Abstract
Background
Evidence-based practices (EBPs) improve housing and health for persons who have experienced homelessness with serious mental illness (PEH-SMI) but are challenging to implement. We tested a strategy to support pilot implementation of a 12-session housing skills training intervention for PEH-SMI, tailored from effective social skills training interventions. We aimed to optimize the implementation strategy and intervention prior to an implementation trial.
Method
We provided training and technical assistance to nine providers to support pilot implementation of this intervention to six groups of PEH-SMI (n = 35) engaged in VA Greater Los Angeles’ homeless services. We used scales and semi-structured interviews with 14 PEH-SMI and all interventionists to inform implementation strategy adaptations, identify factors that impacted implementation, and assess perceptions of the intervention. Attendance was tracked and we observed a random sample of each interventionist's groups to assess treatment fidelity.
Results
Interventionists perceived the implementation strategy and the intervention favorably. However, interventionists often lacked physical space, staff, and resources (e.g., computers) to conduct the intervention. Interventionists found the content valuable for participants and a few suggested that group engagement should be a prerequisite for obtaining housing services. PEH-SMI were interested in the intervention's content and receptive to the group-based format. Participants attended a mean of 4 ± 3/12 groups; all groups observed had acceptable fidelity. Problems with intervention retention were described, suggesting challenges maintaining group participation when participants transitioned between VA homeless services.
Conclusions
To support the implementation of an EBP for PEH-SMI in homeless programs, these data suggest the value of training/technical assistance and strategies that enhance program-level buy-in to address resource concerns. Intervention adaptations, e.g., using a drop-in, open group format, in community-based settings that are easily accessible to PEH-SMI, may also increase adoption. This project was registered as “Improving Housing Outcomes for Homeless Veterans” Trial registration NCT03646149, registered 8/24/2018.
Plain Language Summary
There are effective social skills programs for people with serious psychiatric disorders; we adapted these programs into a 12-session housing skills program for people who had experienced homelessness. We then tested a training and technical assistance package to support the program's delivery by nine providers (e.g., social workers) to six groups of homeless people with serious psychiatric disorders (n = 35). We used surveys and interviews with some participants (n = 14) and all involved providers (n = 9) to understand their perspectives on our training and technical assistance, as well as the program itself; and to assess their views on factors that affected the program's use in real-world settings. We tracked participants’ attendance at the groups and observed a random selection of groups to see if providers adhered to key program elements. We found that participants attended an average of one-third of the program's groups (4/12) but that providers were able to deliver the program to include all key elements. Some providers lacked important resources (e.g., classroom space or computers) to deliver the program as it was intended; they liked the training and technical assistance offered. Participants liked the program's content and format. Difficulties with participant retention may relate to drop-out from homeless services in which the program was delivered. This pilot project suggests that getting buy-in from leaders across levels and structuring the program as a drop-in group, in the community, or in places where attendance is easy for participants may increase its likelihood of being used as part of routine homeless services.
Background
Evidence-based practices (EBPs) improve housing for persons who have experienced homelessness with serious mental illness (PEH-SMI), including permanent supportive housing (PSH: subsidized community-based housing with long-term supports) (Hwang & Burns, 2014) and rapid rehousing (time-limited financial subsidies and supports) (Cunningham et al., 2015). Though implementation approaches focus on strategies to implement housing interventions with fidelity (Kertesz et al., 2017), few have studied strategies supporting the implementation of psychosocial EBPs for PEH-SMI within homeless service settings.
PEH-SMI report higher satisfaction with housing-focused EBPs when they are taught housing skills (Calsyn et al., 2002), which largely involve interpersonal communication (Kopelowicz et al., 2006). Evidence-based social skills training (SST) interventions (Bellack et al., 2004) strengthen interpersonal skills through “behavioral instruction” (role modeling and feedback) and improve functioning. Despite the Department of Veterans Affairs’ (VA's) national rollout of SST in psychiatric rehabilitation settings (U.S. Department of Veterans Affairs, n.d.), little is known about strategies that support SST implementation in housing programs for PEH-SMI. We piloted a training and technical assistance strategy used in VA's rollout to support the implementation of a 12-session SST for PEH-SMI.
Guided by the Consolidated Framework for Implementation Research (CFIR) (Damschroder et al., 2009), we use qualitative data to highlight contextual factors that impacted implementation and suggest modifications to the implementation strategy and intervention for future trials. We used surveys and qualitative data to describe the intervention's acceptability and appropriateness (Proctor et al., 2011); quantitative data to capture feasibility; and qualitative data to examine ways to enhance feasibility.
Method
Setting and Ethics
We conducted this pre-implementation pilot at VA Greater Los Angeles, which has robust services for PEH-SMI. Study procedures were approved by the VA Greater Los Angeles Institutional Review Board and informed consent was obtained from participants. This pilot was registered as “Improving Housing Outcomes for Homeless Veterans” (Trial registration NCT03646149, registered August 24, 2018).
Study Design and Goals
This pre-implementation study assessed an implementation strategy; and acceptability, appropriateness, and feasibility of an adapted version of SST piloted in homeless service settings. This practice has strong evidence in the SMI literature (Kurtz & Mueser, 2008) and we intended to inform optimization of its implementation strategy before a larger trial in homeless service settings. We used a convergent mixed methods design (Fetters et al., 2013), collecting and analyzing quantitative and qualitative data separately but comparing findings of each data type. We initially intended to also explore the intervention's effectiveness (e.g., on social skills) by measuring change from baseline to 6 months. However, physical distancing requirements of the coronavirus of 2019 (COVID-19) pandemic, complicated by technology barriers, did not allow measure administration at 6 months.
Participants
We identified nine staff (seven social workers, one mental health counselor, and one addiction therapist) who worked with PEH-SMI and were interested in piloting the intervention. These staff worked in four homeless service settings: residential rehabilitation for PEHs; PSH for PEHs run by a VA community partner; “bridge housing” that transitions PEHs to longer-term placements; and an SMI-focused intensive outpatient program.
Each interventionist approached PEH-SMI from their programs and identified 4–8 participants willing to receive SST. Eligibility criteria included: homeless experiences (past/current VA homeless service use); mental illness (psychotic disorders, bipolar disorders, major depressive disorders, anxiety disorders, and/or post-traumatic stress disorder [PTSD]) noted in VA's electronic health record (EHR); and serious functional impairments (per interventionists’ clinical impressions). As PEH-SMI were identified by staff as part of their clinical practice, we do not know the number of persons approached or deemed ineligible to participate. Each PEH-SMI (total n = 35) participated in one of six intervention cohorts in one of the four aforementioned settings. Participants were not compensated for intervention participation; they received $20 for each of two assessments.
Social Skills (“Housing Skills Training”) Intervention
Methods used to tailor this intervention (“Housing Skills Training”) from evidence-based SSTs are described elsewhere (Gabrielian et al., 2019); in brief, we used literature review, key informant interviews, a national expert consensus panel, and two focus groups with PEH-SMI to reach consensus on social skills relevant to housing attainment and retention for PEH-SMI. These skills (Table 1) were selected and adapted from a list of skills used in psychiatric service settings for persons with SMI. A meta-analysis shows that social skills interventions have a moderate mean effect size on community functioning and a large weighted mean effect size on social skills (Kurtz & Mueser, 2008) among persons with SMI. The intervention is feasible in VA psychosocial treatment settings (U.S. Department of Veterans Affairs, n.d.) but absent within VA's homeless programs.
Housing Skills Training Session Topics and Implementation Strategy Components
Housing skills training intends to improve housing attainment and retention. The first session was delivered as an individual session or group (4–8 participants/interventionist, consistent with SST for persons with SMI); it introduced participants to the intervention and the interventionists, set expectations, and identified participants’ housing/mental health goals. The remaining 11 sessions were delivered in groups (1–2 times/week, 1 h/group), led by 1–2 interventionists. The 12-session intervention was delivered over 6–12 weeks, dependent on sessions per week. Each session focused on 1 of 19 skills classified into six themes (Table 1); skills were selected by the interventionist(s) to reflect participants’ goals. Sessions were interactive, engaging participants in behavioral instruction, followed by “outside practice” (applying skills in everyday life).
Though the intervention was designed to occur face-to-face, the COVID-19 pandemic began mid-study; five of six cohorts conducted the group face-to-face with precautions and the sixth conducted the group via telephone/videoconference. Cohorts were grouped by interventionist's existing caseloads.
Implementation Strategy
Implementation was supported by training, case consultation, and educational materials (Table 1) that enabled providers’ access to relevant knowledge and information and paralleled supports used in VA's SST roll-out. Training included an asynchronous 3-hour training web course; and a 2-hour in-person or synchronous videoconference training. Trainings reviewed the intervention's approach, detailed the content of individual sessions, and engaged interventionists in experiential learning (simulated groups with feedback). All interventionists received Housing Skills Training educational materials (manual with patient worksheets, group leader discussion points, and implementation tips) and templated progress notes were created in the VA's EHR that captured intervention core components and each session's skill focus. During their first time delivering the intervention, interventionists received case consultation (30 min/week, via videoconference or telephone).
Data Collection
At baseline, we reviewed the EHR and conducted baseline assessments with participants (n = 35). Post-intervention, we administered surveys to 16 participants and all 9 interventionists; and conducted, recorded, and transcribed 30-min semi-structured telephone interviews with a subset of PEH-SMI (n = 14) and all nine interventionists.
Participant Characteristics
We abstracted age, gender, race/ethnicity, and percent service-connection (VA disability rating) from the EHR. We used the EHR “problem list” (diagnosis list) and notes from mental health visits/admissions in the year prior to intervention initiation to abstract the presence or absence of depression, bipolar disorder, PTSD, anxiety disorders, schizophrenia/other psychotic disorders, and alcohol/drug use disorders. The EHR was used to track group attendance (presence/absence of progress notes). As the EHR does not consistently record housing status, we administered the Residential Time-Line Follow-Back (TLFB) Inventory (Tsemberis et al., 2007) to each participant at baseline; the TLFB is a retrospective event history housing status over the past 6 months. We classified residences as “stable,” “sheltered homelessness” (e.g., transitional housing), or “unsheltered homelessness” (Gabrielian et al., 2015).
Implementation Strategy
Interventionist interviews asked about perceptions of the implementation strategy. We sought suggestions to improve the training and technical assistance package to increase practice adoption.
Acceptability and Appropriateness
We captured interventionist and participant satisfaction and perceived utility of the intervention. Interventionists completed an eight-item version of the Perceived Characteristics of the Intervention Scale (PCIS), a reliable measure of perceived characteristics with preliminary support for its validity (Lau et al., 2018); the PCIS aligns with the CFIR construct of intervention characteristics to capture perceptions about an intervention's acceptability and its potential for future implementation and sustainability in routine care. Participants completed the Client Satisfaction Questionnaire (CSQ)-8 (Attkisson & Greenfield, 2004), a measure of satisfaction with interventions with strong internal consistency, item-total correlations, and inter-item correlations.
PEH-SMI interviews included questions about their intervention experiences, including their satisfaction with it, challenges they experienced or how it helped, and suggested changes. Interventionist interviews asked about parts of the interventions that they found more/less useful and ways to enhance the intervention's fit for this population and setting.
Feasibility
We conceptualized feasibility to encompass intervention retention and treatment fidelity (Teresi et al., 2022). Retention was calculated via the number/percent of groups attended; interviews with interventionists assessed perceptions surrounding retention. We observed a random sample of three groups/interventionist to assess fidelity, captured via an adapted version of a fidelity checklist used in VA SST (Likert scales to capture adherence to intervention core components (Supplemental Material S1; U.S. Department of Veterans Affairs, n.d.).
Analyses
To enhance our analytic rigor, we maintained process notes with methodological decisions and analytic products (Wolf, 2003). Interview transcripts were entered into ATLAS.ti Version 9; we used the CFIR Codebook (CFIR Research Team-Center for Clinical Management Research, n.d.), augmented with codes from interview guide, as a top-level codebook for deductive analyses. Two authors (a masters-level psychologist and physician trained in rapid qualitative analyses) applied this codebook to three transcripts, comparing findings to ensure consistent use of the top-level codes, resolving minor discrepancies, and augmenting the codebook with findings from the data. A finalized codebook was developed iteratively with the study team and applied to all transcripts.
We calculated means and standard deviations of the PCIS and CSQ-8 using Microsoft Excel. We calculated the mean, standard deviation, and range of the number of groups participants attended. Derived from VA SST, we used a threshold of >80% of checklist items assessed as “fully completed” to identify if fidelity was adequate/inadequate.
Results
Table 2 characterizes participants’ demographics (typical for homeless-experienced VA Greater Los Angeles users; Gabrielian et al., 2014), as well as diagnoses and number of groups attended. The sample's mean age was 54.3 ± 12.2 years; most (88.6%) were male. Over two-thirds (68.6%) were African American and nearly half (42.9%) had profound functional impairments (≥70% service-connected). Most (77.1%) were experiencing sheltered homelessness. Many participants had drug (79.4%) or alcohol use disorders (55.9%). Results by data collection domain are detailed below and summarized in Supplemental Material S2.
Participant's Demographic Characteristics, Diagnoses, and Number of Housing Skills Training Groups Attended
As assessed by the Residential Time-Line Follow-Back Inventory.
Implementation Strategy
Interventionists highlighted the benefits of the implementation strategy; the training and technical assistance provided access to relevant skills and interventionists felt they gained knowledge required to implement the intervention. As one interventionist stated: “I felt confident … I had all the information.” To increase practice adoption, interventionists suggested augmenting the strategy to reflect local implementation challenges. Often, they lacked needed resources, worsened by the pandemic's physical distancing requirements; as one interventionist asserted, “we don’t have space … or a computer [to use materials provided].” They suffered staff shortages and did not have transportation resources. As one interventionist said, “there aren’t facilitators … we’re short social workers.” Another described, “transportation [is needed] … to be accessible to the population.”
Acceptability and Appropriateness
Survey findings were congruent with interviews regarding the intervention's acceptability and appropriateness. On the PCIS (range 1–5, higher scores indicating better perceptions) (Lau et al., 2018), the mean among interventionists was 4.1 ± 0.4. On the CSQ-8 (range 8–32, higher scores indicating greater satisfaction) (Attkisson & Greenfield, 2004), participants had a mean 26.2 ± 4.1 (high satisfaction) (Attkisson, 2020).
Interventionists across settings perceived the intervention's content favorably; one described, “I think the topics were right on and something that our clients need.” Another said, “I liked the topics that were covered. I liked that it was structured and that there were topics that we could choose from.” Two interventionists described the intervention's content as an important prerequisite for housing. One interventionist stated, “There definitely is a need because people cycle through … people have been through the programs up to 8 times, so I think it is necessary.” Participants were also satisfied with the intervention's content, particularly valuing money management training. As one participant described: “Everyone needs a housing training group … if [they have] been on the street … [they may have] lost the skills of budgeting.” Similarly, another participant described “budgeting, or making a plan of [my] bills, was really helpful to stay focused on housing.” A participant described: “it might cut down a lot on [housing loss] if [learning information taught in this group] was mandatory.”
The format of the intervention was also well received; participants and interventionists liked the structured behavioral instruction and group-based format. An interventionist described that they “liked the camaraderie … it provided a space for [participants] to support each other.” As one participant stated, “we interacted [in the groups] … and we got to role play, which was good.”
Feasibility
Regarding feasibility, participants attended an average of one-third (4.0 ± 2.8/12, range 1–10) of groups; stably housed participants attended more groups than homeless participants (5.1 ± 3.6 vs. 3.8 ± 2.5). Interventionists’ interviews reflected Housing Skills Training's modest retention; as one asserted, “our biggest problem was how to get [participants] to keep coming back.” For groups in residential rehabilitation or bridge housing, program turnover affected retention; as an interventionist described: “it was difficult to retain [participants] the 12 weeks … if they relapsed they got discharged [from rehab].” One interventionist suggested redesigning the group as an individual intervention to enhance retention: “a curriculum that is one-on-one and tailored.” All groups observed had adequate fidelity (>80% of items on our fidelity checklist fully completed).
Discussion
We pilot tested a training and technical assistance strategy to support SST (“Housing Skills Training”) for PEH-SMI. The implementation strategy was valued, but our data suggest adaptations to the strategy and intervention prior to a larger implementation trial. Interventionists and participants found the intervention acceptable and appropriate; they endorsed its content and format. Though recruitment and treatment fidelity were feasible, strategy adaptations may enhance practice engagement.
To implement the intervention with adequate PEH-SMI engagement, enhanced buy-in from program-level leadership may address competing staff responsibilities and resource limitations. Multi-level stakeholder engagement is critical to ensure organizational EBP adaptations and adoption (Holcomb et al., 2021); flexible supports tailored to context (e.g., facilitation) (Ritchie et al., 2014) may enhance adoption. Though we attempted a closed group, PEH-SMI dropped out because they left transitional housing or had competing needs (e.g., food) (Davies & Wood, 2018). A drop-in, open group may increase PEH-SMI engagement and align with the population's needs (Miller & Mason, 2012).
Though telehealth escalated markedly with COVID-19, role plays/behavioral instruction are challenging on telehealth. Implementing this intervention in community-based settings accessible to PEH-SMI, and/or ensuring space and transportation if the intervention is delivered in clinics, may increase engagement.
This work has limitations. This pilot was based in an urban VA; findings may not apply to other settings. However, this pre-implementation work, using mixed methods to obtain information that can be taken to interventionists and leadership to refine an implementation strategy and EBP, may aid implementation trials of other practices for vulnerable populations. Additionally, the timing of this pilot, interrupted by COVID-19, had implications regarding staff and resource limitations. Though we were unable to gather effectiveness data, SST has a strong evidence base (Kurtz & Mueser, 2008).
Conclusions
This pre-implementation pilot collected data that can inform future adaptations to a housing skills training intervention and strategies that optimize its implementation. Intervention adaptations, e.g., a drop-in format, in community-based settings, may be useful, as well as augmenting training and technical assistance with flexible, tailored supports that respond to context. Our findings align with literature highlighting the value of strategies that enable multi-level stakeholder engagement and the importance of tailoring EBPs to align with populations’ vulnerabilities (Pro et al., 2021). More work is needed to partner with PEH-SMI to optimize strategies supporting housing skills training implementation, as well as to identify additional EBPs, and strategies that support their adoption, for consumers who receive care in homeless service settings.
Supplemental Material
sj-docx-1-irp-10.1177_26334895241236679 - Supplemental material for Testing an implementation package in a housing skills training pilot for homeless-experienced persons with serious mental illness
Supplemental material, sj-docx-1-irp-10.1177_26334895241236679 for Testing an implementation package in a housing skills training pilot for homeless-experienced persons with serious mental illness by Sonya Gabrielian, Alison B. Hamilton, Lillian Gelberg, Ella R. Koosis, Lauren Hoffmann, David M. Carlson and Alexander S. Young in Implementation Research and Practice
Supplemental Material
sj-docx-2-irp-10.1177_26334895241236679 - Supplemental material for Testing an implementation package in a housing skills training pilot for homeless-experienced persons with serious mental illness
Supplemental material, sj-docx-2-irp-10.1177_26334895241236679 for Testing an implementation package in a housing skills training pilot for homeless-experienced persons with serious mental illness by Sonya Gabrielian, Alison B. Hamilton, Lillian Gelberg, Ella R. Koosis, Lauren Hoffmann, David M. Carlson and Alexander S. Young in Implementation Research and Practice
Footnotes
Acknowledgments
The views and opinions of the authors expressed herein do not necessarily state or reflect those of the VA or the U.S. Government. They would like to acknowledge the contributions of Drs. Brian Mittman and JoAnn Kirchner to this pilot project. They also thank leadership involved with the VA Social Skills Training (VA-SST) for Serious Mental Illness rollout.
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the U.S. Department of Veterans Affairs Health Services Research and Development (HSR&D) Career Development Award (CDA) 15-074 (principal investigator: Sonya Gabrielian), with additional support from her NIMH/VA Implementation Research Institute fellowship. Alison B. Hamilton is supported by a VA HSR&D Research Career Scientist Award (RCS 21-135).
Supplemental Material
Supplemental material for this article is available online.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
