You can download a brief on the study design and results here.
By Bevin Croft, M.A., M.P.P., Research Associate, Human Services Research Institute
Those of us who are concerned about the state of the behavioral health service system would agree that voluntary, cost-effective services and supports that preclude the need for coerced or institutional treatment should be widely available. Peer respites may be one component of such a system.
This month, two of our nation’s most respected publications, the New York Times and the Journal of the American Medical Association, published opinion pieces calling for a return to institutional settings for people diagnosed with serious mental illness. Both are poorly reasoned, casually dismiss the rights and preferences of an entire subset of the human population, and play on public fears about a link between violence and mental illness. It seems these writings may have made it to the pages of such hallowed publications because they are part of a perfect storm of political forces that include Rep. Tim Murphy’s efforts to expand assisted outpatient treatment (AOT) and dismantle the Substance Abuse and Mental Health Services Administration (SAMHSA), a damning report from the Government Accountability Office (GAO) criticizing SAMHSA and other federal agencies, and a growing concern that prisons and jails play an increasingly de facto role in the behavioral health service system. There’s no doubt that the system needs improvements, but calls for institutional treatment ignore decades of research documenting the effectiveness of community-based services and supports that, if adequately funded and properly implemented, enable people diagnosed with serious mental health conditions to live rich lives and contribute to society in meaningful ways.
As my colleague Laysha Ostrow and I describe in an Open Forum, published online in Psychiatric Services this month, peer respites are short-term residential programs that offer trauma-informed peer-to-peer support for individuals experiencing extreme states. Peer respites are designed for individuals who may be heading toward a mental health crisis that might otherwise lead to an inpatient hospital stay. These programs are largely unstructured and explicitly non-clinical, focusing instead on creating an environment for guests and staff to foster mutual healing relationships. Peer respites create space for guests to move through extreme states in a manner of their choosing. Many use Intentional Peer Support, a trauma-informed practice that focuses on mutual learning relationships and social change, as an organizing framework. Currently, there are 16 peer respites operating in the United States, and four more are in the planning stages. Although a fair amount of research has documented the effectiveness of peer support and residential crisis alternatives, relatively little has examined peer respites in particular. One 2008 study of a peer respite documented higher satisfaction, improved symptoms and social functioning, and higher self-esteem for peer respite users compared with individuals committed to an inpatient hospital.
In the same issue of Psychiatric Services, my colleague Nilufer Isvan and I published some early findings of an evaluation of the Second Story program, a peer respite in Santa Cruz, California. Second Story is one of the first peer respites in the country; it opened in May 2011. The program is funded through a combination of county behavioral health department funds and a grant from the Substance Abuse and Mental Health Services Administration. The SAMHSA grant is slated to end this year, and as the county determines its funding priorities for the coming years, Second Story’s future remains uncertain.
We included 139 individuals who used the peer respite between May 2011 and June 2013 in the analysis. Instead of randomizing participants in the study to create an “intervention group” and a “comparison group”, we used a statistical method called propensity score matching (the details of this are described in the article) to establish a comparison group of 139 Santa Cruz County residents who did not use the program. We compared the two groups (respite users and those who had not used the respite) to ensure they had similar clinical, demographic, and behavioral health service use characteristics. This allows us to reasonably assume that the two groups are similar enough to draw some conclusions about the effect of the program on service utilization.
First, we examined the likelihood of using inpatient or emergency services after the respite start date. Next, we looked at total hours of inpatient and emergency service use for the 98 individuals in either group who used any of those services. In all of our analyses, we took individuals’ clinical, demographic, and behavioral health service use histories into account.
We found that the respite guests had a 70% lower probability of using inpatient or emergency services compared to the group of similar non-respite users. For the individuals who used any inpatient or emergency services during the study period, a longer stay in respite was associated with fewer hours of inpatient and emergency service use.
A closer look at the results suggests a complex relationship between respite and inpatient and emergency service use. Although respite guests were on average less likely than the non-respite group to use inpatient or emergency services, each additional day spent at Second Story increased this likelihood. That is to say, respite guests with longer stays at Second Story had a higher likelihood of using inpatient or emergency services than guests with shorter stays. Similarly, the length of respite stay had diminishing returns in terms of reduced hours of inpatient and emergency service use, with negligible decreases predicted beyond 14 days of respite.
In the article, we discuss some possible reasons behind these findings. Because these analyses relied solely on county administrative data, it is very likely that factors we couldn’t measure had an impact on the relationship between peer respite and inpatient and emergency service use. Although we had rough measures of housing status, for example, we know from our qualitative work that many respite guests experience instability in housing, and that this instability is an important factor in decisions to visit both peer respites and emergency rooms.
Although the findings had some important nuances, they point to a high degree of promise for peer respites to decrease the behavioral health system’s reliance on coercive and isolating interventions like inpatient hospitalizations.
In the coming year, we plan to repeat the analyses described above with four years of data rather than two. Using a larger sample size, we hope to better understand the relationship between peer respites and inpatient and emergency service use. Also, these results reflect only one part of the Second Story program evaluation. We’ve been working with a group of peers who have been meeting with Second Story guests these past four years to ask about their experience with the program and their lives in general to understand how—if at all—the program helps people to achieve a higher quality of life. To explore these questions, we used both surveys and open-ended interviews to capture the range of possible experiences with the program. If you’re interested in learning more about this kind of research, you can find a toolkit on evaluating peer respites written by Laysha and myself here.
We hope the results of this and other future research will help to answer some of the important questions we uncovered in this first analysis and paint a richer picture of how peer respites can improve people’s lives and contribute to a behavioral health system that is more humane and less traumatizing. This research might help to counter the push to bring back asylums and other such misguided, ahistorical, unscientific calls for reform.