College programming for students in addiction recovery: A PRISMA-guided scoping review.

BACKGROUND
The health and well-being of students in recovery from substance use disorder are increasingly being recognized as a priority on college campuses. This scoping review maps the state of the existing literature evaluating collegiate recovery programming to highlight research gaps and inform policy.


METHOD
We conducted a systematic search of articles related to collegiate recovery programming published before August 2020. The 15 extracted study characteristics included publication type, study design, primary outcomes, reporting of behavioral addictions, mutual-help group attendance, sample demographic information, school size, ownership, and funding source.


RESULTS
The PRISMA-guided search strategy identified 357 articles for abstract review; of 113 articles retained for full-text review, 54 studies met criteria for inclusion. Primary outcomes were coded into four domains: clinical, recovery experience, program characterization, and stigma. Most (57%) used quantitative observational designs and 41% employed qualitative research designs. Government or foundation grants funded 11% of the studies.


CONCLUSION
The domains identified offer a framework for healthcare providers, college administrators, and researchers to understand and improve programs, thereby better serving this vulnerable student group.


Prior reviews on CRPs
Though there are previous descriptive reviews on collegiate recovery programming (e.g., Bugbee et al., 2016;Depue & Hagedorn, 2015;, this is the rst study to provide a comprehensive scoping review. As noted in Arksey and O'Malley (2005), scoping reviews are distinguishable due to their comprehensiveness and attendant ability to identify gaps in the literature and to inform policy. Scoping reviews provide a "lay of the land" of the research in an area of study without assessing research quality. Conversely, systematic reviews provide a speci c answer to a research question from a narrow range of quality-assessed studies. Early CRP reviews have focused primarily on post-SUD treatment considerations (Morgan & Cavendish, 1988) and the recovery school movement (White & Finch, 2006). More recent literature reviews have focused on how CRPs t into the recoveryoriented system of care model (Bugbee et al., 2016;Harris, Baker, Kimball, & Shumway, 2008; and how CRPs help students succeed (Brown, Ashford, Heller, Whitney, & Kimball, 2018;Reed et al., 2020), However, none of these reviews was systematic. The two systematic reviews completed did not report any quantitative ndings related to CRPs (Ashford, Brown, Eisenhart, Thompson-Heller, & Curtis, 2018;Hennessey & Finch, 2018). One was a combined report on recovery schools (high school and college) and though some controlled trials had been conducted and reported on for recovery high schools, no outcomes from controlled trials were found for CRPs (Hennessey & Finch, 2018). A second review paper reported major CRP themes identi ed from a metasynthesis of qualitative ndings (Ashford, Brown, Eisenhart, et al., 2018).

Aims
Following Arksey and O'Malley (2005), the purposes of this review were to identify the initial research question, search for relevant studies, select studies, chart the data, and collate, summarize and report the studies utilized in the review. Lastly, the review aimed to identify gaps in the research to be addressed in future investigations. The ndings will be important for guiding policy and practice as they relate to college students in recovery at institutions of higher education. The reviewed literature will serve as a resource for CRP directors, college healthcare providers, college administrators, researchers, and other professionals who work with students in SUD recovery.

Data Sources and Searches
In adherence to guidelines for scoping reviews (Peters et al., 2015) and Preferred Reporting of Items for Systematic Reviews and Meta-Analyses (PRISMA) procedures (Moher, Liberati, Tetzlaff, & Altman, 2009), we developed a searching scheme to identify articles related to collegiate recovery programming. Three online databases were searched -PsycInfo, PubMed, and Web of Science-using search terms including substance use, college, student, and recovery (See Appendix A for full list of search strings) on July 8, 2020. Restrictions were not imposed based on dates, languages, or article types. To identify any further articles not located through these databases, we manually searched through relevant references, contacted experts in the eld, and set up alerts on Google Scholar to notify us when any relevant new research was made available, concluding the search on August 14, 2020. Figure 1 outlines the process of article selection utilizing the PRISMA guidelines (Moher et al., 2009). We excluded studies that did not focus on college students in recovery and studies that did not generate any de nitive statistical results regarding collegiate recovery programs or college students in recovery from substance use disorder. Also excluded were several article types: systematic reviews, literature reviews, and book reviews. Case reports and case studies were generally excluded with the exception of one study with rigorous experimental design and comprehensive results. One study written in Spanish that examined post-treatment college students in Mexico was excluded because it did not focus on recovery speci cally (not because it was written in Spanish). Finally, we examined all studies for the potential of duplicate results and eliminated any articles reporting ndings on the same data. Publicly available theses and dissertations were included but other unpublished works were not considered.

Study Selection
After any duplicate articles were removed, two researchers (NV, MR) independently reviewed all abstracts against the eligibility criteria. To be included, the abstract had to be (a) recovery-oriented (i.e., programming or services related to substance use disorder recovery), and (b) focused on college students. Any dispute was resolved through discussion. Full-text articles were then obtained for all included abstracts and subsequently independently reviewed by two researchers (NV, MR). Only articles that focused on college students in recovery, or recovery programming on college campuses, and reported thorough quantitative or qualitative results were included. Again, any disagreements were resolved through discussion between the researchers, and if agreement could not be attained, the senior author (KH) was brought in for resolution.

Data Synthesis and Analysis
The scoping review summarized and categorized articles across these 15 characteristics. It provides an overview of the literature without evaluation of article bias or research rigor.

Results
The initial search revealed 357 articles for abstract review, of which 244 were excluded after examination ( Figure 1). After separately reviewing the full-texts of the remaining 113 articles, NV and MR collectively identi ed 54 studies for nal inclusion and excluded 59. A list of the articles excluded and the reason for exclusion are provided in Appendix B. Table 1 offers a detailed summary of study characteristics.
During the iterative review process, we found primary outcomes fell into four major domains: clinical, lived experience (students in recovery answered open-ended questions regarding their experiences), program characterization, and stigma. All of the included articles are listed in Table 2 and summarized by CRP outcome domain/ category and study design. The majority of studies examined clinical outcomes and lived experiences of college students in recovery. Most were journal articles (32/54, 59%) or dissertations/theses (17/54, 31%), with four book chapters and one case study. There were 12 studies published prior to 2010, 11 studies published from 2010 to 2015, and 31 studies published after 2015.
Only one study was conducted outside the United States.

Study designs
The majority of studies used an observational design (31/54, 57%), including 17 cross-sectional, 9 prospective cohort, and 5 "other" observational designs. The remaining studies used a qualitative design (22/54, 41%), including 15 semi-structured interview, 3 focus group, and 3 "other" qualitative designs. The scoping review identi ed just one randomized controlled trial. We highlight some of the study designs in the explanation of primary outcomes below.

Funding source
A large number of studies either did not report funding (32/54, 59%) or were unfunded (7/54, 13%). When reported, funding was most often obtained from internal university funding sources (5/54, 9%) and philanthropic sources (4/54, 7%). Few studies reported government funding sources (6/54, 11% [1] ), with two from the National Institute on Drug Abuse (NIDA -though these were from the same grant), one from the National Institute on Alcohol Abuse and Alcoholism (NIAAA), one from the Substance Abuse and Mental Health Services Administration (SAMHSA), and two studies that reported "other government" funding sources.

Lived experience
Lived experience refers to studies in which students answered open-ended questions about their recovery experiences. Studies with the outcome of lived experience that were qualitative designs and coded information collected in semi-structured interviews to identify themes reported among CRP students were very common (16/54, 30%) in the review (Bell et al., 2009;Iarussi, 2018;Kollath-Cattano et al., 2018;Terrion, 2013;Walker, 2017;Whitney, 2018;Woodford, 2001;Workman, 2020). Other studies of lived experience investigated CRP alumni (Lovett, 2015), recovery discourses (Whitney, 2018), what made student recovery possible (Washburn, 2016), why students joined a CRP Laudet, Harris, Kimball, Winters, & Moberg, 2016), and the role of recovery identities among CRP members (Hoffman, 2020). A qualitative study using focus groups identi ed needs speci c to students in recovery on campus (Wor er, 2016).

Program characterization
About a quarter of the studies (13/54, 24%) examined the general characteristics of CRPs. Although some aimed to demonstrate the importance of the CRP as an essential resource on campus (Carlson, 2018;Watts, Chowdhury, & Holloway, 2019), others aimed to provide basic and program-speci c information Beeson, Whitney, & Peterson, 2017;Gueci, 2018a). Study designs were a mix of qualitative and observational designs.
[1] Note that percentages may not add to 100% due to rounding error.

Discussion
This scoping review summarizes the research literature regarding college student substance use disorder recovery related programming. Nearly all of the studies to date have used either observational or qualitative research designs. Like many other emergent literatures, randomized clinical trials generating e cacy estimates are rare in the CRP literature. Sample sizes were generally small. Nearly one-quarter of the studies did not report IRB approval, though some of these may have had such approval. Extramural funding was scarce among CRP investigations with only 11% reporting investment from a National Institute of Health funding organization, foundations, or other government funders. The lack of international studies on programing relating to university students in recovery suggests that CRPs may not exist in other countries outside the US, although it could also mean that CRPs have not attracted interest from researchers in other countries.

Evidence regarding CRPs
Though there may be evidence regarding the core elements of recovery programming in research on adult populations, how those elements operate speci cally among college students has not been effectively evaluated. Hence, it may be presumed that a blending of evidence-based interventions including recovery housing (Jason & Ferrari, 2010), peer recovery supports (Laudet & Humphreys, 2013), continuing care treatment programming (McKay, 2009), and mutual-help group (Kelly, Humphreys, & Ferri, 2020) facilitation are driving the effectiveness and growth of CRPs. Until additional controlled trials are completed, we cannot draw conclusions that are certain. There may be great di culty in conducting randomized controlled trials (RCTs) in this population, as the multi-component nature of CRPs does not lend itself to a comprehensive RCT design. Likewise, an important facet of CRPs is that they attract current students interested in recovery as well as individuals in sustained recovery back to college to begin or nish a degree who may not otherwise attend due to risk of relapse in an abstinence-hostile environment. Hence, this potentially confounding developmental difference in students should be accounted for in studies. Additionally, self-selection bias (e.g., those more motivated to change also may be motivated to engage in CRP) may cast doubt on CRP studies shown to improve student outcomes.
This limitation has been overcome in adult studies on Alcoholics Anonymous (Humphreys, Blodgett, & Wagner, 2014), and perhaps could also be handled in CRP research; for example, through studies that randomize arriving students to receive or not receive a tailored welcoming intervention to a CRP. Future studies could also exploit the exogeneity in availability (e.g., schools with and without CRPs) or using propensity score matching to evaluate CRP effectiveness while addressing the risk of selection bias.
Though CRP research is still in its infancy, some observational evidence suggests that programming for college students in recovery may be successful in helping students and CRP alumni to sustain abstinence (Bennett, McCrady, Keller, & Paulus, 1996;Botzet et al., 2008;Brown et al., 2019;Cleveland et al., 2007;Laudet et al., 2015). Likewise, increased GPA, retention in school, and graduation rates among students in recovery, compared to same school general population students, offer some moderate evidence of CRP e cacy in educational attainment Botzet et al., 2008;Moore, 1999;Watts, Tu, et al., 2019). Lastly, sobriety related social support was associated with reduced drug and alcohol-related cravings among CRP students (Cleveland & Harris, 2010a, 2010bWiebe et al., 2018;Zheng et al., 2013). Though more research is needed to explore these observational ndings, results may have implications for college administrators and researchers. For example, integrating interventions aimed at increasing social support for non-drinking lifestyles (e.g., sober tailgates and substance-free social outings) among students both in recovery and not in recovery has the potential to create a safer campus environment for all students.
Mutual-help group participation was recorded in only 37% of the studies. Because mutual-help has been shown to be effective among the general population (Kelly et al., 2020) and emerging adults (Kelly, Stout, & Slaymaker, 2013), the limited research on this data point suggests that mutual-help participation should be prioritized and collected with basic demographic information in future CRP studies. Similarly, behavioral addictions (i.e., gaming/internet, sex, food, gambling) were recorded in only 11/54 studies despite the high co-morbidity of behavioral addictions and SUDs (Cohen et al., 2010;Laudet et al., 2015), and the recent focus of examining these conditions in combination among CRP students Monsour, Kimball, & Hensley, 2020). Future research should aim to record information on behavioral addictions among all CRP students.

Qualitative research
Qualitative research is abundant in the CRP literature. Some of the common themes reported among these studies include the importance of on-campus mutual help meetings (Whitney, 2018), the role of the student-drop-in center (Ashford, Brown, Eisenhart, et al., 2018), and the role of community and social support among CRP students . Likewise, other qualitative investigations looked at themes related to the importance of CRP seminars and addiction education programming (Bell et al., 2009;Casiraghi & Mulsow, 2010), managing emotions (Lovett, 2015), academic success (Terrion, 2013), enhancing overall wellness (Iarussi, 2018), and recovery housing and diversity in CRP programs (Woodford, 2001). These ndings should act as a springboard for new quantitative research projects examining these common topics in detail.

Stigma
Research on stigma surrounding substance use disorder as it relates to college students was the focus of some articles in the CRP literature. The ndings indicated that CRP-related "recovery ally" trainings reduced stigma and improved self-reported ally related behaviors (i.e., use inclusive language, make others aware of CRP) for individuals completing the training (Beeson et al., 2019). These trainings have also been shown to increase empathy for students in recovery among student allies (Gueci, 2018a). A nal stigma related study used a unique methodology incorporating photographs to capture common themes to identify sources and consequences of stigma among students in a CRP (Spencer, 2017).
Sources of stigma included discrimination and expectation of rejection while consequences of stigma included loneliness, isolation, and fear of missing out. Future studies should not only examine the personal experiences of stigma, but also the overall level of stigma toward individuals with substance use disorder among all students at schools with and without CRPs. These ndings on school level stigma may provide policy recommendations for school administrators and lawmakers contemplating CRP implementation at the university and state level.
Knowledge gaps and informing policy, practice, and research As expected, this scoping review identi ed numerous gaps in the literature regarding programming and services related to college students in recovery from substance use disorder. Major gaps included a lack of 1) controlled trials, 2) mobilizers and mechanisms of behavior change (i.e., what components of CRPs lead to better outcomes and how speci cally they achieve those better outcomes ), and 3) implementation science research designs. Other important gaps in the literature included a lack of research on sociodemographic differences among students in CRPs, racial disparities among students in CRPs, and community college programming. Simply understanding how many CRP programs are currently operating and basic information on programming would be a huge step forward for research in this area. Though we did not implement a data point for studies examining CRP participants' with cooccurring conditions and criminal justice involvement, it should be noted that gaps were apparent on studies in these areas. Additionally, conceptual models were seldom applied to inform research design, and data collection. Taken together, these gaps provide researchers with important next-steps in this area of inquiry.

Controlled trials and implementation science studies
As noted, there has only been one randomized controlled trial including CRP students to date. However, it is important to note that this study did not randomize on CRP participation; rather all study participants were CRP students and were randomized into a bio-feedback control and treatment condition to evaluate levels of craving (Eddie et al., 2018). This highlights the most glaring de cit in the CRP literature: lack of funded (R01, R03, or R21 NIH funding mechanisms) controlled or matching trials severely limits the evaluation of CRP effectiveness and necessitates the urgency of research in this area. The only NIH funded study aimed speci cally at college students in recovery is now more than 7 years old and was not a controlled trial (Laudet et al., 2015(Laudet et al., , 2016. Without controlled trials and funded research, evidence-based practices and standards for CRPs cannot be ascertained. In order to suggest elements of CRPs contain interventions and practices that are evidence-based, the NIH and other funding sources must prioritize the understanding of the recovery phase of substance use disorder among college students. Implementation science is the study of methods to promote the systematic uptake of research ndings and other evidence-based practices and also to use these ndings to enhance routine practice and improve the quality and effectiveness of health services (Eccles & Mittman, 2006). Colleges represent a unique context in which to conduct implementation research due to their differing models of leadership, school year calendar, and student culture, and the diversity of students and staff (Owens et al., 2014). Likewise, colleges themselves may be in uenced by the local community, powerful alumni, the U.S. Department of Education, and funding at state and federal levels (Clotfelter, 2003;Jongbloed & Vossensteyn, 2016). Unfortunately, these contexts through which universities operate often do not work in harmony to support implementation of SUD recovery programming . The rst step of determining the implementation strategy for CRPs is identifying a framework to guide data collection, analyses, and interpretation such that contextual factors can be identi ed and studied. Hence, implementation science investigations are needed to identify barriers and facilitators to CRP implementation using well-validated frameworks (i.e., Consolidated Framework for Implementation Research [CFIR] or Reach Effectiveness Adoption Implementation Maintenance [RE-AIM]).
Sociodemographic differences, racial disparities, and gender differences This scoping review exposed a severe lack of research on underserved student groups and highlights the need to know more about CRP students of color, women students, and low-income and rst-generation students. We found no articles examining differences between socio-economic groups, though one study reported that a third of students had experienced homelessness in their lifetime (Laudet et al., 2015). Although three-quarters of the studies reported CRP participants' racial/ethnic background, 25% of the studies did not. None of the studies examined racial disparities or racial differences on any outcomes among college students in recovery. Only one study focused speci cally on women (Walker, 2017) and one examined sex differences (Smith et al., 2018). Though most studies 63% reported gender, 37% of the studies did not, and only two studies reported transgender/non-binary students (Carlson, 2018;Watts, Tu, et al., 2019). Likewise, only three studies reported student sexual orientation Ashford et al., 2019;Watts, Tu, et al., 2019). To inform college healthcare providers and policy, research is needed on these underrepresented students in terms of their use and outcomes of CRP participation. These ndings would provide guidance on the unique programming needs and inform tailored interventions programs for these underserved student populations.

Co-occurring conditions and criminal justice involvement
Though the evidence is preliminary, it appears that co-occurring conditions (mental health disorders and SUDs) are the norm among CRP participants. In connection with the ndings presented on behavioral addictions, multiple studies have documented a relationship between SUD and eating disorders Laudet et al., 2015), SUD and other mental health conditions Laudet et al., 2015;Odefemi-Azzan, 2020), and poly-substance use disorders (more than one reported SUD - Cleveland et al., 2007;Laudet et al., 2015). How these conditions confound CRP involvement or in uence recovery-related outcomes will be vital for future research.
Interestingly, there were two reports showing a high level of previous criminal justice involvement (58%-66%) among CRP members (Cleveland et al., 2007;Laudet et al., 2015). These ndings have implications for both CRP and criminal justice system researchers. First, among CRP researchers, investigations into internal processes among students leading to engagement in CRPs and desistance from the criminal justice system are warranted. Second, among criminologists and economists, research evaluating the cost savings associated with desistance from the criminal justice system among these students deserves immediate attention in the literature. Relatedly, research examining the effectiveness of reallocation of resources from the criminal justice system into CRPs should be prioritized.

Conceptual models
Though not collected as a data point, a small number of studies implemented a conceptual model to guide ndings. One such study included an effort to create a conceptual model using a systems-based community approach (Harris et al., 2008) and another incorporated the continuum of care model to treat addiction (Laitman et al., 2014;Laitman & Lederman, 2008). Others have suggested an integrated behavioral health model to treat all co-occurring mental health disorders along with SUD in CRPs . One study suggested a recovery model speci cally for community colleges (DiRosa & Scoles, 2020). Lastly, one study aimed to understand CRPs using a socio-economic model which categorizes CRPs into outcomes at the individual, interpersonal, organizational, and community levels (Beeson et al., 2017). Unfortunately, none of these models conceptualizes CRP programming in a comprehensive manner. Because CRPs can draw from research in different disciplines, including criminal justice, public health, education, and each is interested in their own models and preferred outcomes (i.e., recidivism, relapse, retention), this research gap is profound. Future research should prioritize creating a socioecological model of CRP outcomes to ll this gap.

Limitations
This review needs to be understood in light of limitations inherent in the study design. We did not review the quality of the studies included or investigate the research methods for potential bias; thus, we are limited in the conclusions we can draw. However, this limitation is standard to scoping reviews. This review was also limited by the amount of research available and the focus on largely public university samples in North America.

Conclusions
This scoping review was intended to give a current "lay of the land" as it relates to research on collegiate recovery programming. The studies included in the review are intended to provide an expansive overview of the literature and add to the general understanding of students in recovery from SUD. The review identi ed gaps in the research on programming for students in SUD recovery and highlighted areas for future inquiry. The available evidence on CRPs is minimal when compared to the extensive literature base on prevention and reduction of substance use on college campuses. Given such a disparity and the growing need for recovery-oriented services on college campuses, evaluations of CRP effectiveness are needed. The domains identi ed in this review offer a potential framework for healthcare providers and researchers, and will help to inform policy and practice to improve outcomes for this marginalized student group. Noel Vest designed the study, conducted the literature search, coded study characteristics, conducted the statistical analyses, and contributed to a majority of the manuscript writing. Meg Reinstra aided in formulating the search terms, conducted the literature search, coded study characteristics, created the PRISMA chart, and wrote the rst draft of the methods. Christine Timko helped formulate the search strategy and edited the manuscript. John Kelly provided edits to the nal draft of the manuscript. Keith Humphreys was the senior author and provided guidance, feedback, and manuscript edits throughout the project. All authors contributed to and have approved the nal manuscript.
Con ict of Interest: No con icts of interest reported.