1	2
6	−1	2	2	61.01
	−1	0	3
	−1	1	1
7	−1	1	1	63.22
	−1	1	2
	−1	1	3
Control Participant #	Pre-Intervention GAS	Post-Intervention GAS	Weight	T score
1	−1	0	3	56.61
	−1	1	2
	−1	1	1
2	−1	0	3	52.20
	−1	0	2
	−1	1	1
3	−1	1	1	52.20
	−1	0	2
	−1	0	3
4	−1	0	3	58.81
	−1	1	2
	−1	2	1
5	−1	1	2	61.01
	−1	1	3
	−1	0	1
6	−1	1	1	69.82
	−1	2	3
	−1	1	2
7	−1	1	2	63.22
	−1	1	1
	−1	1	3
8	−1	0	3	52.20
	−1	0	2
	−1	1	1

Similarly, a Mann Whitney U test found a statistically significant difference between intervention (M = 47.85, SD = 12.30) and control group (M = 35.50, SD = 5.52) MANSA scores at program end (U = 10.50, p < 0.04) with a large effect size (d = 1.30). This suggests that intervention group participants rated their quality of life higher at program end than did control group participants (see Table 6).

Table 6.

Participant Manchester Short Assessment of Quality of Life Scale (MANSA) quality of life scores.

Intervention participant #	Pre-intervention MANSA	Post-intervention MANSA
1	26	31
2	43	65
3	32	37
4	56	62
5	41	45
6	36	46
7	40	49
	M = 39.14	M = 47.85
	SD = 9.45	SD = 12.30
Control participant #	Pre-intervention MANSA	Post-intervention MANSA
1	44	37
2	50	46
3	40	37
4	44	38
5	34	33
6	35	35
7	30	30
8	30	28
	M = 38.62	M = 35.5
	SD = 7.06	SD = 5.52

At a six-month follow-up, four (57.14%) intervention group participants had transitioned into supportive housing while only two (25%) of the control group participants had transitioned. A case manager reported that, at the 6-month follow-up, problems encountered by the intervention group participants who had transitioned into supportive housing included loneliness and anxiety. In addition to loneliness and anxiety, control group participants who had transitioned were also reported to have problems with hoarding and medication compliance.

Discussion

Our research question asked whether a group of homeless men with chronic mental illness and substance use histories who received the SMART Program intervention would have higher transition rates to supportive housing, attain greater progress toward apartment living skills, and report higher quality of life compared to control group participants who did not attend the program. Our data analysis found statistically significant differences between control and intervention group GAS T scores and MANSA scores, suggesting that the intervention group participants made greater progress toward apartment living skill goals and reported higher quality of life at intervention end. At a six-month follow-up, four (57.14%) intervention and two (25%) control group participants had transitioned to supportive housing, suggesting that SMART Program attendance may have helped intervention group participants better learn needed apartment living skills, report greater life satisfaction, and achieve supportive housing transition. Although both groups received treatment-as-usual, it appears that the addition of the SMART Program helped intervention group participants to make greater progress compared to controls.

Limitations

One study limitation was a small sample size and a 25% attrition rate. Our attrition rate was similar to or lower than other intervention studies with homeless adults having mental health and substance use histories (Hobden et al., 2011). Although this population, once sheltered, commonly receives medical and psychiatric health care, medication compliance is fragmentary and is often not required for shelter placement. Reduction of substance use is encouraged but is also difficult to achieve, particularly when substances such as K2 (synthetic cannabinoids) have become epidemic international problems driven by ready availability and inexpensiveness (Aikins et al., 2015). An acknowledged limitation of this study involved our inability to collect information about participants’ substance use.

A power analysis suggested that our sample size of 20 participants was sufficient to detect between-group differences at 6.6 points on GAS T scores—a power level that decreased after a 25% attrition rate. Yet, despite our small sample, our analysis still found a statistically significant difference between control and intervention group GAS T scores and MANSA scores at intervention end, suggesting that it was likely the SMART Program intervention that caused observed differences. Although we deliberately limited group size to 10 participants to enhance learning in a sample of adults with limited education and cognitive problems, we should have enrolled a minimum of 15 participants per group to achieve a study sample size of 20 after attrition.

Another limitation was a lack of randomization. We chose not to use a randomized waitlist control group because we were unable to run the SMART Program within 6 months of study completion. Although there was considerable interest in attending the SMART Program, we found that many residents were unable to attend the program at a regularly scheduled time. We therefore enrolled the first 10 residents who were both interested and able to attend. Residents who were interested but could not attend due to scheduling conflicts agreed to serve as controls. The lack of randomization limits the generalizability of the findings; however, all participants—both control and intervention—shared similar demographic variables which mirrored US national statistics on homeless adults with chronic mental illness and substance use disorders (US Department of Housing and Urban Development, 2015).

A final limitation involved the inability to collect outcome measures, other than transition rate, at a 6-month follow-up. Once shelter residents transitioned to supportive housing they were relocated at distances that were difficult to access without additional data collectors. As a result we were unable to collect 6-month post-intervention GAS and MANSA scores, which may have provided information about the SMART Program’s effectiveness over time.

Future research

Future research should involve larger sample sizes, randomization, and 6-month and 1-year follow-up measures. Future research should determine if the SMART Program would benefit formerly homeless adults now residing in supportive apartment living who have difficulty with community participation, and apartment, money, and chronic illness management. The SMART Program is currently being expanded to include 12 new modules addressing supportive education and employment for homeless adults who have an interest in and are able to resume student/worker roles. These modules will be assessed for effectiveness and satisfaction in future research. Access to all SMART modules will be available at no cost on the CUMC Occupational Therapy Program’s website.

Conclusion

This study found that a supportive housing program for homeless adults with chronic mental illness and substance use histories helped participants achieve housing goals and higher quality of life scores when compared to a control group who did not attend the program. Although occupational therapists in the US are increasingly providing services to this population, the effectiveness of our services must be documented to ensure service reimbursement and secure our role in this community-based arena.

Key findings

A large sub-group of the homeless population has chronic mental illness and substance use disorders that have contributed to a gap in knowledge about independent apartment living skills.

What the study has added

Occupational therapy-developed housing programs for adults with histories of homelessness, mental illness, and substance use can help this group attain needed skills in preparation for the transition to supportive housing.

Footnotes

Acknowledgments

We thank Lee Ann Westover, Joseph White, Etta Graham, and Lauren Zavier for their assistance in this work.

Research ethics

This study was approved by Columbia University Medical Center’s institutional review board and the ethics committee of the shelter agency. All participants provided consent.

Declaration of conflicting interests

The authors declare that there is no conflict of interest.

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

References

Aikins

Hoefinger

Guarino

(2015) Conducting rapid street assessment of drug users in New York City using oral fluid and brief interviews: A feasibility study. Journal of Addictive Diseases 34(2–3): 185–197.

Berkeley

Mastropieri

Scruggs

(2011) Reading comprehension strategy instruction and attribution retraining for secondary students with learning and other mild disabilities. Journal of Learning Disabilities 44(1): 18–32.

Biddix

Chung

Park

(2015) The hybrid shift: Evidencing a student-driven restructuring of the college classroom. Computers and Education 80: 162–175.

Boisvert

Martin

Grosek

(2008) Effectiveness of a peer-support community in addiction recovery: Participation as intervention. Occupational Therapy International 15(4): 205–220.

Canazza S and Foresti GL (2013) A multimodal learning system for individuals with sensorial, neuropsychological, and relational impairments. Journal of Sensors, 2013. Article ID 564864. Available at: http://dx.doi.org/10.1155/2013/564864 (accessed 6 June 2016).

Department for Communities and Local Government (2016) Statistical data set: Live tables on homelessness. Available at: www.gov.uk/government/statistical-data-sets/live-tables-on-homelessness#statutory-homelessness-live-tables (accessed 14 June 2016).

Dworsky

Napolitano

Courtney

(2013) Homelessness during the transition from foster care to adulthood. American Journal of Public Health 103(S2): S318–S323.

Gaetz S, Gulliver T and Richter T (2014) The state of homelessness in Canada 2014. Homeless Hub Paper #5. Available at: www.homelesshub.ca/sites/default/files/SOHC2014.pdf (accessed 14 June 2016).

Greenwood

Stefancic

Tsemberis

(2013) Pathways Housing First for homeless persons with psychiatric disabilities: Program innovation, research, and advocacy. Journal of Social Issues 69(4): 645–663.

10.

Gutman

Diamond

Holness-Parchment

(2004) Enhancing independence in women experiencing domestic violence and possible brain injury: An assessment of an occupational therapy intervention. Occupational Therapy in Mental Health 20(1): 49–79.

11.

Gutman SA, Raphael-Greenfield EI and Simon PM (2015) Feasibility and acceptability of a pilot housing transition program for homeless adults with mental illness and substance use. Occupational Therapy in Health Care 30(2): 124–138.

12.

Helfrich

Fogg

(2007) Outcomes of a life skills intervention for homeless adults with mental health illness. Journal of Primary Prevention 28(3–4): 313–326.

13.

Helfrich

Aviles

Badiani

(2006) Life skill interventions with homeless youth, domestic violence victims and adults with mental illness. Occupational Therapy in Health Care 20(3–4): 189–207.

14.

Hobden

Forney

Durham

(2011) Limiting attrition in longitudinal research on homeless adolescents: What works best? Journal of Community Psychology 39(4): 443–451.

15.

Hyman S, Aubry T and Klodawsky F (2011) Resilient educational outcomes: Participation in school by youth with histories of homelessness. Youth & Society 43(1): 253–273.

16.

Khandor

Mason

Chambers

(2011) Access to primary health care among homeless adults in Toronto, Canada: Results from the Street Health Survey. Open Medicine 5(2): 94–103. Available at: http://www.openmedicine.ca/article/viewArticle/368/407 (accessed 6 June 2016).

17.

Kiresuk

Sherman

MRE

(1968) Goal attainment scaling: A general method for evaluating comprehensive community mental health programs. Community Mental Health Journal 4(6): 443–453.

18.

Office of National Drug Control Policy (n.d.) Chapter 3. Integrate treatment for substance use disorders into mainstream health care and expand support for recovery. Available at: www.whitehouse.gov/ondcp/chapter-integrate-treatment-for-substance-use-disorders (accessed 6 June 2016).

19.

Ottenbacher

Cusick

(1990) Goal attainment scaling as a method of clinical service evaluation. American Journal of Occupational Therapy 44(6): 519–525.

20.

Ottenbacher

Cusick

(1993) Discriminative versus evaluative assessment: Some observations on goal attainment scaling. American Journal of Occupational Therapy 47(4): 349–354.

21.

Portney

Watkins

(2015) Foundations of Clinical Research: Applications to Practice, 3rd ed. Philadelphia, PA: FA Davis.

22.

Priebe

Huxley

Knight

(1999) Application and results of the Manchester Short Assessment of Quality of Life (MANSA). International Journal of Social Psychiatry 45(1): 7–12.

23.

Raphael-Greenfield

Gutman

(2015) Understanding the lived experience of formerly homeless adults as they transition to supportive housing. Occupational Therapy in Mental Health 31(1): 35–49.

24.

Shefler

Canetti

Wiseman

(2001) Psychometric properties of goal-attainment scaling in the assessment of Mann’s time-limited psychotherapy. Journal of Clinical Psychology 57(7): 971–979.

25.

Substance Abuse and Mental Health Services Administration (2016) Homelessness and housing. Available at: www.samhsa.gov/homelessness-housing (accessed 6 June 2016).

26.

Treatment Advocacy Center (2014) Eliminating barriers to the treatment of mental illness. How many individuals with a serious mental illness are homeless? Available at: www.treatmentadvocacycenter.org/problem/consequences-of-non-treatment/2058 (accessed 6 June 2016).

27.

Tsemberis

Kent

Respress

(2012) Housing stability and recovery among chronically homeless persons with co-occurring disorders in Washington, DC. American Journal of Public Health 102(1): 13–16.

28.

US Department of Housing and Urban Development (2015) 2015 Annual Homeless Assessment Report (AHAR) to Congress (November 2015). Available at: www.hudexchange.info/resources/documents/2015-AHAR-Part-1.pdf (accessed 6 June 2016).