Challenges and Rewards of Conducting Research on Recovery Residences for Alcohol and Drug Disorders

Abstract

Although research shows treatment for alcohol and drug problems can be effective, persons without stable housing that supports recovery are at risk for relapse. Recovery residences (RRs) for drug and alcohol problems are a growing response to the need for alcohol- and drug-free living environments that support sustained recovery. Research on RRs offers an opportunity to examine how integration of these individuals into a supportive, empowering environment has beneficial impacts on substance use, housing, and other outcomes, as well as benefits for the surrounding community. Research can also lead to the identification of operations and practices within houses that maximize favorable outcomes for residents. However, research on RRs also presents significant obstacles and challenges. Based on our experiences conducting recovery home research for decades, we present suggestions for addressing some of the unique challenges encountered in this type of research.

Introduction

Research over the past several decades has shown a consistent albeit moderate impact of treatment on substance use disorders (National Institute on Drug Abuse, 2012). One factor affecting the success of treatment is the availability of recovery capital, which includes the economic and social resources necessary to access help, initiate abstinence, and maintain a recovery lifestyle (Cloud & Granfield, 2008; Laudet & White, 2008). Individuals with substance use disorders who are unemployed, do not have stable housing, or are involved in the criminal justice system are particularly vulnerable given their limited access to recovery capital. RRs, such as Oxford Houses™ (OHs), sober living houses (SLHs), and other types of recovery homes for alcohol and drug problems can help increase recovery capital by providing affordable, alcohol- and drug-free living environments and peer support for recovery (Jason, Mericle, Polcin, & White, 2013).

Types of Recovery Residences

There are different types of recovery residence models that vary in terms of administration, services offered, type of residence, and staffing. The National Association of Recovery Residences (National Association of Recovery Residences, 2012) is an organization that provides advocacy and standards for RRs and has devised four levels of RRs based on these factors. All levels provide an abstinent living environment and social support for recovery within a communal living arrangement. Level I residences are democratically run by resident peers, offer no on-site services, are small facilities located in residential neighborhoods, and do not employ on-site staff. Resident fees usually cover financing of these homes and residents are free to live there as long as they wish. OH’s are good examples of these residences. Level II residences are similar to level I houses in most respects, but they typically have an on-site house manager who oversees house operations. The manager is typically paid or receives reduced rent but is considered a recovering peer, not a professional service provider. Good examples of these residences are SLHs, many of which are located in California. A key difference between level III houses and the first two levels is that they often offer on-site recovery support and other services and employ paid staff. Some of the recovery homes in Philadelphia studied by Mericle, Miles, Cacciola and Howell (2014) are could be considered Level III residences. Level IV houses tend to have an organizational hierarchy, offer on-site clinical services delivered by certified and licensed professionals, and are often larger facilities licensed as treatment programs.

This paper primarily addresses research conducted on the first three levels because they have been studied less than level IV residences. RRs present unique challenges, including recruitment of individuals across multiple sites, describing common and unique characteristics of individual homes, tracking participants for follow-up interviews, enlisting homes as partners in research, and implementing the most appropriate research designs.

Our collective experiences studying recovery homes draw primarily upon research conducted on three types of recovery homes: Oxford Houses (OHs), California Sober Living Houses (SLHs) and a mix of level II and III recovery homes in Philadelphia. OHs are a good example of level I homes. They began as a grassroots movement in the late 1970’s and have seen continued growth over the past four decades. Currently, there are over 1,700 houses nationwide. Research on OHs has been conducted by a team from DePaul University over the past several decades and they have documented favorable outcomes relative to control groups for persons who entered OHs after leaving treatment and criminal justice institutions (Jason, Olson, Ferrari, & Lo Sasso, 2006; Jason, Olson, & Harvey, 2015; Jason, Salina, & Ram, in press). An additional study documented good outcomes for persons who were current residents and followed up over three 4-month intervals (Jason, Davis, & Ferrari, 2007). The OH network is the only level I or level II RR model that has been endorsed by SAMSHA as being effective as an aftercare service for persons completing long-term therapeutic community treatment (see the Substance Abuse and Mental Health Services Administration National Registry of Evidence-Based Programs and Practices (Substance Abuse and Mental Health Services Administration, 2013). Although outcome studies of SLHs and other types of level II RRs have been conducted, there is a need for randomized designs that would make them eligible for the SAMSHA endorsement. There have been a number of randomized trials over the past decade supporting the effectiveness of level III and IV residences (e.g. (Cheng, Lin, Kasprow, & Rosenheck, 2007; Greenwood, Woods, Guydish, & Bein, 2001; Milby, Schumacher, Wallace, Freedman, & Vuchinich, 2005; Sacks, Chaple, Sacks, McKendrick, & Cleland, 2012).

California SLHs are the predominant recovery residence model in California and are good examples of level II houses. The earliest forms of SLHs emerged in the 1940’s in response to housing need among groups of persons attending Alcoholics Anonymous (Wittman & Polcin, 2014). There are currently about 800 homes in California associated with two different organizations, the Sober Living Network and the California Association of Addiction Recovery Resources. A team of researchers at the Alcohol Research Group has conducted studies on SLHs over the past decade. Using an “intent to treat” design assessing all individuals entering SLHs over an 18-month time period, researchers documented significant improvement in multiple areas of functioning (e.g., reduced substance use, reduced arrests, increased employment) that were maintained over the 18-month time period (Polcin, Korcha, Bond, & Galloway, 2010a; Polcin, Korcha, Bond, & Galloway, 2010b). These studies are the only investigations of level I or level II houses to assess longitudinal outcomes for the heterogeneous mix of all individuals who are referred to recovery houses from multiple referral sources. Previous longitudinal studies have focused on specific subgroups, such as persons entering residences after completing long-term residential treatment (Jason et al., 2006).

A final group of researchers examining RRs studied recovery homes in Philadelphia (Mericle, Miles, Cacciola, & Howell, 2014; Mericle, Miles, & Cacciola, 2015). These residences differed from OHs and SLHs in that some were publicly financed, had limits on how long residents could stay, and offered a variety of on-site services. Researchers documented that these houses generally fit into level II and III houses and used a peer oriented “social model” approach to recovery to varying degrees.

Despite their increasing numbers and potential influence, relatively few research teams beyond the aforementioned groups have studied RRs. One aim of this paper is to discuss the rewards of engaging in this work in terms of the beneficial impact that RRs can have on individuals struggling with substance abuse and housing instability. We also identify the important role that RRs can play in systems of care for persons with substance abuse disorders.

We suggest there is a balance in this work between maintaining an objective, scientific perspective and recognizing how the personal experiences and reactions we have as we as we interact with participants motivates us to do this work. A second aim is to identify the obstacles and challenges we have encountered during our research on RRs and ways of handling these issues that can help prepare future researchers for this work.

Recovery housing as a priority for addiction research

After decades of emphasizing acute care and brief intervention models for treating persons with alcohol and drug disorders, researchers, policymakers and providers are increasingly focusing on services that can help individuals sustain long-term recovery in the community (McLellan, 2002; Scott, Dennis, Laudet, Funk, & Simeone, 2011). A major problem with acute care interventions is that the improvements made during treatment are often short-lived, particularly if the individual does not have access to an alcohol and drug free-living environment that supports recovery. Many of the strategies to improve continuing care services after treatment have included ongoing case monitoring and phone based interventions (Dennis, Scott, & Funk, 2003; McLellan, McKay, Forman, Cacciola, & Kemp, 2005; McKay, 2005). However, we posit that a critically important component of successful long-term recovery is access to an alcohol- and drug-free living environment that includes social support for recovery. RRs are good examples of these types of services and they are rapidly increasing in numbers (National Association of Recovery Residences, 2012). However, only a few research teams have engaged in studying resident outcomes.

Although research has been conducted on self-help groups such as Alcoholics Anonymous (Ye & Kaskutas, 2009), only a few research teams have examined RRs. Many treatment and research professionals have limited knowledge about them and these types of services receive scarce if any attention in graduate training programs. When grant applications to study RRs are submitted to funding sources reviewers may have limited knowledge about them and a variety of incorrect assumptions. It is therefore critical for applicants to clearly describe the organization and operations of the RRs to be studied along with the potential benefits to residents. Because they are less familiar to reviewers, applicants have an opportunity to highlight innovation. There is a clear need for more dissemination of information about RRs as an adjunct or alternative to treatment. There is also a need for more dissemination of the growing evidence base for RRs.

When researchers decide to study RRs, they face a variety of potential obstacles. In addition to the difficulty acquiring funding, there are problems such as identifying the population of recovery homes from which to sample, recruiting residences, characterizing different types of residences, enlisting residents as partners in research, tracking research participants for follow-up interviews, and considerations for research measures and designs. Suggestions for addressing these issues are discussed below based on our experience studying RRs for over a decade. We also discuss some intrinsic personal rewards in this work that go beyond the satisfaction garnered by addressing gaps in service delivery and research.

Identifying and Sampling Recovery Residences

A unique challenge to studying RRs is that, compared substance abuse treatment, they are a less well-understood phenomena and information to characterize and identify them is still evolving. Treatment programs that are licensed by states are typically readily identifiable. However, recovery residences are usually not licensed and therefore harder to identify. That being said, a number of strides have been made with respect to defining RRs, delineating different types of them, and identifying where they are located--factors integral to highlighting the potential uniqueness of a particular study and to addressing the generalizability and impact of findings from the study.

For a number of years, Oxford House, Inc. has maintained a directory of houses and operated a website providing information about the OH model as well as a listing of houses that can be searched by state. These resources have proved to be invaluable to researchers in terms of locating and recruiting houses into research studies. The formation of the National Alliance for Recovery Residences (NARR) has represented an important step forward for non-OH residences. Standards developed by NARR (National Association of Recovery Residences, 2012) categorize different types of RRs (including OHs) based on their organizational structure, physical characteristics, staffing, and services provided. In addition to providing a framework for understanding RRs, NARR provides support to statewide and regional affiliate organizations in their efforts to certify that residences operating within their geographic purview do so in accordance with the NARR standards. The NARR website lists states with RR organizations affiliated with NARR. Unfortunately, there is currently no national directory of residences implementing the NARR standards—a substantial impediment to research on non-OH RRs. However, many state-level affiliates of NARR, such as the Georgia Association of Recovery Residences and California’s Sober Living Network, do maintain a listing of certified residences which can be used to identify potential research sites. However, a significant challenge noted by Mericle, Miles & Cacciola (2015) is that some RRs close after short periods of time and some of these residences later reopen or relocate, which makes them difficult to track. In addition, residences that are not affiliated with any recovery residence organization are difficult to study because there is no systematic way to know of their existence. Generalization of research from studies of homes associated with recovery house organizations that monitor quality to non-affiliated homes is questionable at best.

Engaging Recovery Residences as Partners in Research

Locating RRs and developing a sampling plan is only a first step. Enlisting the support of decision makers and key stakeholders is a critically important step in the research process (Henderson, Sword, Niccols, & Dobbins, 2014; Ross, Lavis, Rodriguez, Woodside, & Denis, 2003), and studying RRs and those who live in them requires the support and involvement of those in charge of running them. Because RRs vary in their organizational structure and staffing, this may be one person or it may be a variety of individuals, and identifying critical gatekeepers can often be a challenge. However, prior to embarking on a study, it is critical to identify who does what within a particular residence and who else may be involved in decisions that are made in the residence; “buy-in” is essential at every level.

The successful implementation of studies on OHs illustrates the importance of RRs having an organizational structure that supports research. OHs are part of a national organization that has local chapters within states. Chapters include volunteers and in some states paid staff who help new houses successfully launch their homes and implement standards required of all OHs. Chapters are also available to help homes that are struggling with difficult internal issues (e.g., noncompliance with house rules and regulations, dropout, etc.) or responding to pressures from the surrounding community. OH staff and volunteers at the national and chapter level have been instrumental in supporting research by educating consumers about its importance and encouraging resident participation in studies. In addition, each year there is an annual OH conference where staff, volunteers, residents, and researchers come together to celebrate recovery and share information, experiences, ideas, research results, and plans for the future. Interaction among these groups helps generate new research ideas, interpret the meaning and practical implications of research findings, and disseminate findings within the organization.

Regardless of organizational support, it is critical to engage those in the residence overseeing the day-to-day operations of it. These individuals may be most knowledgeable about what actually happens in the residence (as opposed to what may written in a manual or brochure), and they are the gatekeepers to the residents living there. Although directors or clinical staff in treatment programs may have had some prior exposure to research or training in research methods as part of their formal education, it is unlikely that the person in charge of the operations of the residence has had these experiences (Mericle et al., 2015). Given the constraints of resources available to RRs and their operators, it is important to understand the limitations of their ability to coordinate research protocols and to take this into account when designing studies. Although failure to adhere to study procedures may seem like resistance, it is more likely the case that the researcher needs to do more education about why the procedures are necessary and to provide more support to the person in charge of the residence to ensure that the research procedures can be carried out.

Individuals in the upper-tier of the organizational hierarchy must be clear about their role in the research. It is important for these individuals to understand that the purpose of scientific research is to answer important questions. They need to understand that research is often theory-based and hypothesis-driven, meaning that researchers have ideas about potential outcomes of the research, but that a hallmark of scientific research is objectivity, meaning that researchers should strive to be distanced from what they study so that findings depend on the nature of what was studied rather than on the beliefs and values of the researcher (Payne & Payne, 2004). Individuals who own and operate RRs understandably have strong beliefs about the value of what they are doing as well as a financial stake in being able to claim that scientific evidence supports it. Researchers can feel considerable pressure to report only findings that support RRs and ignore or minimize negative findings. To the degree that this occurs, data that could be used to improve RRs and delineate their role within larger recovery systems can get lost. The inherent difference in perspectives can potentially create obstacles in the research process, particularly if the owner/operator is unfamiliar with research and the need for it to be objective. However, those who own/operate RRs help people successfully recovery from addiction, often in the face of tremendous obstacle and barriers (Mericle, Miles, & Way, 2015). It important to underscore that research provides information that can be used to validate the areas of strength of RRs, but it can also be used to point out issues that need attention to improve operations and, most importantly, resident outcomes.

Addressing Specific Issues

Research on RRs entails all of the challenges of community-based research, starting with the selection of research methods that will result in collection of valid and generalizable data. Beyond that, there are specific considerations for coordination of activities across multiple research sites, locating study participants for research interviews, and utilizing personal experiences as a way to understand the impact of RRs on residents’ lives. The issues discussed below and the suggestions for addressing them are based on our collective experiences studying RRs for the past two decades.

Study Designs

RRs emerged organically as grassroots movements among persons in recovery and that history has implication for selection of optimal research designs. In a variety of publications Borkman and colleagues (Borkman, 1999; Borkman, Kaskutas, Room, Bryan, & Barrows, 1998) pointed out that mutual help services that are based on peer support cannot be adequately understood from the vantage point of researchers and professionals alone. Understanding the rationale for these services, their operations, and ways they are helpful must draw on the experiences and views of participants. Our research on RRs has demonstrated a pluralistic research program designed to understand the many facets of RRs, including resident experiences and perceptions as well quantification of resident outcomes. Qualitative methods have included focus groups, qualitative interviews, and observation of the physical and social characteristics of houses. These methods have helped us understand not only what is occurring in RRs, but also how and why, which has provided information that has informed the development of formal hypotheses addressing outcomes and mechanisms of action.

Quantitative studies that test a priori hypotheses have used different designs, each with strengths and weaknesses. Randomized designs that compare outcomes of individuals receiving different services have the advantage of showing causality. Such designs have been used to study samples of persons entering RRs after leaving controlled environments such as residential treatment (Jason et al., 2007) and criminal justice incarceration (Jason et al., 2015). Randomization has also been used to study an add-on intervention, (motivational interviewing case management) after individuals enter RRs (Polcin, in press). In the first two instances the RRs operated as a type of aftercare or post-release intervention after release from a controlled environment. In the other randomized study, the add-on intervention is being studied in terms of effects on outcome, not the effects of the RRs.

There have been no randomized trials that included samples of all persons entering RRs. Such studies would include individuals without recent residential treatment or incarceration. As described elsewhere (Polcin, in press), there would be a number of challenges in such studies. Refusal rates for participation in the research would likely be much higher than that of persons leaving a controlled environment, which would create problems with generalization of results.

Unlike residential treatment or criminal justice populations, there is no readily available comparison group (e.g., aftercare or probation or parole as usual). These limitations indicate that there is an important role for descriptive longitudinal and other quasi-experimental studies that document outcomes over time. Although such designs do not prove that the improvements residents make are due to their residence in a RR, there are a variety of assessment strategies that can be used to increase confidence about the role of RRs in influencing outcomes (Polcin, in press). These include multivariate techniques that can parse out the relative influence of various factors in predicting outcome over time, testing whether theoretically relevant variables are significant predictors of outcome, and conducting propensity score matching (PSM) analyses. PSM estimates the effect of an intervention by controlling covariates that predict receiving the treatment versus not receiving it. The challenge of using this procedure in research on RRs is that it can be difficult to identify all of the potential factors that might influence entry into a recovery residence. For a more complete description of PSM methods see Ye and Kaskutas (2009).

Coordination of Study Procedures

A major challenge in RR research is coordination of research procedures across multiple houses in different locations. Houses can vary by size, mix of residents, services offered, location, and neighborhood characteristics. There are many types of recruitment strategies that researchers can pursue, but they need to pay careful attention to their selection of houses, types of residents within houses, and nesting effects of individuals within houses. In general, researchers either need to recruit enough houses to implement multilevel designs that assess house differences of ensure that all houses are similar enough to make differences inconsequential. Whatever is decided, specification of the houses sampled and the limitations of that sample need to be explicit in the dissemination of findings.

There can also be significant logistical issues related to the fieldwork when recruiting residents from multiple houses, particularly when these houses are not in close proximity to one another. For example, in an ongoing study of RRs in Los Angeles researchers travel to and from the homes to conduct research interviews and deliver an add-on intervention that is part of the study has been time consuming due to traffic congestion and the large geographic area. There can also be challenges in terms of finding an appropriate place to meet to conduct research interviews. Houses often do not have offices and ensuring sufficient privacy can be an issue. This contrasts markedly with studies that take place at treatment programs where office space may be readily available. Potential options that researchers can consider to address these issues include renting additional office spaces near houses, pursuing part-time rental space from existing programs located near houses, conducting research and intervention interviews by phone when in-person meetings are not feasible and meeting with residents in public places when they provide sufficient privacy (e.g., coffee shop or park).

Participant Retention

Longitudinal designs are essential to assessing resident outcomes. Barriers to collecting longitudinal data on substance abusing and recovering populations include transient and precarious housing situations, difficulty obtaining accurate contact information from government and public databases, and a lack of familiarity with the communities in which participants reside (Gilliss et al., 2001). Following up with individuals in RRs poses additional challenges because many are involved in the criminal justice system. Because many come directly from incarceration or from residential treatment, they do not have a stable address where they can be located after they leave the RR. In addition, many have been cut off from family and friends during incarceration, which limits the number of persons that researchers can contact to locate participants. Despite these challenges, those researching RRs have generally been able to achieve acceptable follow-up rates even when residents were followed for more than a year after their baseline interviews (Jason et al., 2006; Jason et al., 2015; Polcin, Korcha, Bond, & Galloway, 2010a).

A number of resources have been developed to increase follow-up rates with substance abusing and recovering populations (Hall et al., 2003) (Scott, 2004). We have found these all to be invaluable resources. However, some things that we have found particularly useful are rapport-building at recruitment, extensive collection of collateral contacts (family members and friends of the participants who may be more likely to own property, have cell phones, social networking accounts, or online presences, and be easier to reach), provision of trinkets (e.g., a key chain, pen, calendar), use of graduated incentives, maintaining contact between interviews, use of social media, ongoing and rotating use of multiple paid and publicly available databases, and employing culturally competent trackers. Although there are generally no set lengths-of-stay in RRs, they are typically used by residents as a step on the way to living independently on their own or back with their own families. As such, the researcher may be recruiting from the RR but that may not where the resident will be at follow-up.

During the recruitment process, researchers need to ensure that residents understand that participation in the research involves follow-up contacts (regardless of where the resident may be living). Residents who will not be available for follow-ups or are unwilling to provide personal and collateral contact information may not be appropriate for the research study. Providing residents with a trinket that includes a way to get in touch with researchers at the outset can serve as a reminder about the study and that researchers will be following up with them. Including study information on a project website or social media page (Mychasiuk & Benzies, 2012) can also facilitate the resident contacting the researcher with updates to their locating information. Researcher-initiated contact between interviews (at a set midway point or via birthday and holiday cards) is another useful way to keep residents engaged with the study and to verify or collect additional contact information for when their follow-up interview is actually due. We have also had success using paid and publicly available data bases. However, it is important to note that many databases scan real property records, credit reports, and published telephone records, and transient and at-risk individuals are less likely to purchase property or take out lines of credit which can often limit the relevance of these databases. Thus, it is important to include a variety of free and paid databases. Table 1 identifies resources that we have found helpful using a regular rotation to ensure collection of the most current information (Callahan & Jason, 2013).

Table 1.

Online Resources.

Government Databases

Pacer.gov

Cookcountyassessor.gov

Vinelink.gov

Publicrecordcenter.com(portal to nationwide state and federal databases)

Property records

Ssnvalidator.com(social security number validator)

Skipmax.com

Masterfiles.com

Merlindata.com

Social Networking

Whatsmyipaddress.com

Tracersinfo.com

Didtheyreadit.com

Melissadata.com

Facebook, linkdin: You may search a telephone number on these networking sites to find accounts linked to that number.

It is critical to employ research staff who have familiarity with the neighborhoods in which participants and collaterals reside. Attending recovery oriented community events and meetings can be a way to locate some study participants. However, research staff must be willing and able to engage with residents and their collaterals in the community wherever that may be.

Rewards of Recovery Residence Research

Despite the challenges involved in conducting research on RRs, we want to remark on the importance of this work and the personal gratification that comes with it. The research conducted thus far on RRs underscores immense promise for this work and the favorable results that we have found for individuals with limited resources have been personally gratifying. For example, many individuals with substance use disorders have found themselves in criminal justice settings and our work has shown that RRs can be useful resources for many of these individuals (Jason et al., 2015; Polcin, 2006). In addition, we are currently developing interventions that can enhance recovery homes so they are more responsive to the needs of ex-offenders (Polcin, Korcha, Bond & Galloway, 2010c). Research staff working with residents referred from the criminal justice system hear remarkable stories of recovery and successful adaptation to the community after months or even years of incarceration. Even as an objective researcher, it is difficult not to be moved by stories of residents in these houses who are going gains in employment, reconciling with family members and friends, and getting a second chance to accomplish so many things that fell to the way side when addiction took hold of their lives.

Documenting resident successes leaves many researchers with a sense of personal gratification of doing the work; it is a reminder that those in the most need can achieve remarkable transformations in spite of the systemic and personal challenges they face. However, we also hear about the struggles and hassles of everyday life, and also tragedy and setbacks, such as rearrests and re-incarceration. Accurate documentation of ways that residents are struggling can nevertheless be helpful because we are providing information about the limitations of RRs and areas where they may need to be improved. Researchers can play vitally important roles in terms of articulating how RRs should be used, what type of RR is best for different types of individuals, and ways residences might be improved. Our motivation to study RRs is enhanced by the commitment of RR operators and providers who so passionately, and despite great obstacles, open or otherwise foster recovery communities in their houses (Mericle et al., in press) (Troutman, 2014). Researchers can support their efforts by disseminating studies showing favorable resident outcomes and support from local neighbors and communities (Jason, Roberts, & Olson, 2005; Polcin, Henderson, Trocki, Evans, & Wittman, 2012).

Expanding Research on RRs

A recent review of the evidence base for recovery housing noted that, although studies have consistently shown positive outcomes, replication of study findings with greater specificity and in more settings is needed (Reif et al., 2014). There is an urgent need to expand the evidence base on RRs to include research on various types of RRs and in various geographic locations. In addition to increasing the focus and geographic diversity of studies, studies must also begin to address more nuanced questions about RRs—questions moving beyond whether RRs work but how they work and what type of recovery residence works best for whom. Although some cost-benefit analyses on recovery residences have been favorable (Lo Sasso, Byro, Jason, Ferrari, & Olson, 2012), there is a need for more research in this area that can be used to influence policymakers to fund recovery houses. Addressing these types of questions will lead to the development of evidence-based practices for RRs more generally and to enhancements to services for more vulnerable, potentially harder to serve, populations in particular.

One issue needing more attention is the question about the length of time in the home that is necessary for different residents to maximize beneficial outcome. For some residents who prematurely leave RRs, dropout leads to relapse. We do not yet know which recovery home characteristics are associated with optimal lengths of stay and sustained recovery. To address these important questions, research is needed that conceptualizes recovery homes as evolving social networks that vary in their ability create and maintain residents’ social integration. We need to better understand how social networks within the homes help residents remain long enough to learn how to maintain their sobriety. These dynamics could then be linked to changes in mediating outcomes such as abstinence self-efficacy, house dropout, costs and benefits, and, ultimately, relapse or continued abstinence. In addition, many persons residing in RRs have a variety of problems in addition to substance abuse, such as homelessness, past criminal justice involvement, and other chronic illnesses. Therefore, researchers with these interests have samples of persons living in RRs that can be recruited for their investigations.

Research must also address how RRs should be integrated into the existing formal service delivery system. RRs provide recovery-supportive housing but have historically operated outside the formal substance abuse and housing continua of care. Reviews highlighting the need for and the effectiveness of recovery support services (Laudet & Humphreys, 2013) have included RRs in their discussions of how the acute-care approach substance abuse treatment needs to expand into a more recovery-oriented system of care, but barriers to funding recovery support services persist, potentially limiting the promulgation of these services. These types of discussions are particularly timely given recent changes in funding of substance abuse treatment more generally with the passage of the Mental Health Parity and Addiction Equity Act and the Affordable Care Act (McLellan & Woodworth, 2014). Similar discussions need to be taking place within the housing services system about how RRs fit within programs administered by the US Department of Housing and Urban Development (HUD). Work in this area would be enhanced by research highlighting the housing needs and housing outcomes of those living in RRs and by economic evaluations of RRs.

Conclusion

Although the number of RRs are rapidly growing, many addiction treatment practitioners have limited knowledge about them. Even fewer researchers know about them and the number of investigative teams studying them are limited. Because RRs are not well understood among potential funding agencies it can be a challenge to acquire the resources needed to study them. One way of addressing this challenge in grant applications is to emphasize the innovation and underutilization of RRs. The case for the significance of studying RRs can be framed in terms of a response to the growing trend in the addiction field to emphasize services in the community that can help sustain long-term recovery.

We have reviewed a variety of challenges facing researchers who undertake in this work along with ways these challenges can be and have been addressed. In addition to funding issues, we have discussed considerations for research designs, coordinating research procedures across multiple sites, and finding residents for follow up interviews. Of particular importance is forming a collaborative alliance with RR organizations and the individuals in leadership positions within houses. RRs emerged as a grassroots movement rather than a professional derived intervention and understanding their operations and the ways they are beneficial to residents requires input from providers and residents.

This paper has identified several ways to look at the rewards of this work. First, collection of objective data on operations and outcomes is gratifying because it highlights the utility and effectiveness of RRs as a substance abuse service that can address one of the most urgent goals confronting the field, sustaining long-term recovery. Second, it is equally gratifying to identify limitations and areas where RRs can be improved because this will ultimately improve services delivered in RRs and resident outcomes. It can be difficult for providers to confront these limitations. We are only beginning the process of describing the limitations of RRs and identifying what type of resident is best for what type of RR. However, there are potentially enormous rewards for RRs and the researchers who provide objective data that can be used to target different types of RRs to specific resident characteristics and modify operations to maximize resident outcomes. Perhaps the most rewarding aspect of this work comes from our interactions with residents. The stories they share with us about their struggles with addiction, celebrations of recovery, and hopes for the future are compelling. Fully aware of how challenging this can be, we hope that other researchers will join us in this important and rewarding research on RRs so that the field can better meet the diverse and ongoing needs of individuals in recovery from addiction.