Homelessness and Mental Health

The At Home/Chez Soi study was a multisite randomised controlled trial demonstrating the effectiveness of a “housing first” intervention in Canada. The “housing first” model is related to the Pathways to Housing model that originated in New York City. The concept is to immediately provide permanent housing at a location largely chosen by the participant and also to provide wrap-around supports.¹ Positive results from the At Home/Chez Soi trial were previously summarised in The Canadian Journal of Psychiatry (CJP) by Aubry et al.² People receiving the housing first intervention had superior housing outcomes and showed benefits in functioning and quality of life compared to a treatment-as-usual comparator. Such interventions, of course, are not a panacea. Earlier this year, Adair et al.³ reported the occurrence of both positive and negative housing trajectories in the housing first and treatment-as-usual arms of At Home/Chez Soi. In the current issue of the CJP, Aquin et al.⁴ examine 2 additional outcomes, suicidal ideation and attempts, in the At Home/Chez Soi cohort. Suicidal ideation was found to diminish over time, but no differences between the intervention and control groups were observed. No differences were found in suicide attempts either. The improvement in suicidal ideation may be due to regression to the mean. At Home/Chez Soi recruited its participants from shelters, hospitals, and agencies. At the time of recruitment, the participants may have had higher levels of suicidal ideation than is usual for them, something that could cause regression to the mean as those participants returned closer to their long-term baseline over time. Resources available to the respondents in both study groups (the treatment-as-usual group was not denied services) may also have led to improvement over time. Another possible explanation is the “Hawthorne effect”: the tendency for better outcomes to occur when one is aware that one is being observed, as in the context of a study.

Another study published in this issue looked at suicides among homeless and precariously housed residents of Toronto.⁵ Coroner records were reviewed for all suicide deaths in Toronto from 1998 to 2012. The study found that 1.8% of people dying by suicide were homeless, a large number given that only about 0.2% of the Toronto population is homeless at a point in time. Homeless people dying by suicide were younger and more likely to be male than nonhomeless people dying by suicide. They were also more likely to have been seen by an emergency department or outpatient psychiatrist in the week prior to their death and to have died by falling or jumping. These results provide a descriptive comparison of suicides by homeless and nonhomeless persons, which is valuable information for formulating preventive strategies. However, since inclusion in the study was restricted to those people who died by suicide, the study could not estimate or compare suicide rates in different population sub-groups. Because there are many causal pathways to suicide (it is a common outcome upon which many causal pathways “collide”), associations between various predictors are subject to distortion, and therefore are not reported.^6,7

People who are marginally housed, something that is often associated with severe poverty, share many of the challenges of homeless persons. Marginal housing often coexists with extreme poverty, with both having major implications for health care delivery. These problems led to the initiation of the Hotel Study, findings from which are also reported in this issue.⁸ The study looked at data from a variety of sources, including hospital records, police records, and a prospective cohort study following 375 residents of Vancouver’s downtown east side between 2008 and 2015. The authors used a standardised mortality ratio (SMR) to quantify mortality in the cohort. This involves counting the observed and expected number of deaths in a cohort. The expected number of deaths is estimated using available age- and sex-specific mortality rates for the Canadian general population. The SMR is then calculated by dividing the observed number of deaths by the expected number of deaths. In this case, the SMR was 8.4, highlighting the vastly elevated risk of death in this population. Management of human immunodeficiency virus (HIV) was found to be effective, whereas management of hepatitis C and psychosis was less so. A key finding was that disease-specific management did not seem to convey benefits across different diseases for participants having multimorbidity. The authors call for the development of innovative services that can better deal with multimorbidity as opposed to relying on disease-specific management strategies.

It is important to have the kinds of data reported in these studies in order to develop more effective services and policies. However, it is also important to be able to measure health status in homeless, marginalised, and precariously housed people. Without an ability to monitor the health status of this population, how could the impact of interventions and other changes over time be assessed? Unfortunately, descriptive data have been scarce. Historically, available evidence has mostly come from small, local studies. This issue contains a report of the 2015 National Canadian Homelessness Youth Survey, the largest effort of its kind to describe the health status of homeless youth in Canada.⁹ Accomplishing such a study is a very difficult challenge. Most health surveys start with a list, or sampling frame, representing all members of a target population. They then select, usually through a random selection process, a sample from this sampling frame for participation in the survey. Typical sampling frames for population-based studies are based on lists of telephone numbers, addresses, or administrative databases, but no such frame exists for homeless people. The survey sought to address this challenge by working with 57 agencies serving homeless youth in 42 communities across Canada. They were successful in collecting data from 1103 youth, representing Nunavut and all Canadian provinces except Prince Edward Island. Unfortunately, a need for brevity required the collection of data on mental health status to be limited to a questionnaire-style interview. Nevertheless, a disturbing picture is presented: 42% of their participants reported 1 or more suicide attempts, and 85% had high levels of psychological distress.

As all of these studies demonstrate, issues of homelessness are deeply intertwined with those of mental health. These are important concerns for Canadian psychiatrists. The CJP is proud to present this series of studies and to participate in the advancement of knowledge, policy, and practice related to these very difficult problems.