10 ‘Take Aways’ From the Final Report of the At Home/Chez Soi Study

Earlier this month, I wrote a blog post titled 10 Things to Know About the At Home/Chez Soi (AHCS) Study, a homelessness study which I argue is one of the most ambitious randomized controlled trials in Canadian history. The intent of that post was to provide background information about the report prior to its release. Now that the final report has been released, I have compiled another ‘Top 10’ list.

But first, a bit of background:

• All participants in this study were homeless at the start of the program—they were either sleeping in emergency shelters or sleeping outside when the study began. On average, each study participant had experienced a lifetime total of five years of homelessness upon enrollment in the study.

• All participants also “had one or more serious mental illness.” Among other things, this means that this was not a representative sample of Canada’s homeless population, nor was it intended to be. (The Street Health Report 2007, which surveyed a representative sample of Toronto’s homeless population in 2007, found that just 35% of their sample had received a mental health diagnosis from a physician at some point in their lifetime.)

• Study participants were recruited in Vancouver, Winnipeg, Toronto, Montreal and Moncton.

• Participants were interviewed every three months over a two-year period.

• To assess the extent to which participants were using publicly-funded services in the justice, health and social services sectors outside of those provided by the study team, information about the use of such services was collected from national and provincial administrative data sources, as well as from study participants themselves. The information estimated service use by participants both before and after the study started.

• Members of the treatment group received assistance finding an apartment and were provided with a rent supplement (i.e. money to assist with the rent so that no member of the intervention group paid more than 30% of income on rent during the study). The financial assistance in question, known as a “rent supplement,” was generally between $400 and $500 per month. Members of the treatment group were also offered professional health and ‘social work’ support (in some cases medium-intensity, and in other cases higher-intensity).

• Most of the housing used by members of the study’s treatment group was owned and operated by for-profit landlords.

Against this backdrop, here are 10 ‘take aways’ from the final report of the AHCS homelessness study:

1. Histories of both trauma and head injuries are common among homeless persons. At baseline, 38% of participants reported having been sexually abused in childhood, while 55% reported having been physically abused as children. Moreover, almost two-thirds of participants reported “a history of one or more traumatic head injuries involving unconsciousness.” For more on the risk factors associated with homelessness, see this 2009 background document. For more on traumatic brain injuries among homeless persons, see this 2012 piece.
2. A large proportion of homeless persons suffer from chronic physical health problems. I was surprised to learn that more than 90% of study participants “had at least one chronic physical health problem.” Admittedly, it was more than a decade ago that Dr. Stephen Hwang wrote: “Homeless people in their forties and fifties often develop health disabilities that are more commonly seen only in people who are decades older.” Moreover, findings of this general nature were confirmed in The Street Health Report 2007. Nevertheless, I found this “more than 90%” figure from the AHCS study to be rather astonishing. I was not surprised to learn that at least 85 of the study’s participants are known to have died during the study period. For more on mortality rates among homeless and marginally-housed persons, see this 2009 study.
3. In most cases, providing homeless persons with affordable housing and professional support can be effective at ending a person’s homelessness (keeping in mind that this was not a representative sample of Canada’s homeless population). Having personally spent 10 years working with homeless persons in Toronto (including seven as a mental health outreach worker) I did not need the results of a $110-million research study to be convinced of that. But only two randomized controlled trials of this nature had ever been conducted; each had fewer than 100 participants, and both had been done in the United States. As the report’s authors note in the AHCS final report: “Given the difference in social policy and health care delivery between the U.S. and Canada, it is vital that evidence about homelessness interventions be grounded in the Canadian context.”
4. People who consume large amounts of alcohol and drugs can still maintain their housing. Researchers found that study participants “with substance use problems at baseline maintained stable housing to a similar degree as the overall sample.” I think this helps to debunk the ‘old school’ notion that homeless persons must first ‘rehabilitate’ before being provided with access to permanent housing. For a recent article that explores the provision of housing to homeless persons who are heavy users of drugs or alcohol, see this 2013 article. For an open-access on the same topic, see this 2011 piece.
5. Not only is it cost effective to provide a homeless person with affordable housing and professional support, it sometimes even saves taxpayers money. For participants who were costing the broader social support system very considerable sums of money before the study began, the research team has calculated that, from a straight accounting perspective, providing such individuals with housing and professional support can actually save taxpayers money (a point made quite convincingly by Malcolm Gladwell in February 2006). Put differently, though it can cost between $14,000 and $22,000 annually* to provide a homeless person with affordable housing and professional support, for some homeless persons, this approach will result in savings on other publicly-funded services over and above that amount (on such things as psychiatric hospital stays and time spend in detention centres). For the 10% of participants who were using the most services upon enrolment in the AHCS study, every $1 invested in housing and professional support during the course of the study resulted in average savings of just over $2. And across all study participants, every $1 invested in housing and professional support resulted in $0.75 in savings on health, justice-related and social services. (For more on the costs of homelessness, see this 2005 report and this 2012 report.) Even with this evidence of savings in hand, however, there remains a key political challenge when it comes to implementation of such programming. As my colleague Paul Dowling points out, when it comes to providing affordable housing and professional support to homeless persons, “the department that pays and the department that saves are often not the same department.”
6. Large urban centres in Canada already have programs that respond to homelessness; some of these programs, though underfunded, appear to be somewhat effective. Members of this study’s control group, in effect, got the short end of the stick, having been randomly placed into the group that did not receive help finding an apartment from the study team, did not receive a monthly rent supplement from the study team and did not receive professional health and social-work support from the study team. Yet, 31% of these individuals still managed to find housing for the duration of the study through status quo programs (i.e. social programs that were already in place before the AHCS program began); another 23% of them managed to be housed for part of the time. While it should be underlined that they were ‘outperformed’ by the treatment group (members of the treatment group spent more than twice as much time in stable housing than did members of the control group) to an extent I think the success of the control group speaks to the partial effectiveness of homelessness programs that were already in place well before the AHCS study began. In other words, it’s not as though there were no programs that responded to homelessness before the AHCS study got started! For an overview of the development of these programs in the Toronto context, see this 2009 policy report. Just how underfunded are homelessness programs in Canada? As I’ve said before, annual federal funding for homelessness in 2014 (after adjusting for inflation) represents just 35% of its 1999 level.
7. Even with housing and professional support, not every homeless person can maintain their housing. Sixteen percent of the study participants who received help finding an apartment, were offered financial assistance to pay for the apartment and were offered health and social-work support were not able to hang on to an apartment for any significant length of time (in fact, some of them remained homeless the entire time). Going forward, what public policy response is warranted for this group of people? Do they need supportive housing with 24-hour-a-day on-site support, such as that offered at Strachan House in Toronto?
8. Members of this study’s treatment group did not receive one treatment; they received several. Likewise, the study’s control group was not a control group in the narrow sense. Participants in this study’s treatment group were given a bundle of things—they were placed into an apartment, they were given a considerable amount of financial assistance to afford the unit, and they were provided with either moderate- or high-intensive services by a team of health and social-work professionals. Which of these three factors was the largest determinant of their housing success? I suspect we’ll never know. I do not mean this as a criticism of the study; rather, I think it’s simply worth underlining that interpreting the results of a randomized controlled trial is not always a straightforward process. Likewise, we know that members of the control group did not receive the specific intervention offered by the AHCS research project; but there was nothing stopping members of the control group from receiving affordable housing and professional support from another service provider in the city in question. For example, in Toronto, many member s of the control group may have become housed by (and received professional social-work support from) Toronto’s Streets to Homes program. Other members of the control group may have been housed by supportive housing providers who also provide permanent housing and professional support. That said, my guess is that fewer than 10% of members of the control group received affordable housing and professional support on par with the type offered to members of this study’s treatment group. Put differently, even though the control group may not have been a control group in the narrow sense (I am told that is why it was referred to as the study’s “treatment as usual group” as opposed to the study’s “control group”) that does not mean that the study’s findings should not be taken very seriously.
9. The effectiveness of non-profit housing was not a focus of this study. As I argued last October, providing a homeless person with immediate access to permanent housing without requiring that the individual ‘rehabilitate’ (often known as the “housing first” approach) has existed at least since homelessness became a prominent focus of public policy in the 1980s. In the 1980s in Toronto, such housing and professional support was usually provided by non-profit housing providers (such as Houselink Community Homes and Homes First Society). Non-profit housing providers were not part of the “intervention” in the present study, which might lead some naive observers to conclude that non-profit housing providers should not be viewed as part of the solution to homelessness. I would dispute such a suggestion. As I’ve blogged about here, there are advantages of having not-for-profit landlords (as opposed to for-profit landlords) own and operate housing for low-income individuals. I think this underlines the need for future research on the provision of permanent housing by non-profit housing providers who have a history of providing supportive housing (that is, permanent housing combined with professional support) to this population.
10. There may have been a Hawthorne effect (admittedly, this one’s more for the research wonks out there!). If I knew I was part of a study’s treatment group (and was reminded of this every three months by a researcher) and that the goal of the study was to assess the extent to which a well-funded bundle of services and supports would keep me housed, that knowledge in and of itself might provide me with added incentive to maintain my housing. In other words, take the cameras away—and in this study, there were quite literally cameras in some cases—and I might not try as hard to be a success. Conversely, if I knew I was part of a study’s control group (and was reminded of this every three months by a researcher) the knowledge that I was part of the control group might make me feel less confident about securing and maintaining my own housing. Ergo: it’s possible there was a Hawthorne effect on study participants that exaggerated the success that the intervention might actually have in the ‘real world.’ That said, it’s not clear to me how exactly the research team could have realistically gone about measuring or controlling for a Hawthorne effect. At some point in the interview process (possibly near the end of the process) I would have liked it if the researchers had asked members of the treatment group questions such as: “Since you know you were in the treatment group, can you discuss the impact that that knowledge has had on your recent behaviour? Has it motivated you to try harder to keep your housing?” To members of the control group, I would have liked it had they asked something like: “Since you’ve known all along that you’re a member of this study’s ‘treatment as usual group,’ did this knowledge affect your behaviour in any way? Do you think it discouraged you from finding housing?”

In Conclusion. I wish to commend the Mental Health Commission of Canada and the At Home/Chez Soi researchers for carrying out this massive research project. I believe that the findings of both the final report and the many other shorter articles coming out of the project will help Canadians to better understand policy responses to homelessness. Members of the international research c ommunity will also no doubt find this research helpful as they move forward in figuring out appropriate policy responses to homelessness. Dr. Paula Goering (the At Home/Chez Soi Lead Researcher) deserves especially strong praise for her role, along with her large team of researchers, the more than 2,000 research participants, the more than 200 service providers and the more than 260 landlords and property management companies that took part in this initiative. Finally, I wish to note that I received extremely helpful feedback on an earlier draft of this blog post from several individuals whose anonymity I wish to preserve.