More than 400 cities, counties, and states have joined the federal Interagency Council on Homelessness' movement to adopt ten-year plans to end homelessness. This initiative has spurred some communities into adopting evidence-based practices to end homelessness among people with serious mental illness. The council lists "housing first" as an innovative and promising practice for communities to consider for ending homelessness in this high-need population (www.ich.gov) (1). The housing-first model does not stipulate preconditions for obtaining housing, such as being "clean" from drugs or alcohol before obtaining housing, and it does not require adherence to mental health or substance abuse treatment in order to retain housing. Rather, in the housing-first model, participants are offered housing unconditionally; treatment and other services are made available, although adherence to those services is left up to participants. This contrasts with the traditional supportive housing model, which requires treatment adherence as a condition for obtaining housing and maintaining tenancy (1).

In 2003 the U.S. Department of Housing and Urban Development (HUD) sponsored research on the housing-first model for homeless persons with serious mental illness. Results showed that most housing-first participants have a history of severe psychiatric impairment and chronic homelessness. Additionally, the HUD research found that housing-first participants' tenure in the program and involvement in the criminal justice and mental health systems, as well as their substance abuse and mental health histories, are closely related to where they "came from" when they arrived at the housing-first program—for example, from the street, a jail, a hospital, a supportive housing program, or a shelter (2).

Other recent research has found housing to be the most frequently used form of leverage to secure adherence to treatment by persons with mental illness. In one survey of more than 1,000 persons in outpatient treatment settings across the United States, lifetime experiences with being told that housing would be provided if—and only if—one adhered to treatment ranged from 23% to 46% (3). Housing was most often used as leverage to secure treatment adherence with males who had a history of psychiatric hospitalization and who lived in special housing designated for people with mental illness or substance abuse. These individuals were more likely than people without mental illness to have experienced a wide range of other forms of leverage to adhere to treatment, such as having a representative payee control their disability benefits, receiving treatment in lieu of incarceration by the criminal justice system, or having a court-ordered outpatient civil commitment (4).

This study of five housing programs—three housing-first programs and two supportive housing programs—addresses the way in which the housing-first and the supportive housing models are operationalized in practice. More specifically, the research addresses three questions. Is the type of program—housing first or supportive housing—related to an explicit requirement that residents adhere to mental health and substance abuse treatment? Regardless of whether treatment adherence is an explicit requirement of housing, is the type of program related to whether residents perceive that they are being coerced to accept treatment? Is the type of program related to residents' satisfaction with their housing?

Two convenience samples of residents were interviewed for this study. The first sample consisted of 59 residents from the three housing-first programs in the HUD study: Pathways to Housing located in New York City (N=23), Reaching Out and Engaging to Achieve Consumer Health (REACH) located in San Diego (N=19), and Downtown Emergency Service Center (DESC) located in Seattle (N=17) (2). The resident interviews reported here were not part of the HUD study, and these findings have not been reported elsewhere. The second sample consisted of 77 residents from two supportive housing programs in New York City, Project Renewal (N=47) and The Bridge, Inc. (N=30). Structured one-hour interviews were conducted by trained interviewers in the participant's residence whenever possible and took place between October 2004 and May 2005 for the first sample and between October 2005 and January 2006 for the second sample.

All participants were asked two questions about whether they had been explicitly told that adherence to mental health and substance abuse treatment was a requirement of obtaining housing (yes-no response), four questions about whether they perceived that treatment was being coerced (possible scores range from 1, strongly disagree, to 5, strongly agree), and one question about whether they had ever lived in a place where housing was used as leverage to obtain treatment adherence (yes-no response). They were also administered a 19-point scale on housing satisfaction (5). Participation was voluntary, informed consent was obtained, and participants were assured that taking part or declining to take part in the study would have no impact on their housing. The research received institutional review board approval from Policy Research Associates, Inc.

The characteristics of the 136 participants across the five sites show some similarities in that most were male (16 males, or 70%, in Pathways to Housing; ten males, or 53%, in REACH; 14 males, or 82%, in DESC; 28 males, or 60%, in Project Renewal; 18 males, or 60%, in The Bridge), most were 40 or older (N=101, 74%, ranging from 13 persons, or 57%, at Pathways to housing to 15 persons, or 88% at DESC), and most were in the program for six months or longer (data not gathered at Pathways to Housing; 12 persons, or 63%, at REACH; 16 persons, or 94%, at DESC; 41 persons, or 87%, at Project Renewal; and 29 persons, or 97%, at The Bridge) (χ²=12.56, df=3, p<.01). The major difference among sites is the race of participants from the programs (χ²=12.96, df=6, p<.01). Both REACH and DESC were similar in terms of the race of their participants (REACH: ten persons, or 53%, were white and four persons, or 21%, were black; DESC: nine persons, or 56%, were white and four persons, or 25%, were black). A higher proportion of participants from the two New York City sites were black (22 persons, or 49% at Project Renewal; 13 persons, or 43% at The Bridge). However, a higher proportion of participants at The Bridge were white, compared with the proportion of participants of Project Renewal (11 persons, or 37%, at The Bridge and eight persons, or 18%, at Project Renewal). Pathways to Housing did not report the race of participants. (Data for race were not available for all participants.)

Table 1 lists the specific questions participants were asked and shows the mean scores of respondents by program. An explicit requirement of mental health treatment adherence as a condition of obtaining housing was significantly more likely to be reported by supportive housing residents than housing-first residents. Housing-first residents were also more likely to report that substance abuse treatment was a condition of obtaining housing, although this finding was not significant (p=.059). Also, housing-first residents were significantly less likely than supportive housing residents to report a stated requirement of continued treatment adherence being made a condition of housing.

Table 1 also shows that housing-first residents were significantly less likely than supportive housing residents to perceive that their housing was contingent on adherence to mental health and substance abuse treatment. Housing-first residents were more likely than supportive housing residents to agree that they could stay in their current housing if they discontinued mental health services, although this finding was not significant (p=.051). There were no differences between the programs when residents were asked whether they could stay in their current housing if they broke program rules about substance use.

A 19-point scale (5) was used to ask participants about their satisfaction with their current housing, and no significant difference was found in housing satisfaction across the five programs.

Although there are some acknowledged limitations to generalizability introduced by the convenience sample that provides the basis for this research, results showed that compared with residents in more traditional supportive housing programs, those in housing-first programs were significantly less likely to report that adherence to mental health treatment was an explicit requirement of obtaining housing and less likely to report that mental health and substance abuse treatment was a requirement of retaining housing. The housing-first philosophy of meeting residents "where they are" at the moment, whether in treatment or not, is being met in practice according to resident reports in this study. In addition, the fear that residents in housing-first programs would be more likely to engage in substance abuse because of looser program rules was not borne out by the finding there were no differences between housing-first and supportive housing residents when they were asked whether they could remain in their current housing if they broke program rules about alcohol or drug use. It appears that housing-first residents, like supportive housing residents, may be getting a clear message that the program may not require treatment or sobriety as a precondition for offering housing or for retaining the tenant, but the program also does not endorse substance use.

Given the strong and persistent relationship between serious mental illness, substance abuse, and chronic homelessness, innovative programs must be developed and implemented if the goals of the Interagency Council on Homelessness are to be reached by 2010. The innovative housing-first model, at least in the sites studied here, appears to be implemented in practice in a manner consistent with its guiding "no coerced treatment" philosophy. The housing-first programs achieved a level of client satisfaction comparable to that of supportive housing programs.

This research was supported by the Network on Mandated Community Treatment of the John D. and Catherine T. MacArthur Foundation. The authors thank Tariqul Islam, B.A., for his analytic support and Amy Thompson, M.S.W., for her dedication to the coordination of data collection in New York City. The authors also acknowledge that this research would not have been possible without the effort of the interviewers and the cooperation of the five site leaders: Sam Tsemberis, Ph.D., Gary Hubbard, M.F.T., Daniel K. Malone, M.P.H., Craig Retchless, and Jim Mutton, M.S.W.

The authors report no competing interests.