The At Home/Chez Soi project

Mental Health Commission of Canada - Commission de la santé mentale du Canada

The At Home Chez Soi (AHCS) project is a pan-Canadian study on homelessness and mental illness. This $110 million pragmatic randomized controlled trial was funded by the Mental Health Commission of Canada (MHCC) and began recruiting participants in October 2009 simultaneously in five Canadian cities: Vancouver, Winnipeg, Toronto, Montréal, and Moncton.

The AHCS project aimed to provide access to housing to homeless individuals experiencing mental illness under the Housing First (HF) model — a particular approach to developed in New York City by the organization Pathways to Housing.

Jump to: Results | Sub-studies

Housing First – Ending homelessness

The HF model, as implemented in the AHCS project, involved offering people who were homeless and living with a mental illness immediate access to a subsidized apartment of their choice, combined with clinical services. A rental supplement allows participants to live in a decent apartment for 25% or 30% of their income (depending on rent including heating or not). The intensity of clinical support is tailored according to the needs of participants: “high need” participants received Assertive Community Treatment (ACT), while “moderate” need participants received Intensive Case Management (ICM). Services and rental supplements were offered with no conditions, other than agreeing to the visit of a staff person from the program at least once per week.
The project aimed to evaluate the effectiveness and cost-effectiveness of the HF model. Effectiveness was conceived of first and foremost in terms of housing stability, but also in terms of quality of life, mental and physical health status, community functioning, community integration, etc. Recruitment and research follow ups of the participants began in October 2009, with recruitment completed in May 2011, while research follow ups ended in April 2013.
Studies carried out in the United States have found that HF programs significantly increase housing stability, while the costs of the intervention are largely, if not completely, offset by a reduction in spending on acute health care services, shelters, and justice and correctional services.

The At Home/Chez Soi project in Montréal

In Montréal, 469 participants were recruited between October 2009 and May 2011, and assigned to different groups depending on their level of need, either high need (HN) or moderate need (MN). HN participants were randomly assigned either to HF with an Assertive community treatment (ACT) team (N=81), or they were assigned to treatment as usual (TAU, N=82). Participants with moderate needs were randomly assigned to one of the two Intensive Case Management (ICM) teams (N=204) or to TAU (N=102). Participants in TAU formed the control group, and did not receive an HF intervention nor a subsidized housing from the project: they continued to use services already available to them within the community. Over a 24-month period, participants met with researchers every three months in order to complete a research interview. The Montreal site achieved exceptional follow-up rates of 91% overall, which increases the reliability of its findings.

Socio-demographic portrait of participants at baseline
● 44.1 years old: average age of participants;
● 32.3% of them are women;
● 78.4% had worked for at least one consecutive year in the past;
● 29 months: average longest uninterrupted homelessness period;
● 52 months: average time spent homeless during lifetime;
● 35 years old: average age during first homelessness episode;
● 21% have at least one child;
● 40% have been hospitalized more than 2 times within a single year for a mental illness in the past 5 years;
● 29% were involved with the judiciary system during the past 6 months;
● 33% are at moderate or high risk of suicide.

Housing the participants

Participants assigned to HF were quickly offered help in accessing an apartment of their choice (in terms of neighbourhood, type of building, etc.), with a rent subsidy set so that they only had to pay 25 per cent or 30 per cent of their income (the latter if the rent included heat) for the rent. In Montréal, the rent subsidy averaged $375 per month. Participants were able to remain housed after the end of study period, March 31st 2013, thanks to rent supplements from the MHCC and the Société d’habitation du Québec (SHQ).

The teams

Several institutions in Montréal partnered to implement the AHCS project: the CSSS Jeanne-Mance established and managed the ACT team as well as one of the two ICM teams; the Centre Hospitalier de l’Université de Montréal (CHUM) provided the service of psychiatrists and, later, a coordinator for the ACT team; Diogène, a community organization, provided the remaining ICM team; and the Douglas Institute established and coordinated the housing team as well as the research team. While the housing team maintained its own office, the research team was located in the same offices as the CSSS Jeanne-Mance teams in order to facilitate interviews. A community-based organization that includes a large men’s shelter but provides a variety of other services as well, Welcome Hall Mission, also managed the (partial) rent payments to landlords by HF participants who had consented to have their share of the rent paid directly to their landlord from their account.

Results of the At Home project

The final report of the At Home/Chez Soi Project downloaded here, or can accessed via the Mental Health Commission of Canada webpage

The project yielded three key conclusions.

  1. The Housing First approach is feasible in Montreal. The housing team found 73 landlords located in many different neighborhoods of Montreal who were willing to offer one or more of their apartments to Chez Soi participants. Almost all of the 285 participants who were recruited to an HF group were housed within about 3 months of recruitment, in apartments of their choice. The teams continued to provide services to the great majority of participants until the end of the study.
    As a result of sustained implementation efforts, the teams managed to follow aspects of the Housing First model more and more closely. Such higher fidelity to the model was found to be associated with greater effectiveness at the national level.
  2. The Housing First approach is effective. Housing First (HF) participants obtained and retained housing to a much greater extent than those in the treatment-as-usual (TAU) groups. Among high-need participants, in the last 6 months of the study, 60% of those in the HF group were housed all of the time, compared to 31% for the TAU group; and 21% none of the time, compared to 59% for the TAU group. The differences were larger among moderate-need participants: 72% housed all of the time, compared to 29% for the TAU group; and 7% none of the time, compared to 52% for the TAU group. Housing quality was similar on average, but with fewer extremes, for HF residences compared to TAU residences. Furthermore, moderate-need participants in Housing First reported a greater improvement in self-reported quality of life than their treatment-as-usual counterparts. The high-need HF group, and one of the moderate-need HF groups also showed more improvements in measures of interviewer-rated community functioning than comparable TAU groups. In-depth interviews conducted at study entry and 18 months later on a representative 10% sample of both high-need and moderate-need participants reveal many more benefits. Among others, HF participants were much more likely to report improvements in their mental health, decreases in stress and anxiety, greater re-establishment of connections with their family members, and reductions in substance abuse. For people with moderate needs especially, overall life trajectories were much more likely to be positive, compared to those of people receiving services as usual.
  3. Housing First reduces the costs of other services. On average, the intervention cost $22,882 per person per year for high need participants and $14,029 for moderate need participants. Over the two-year period following participants’ entry into the study, every $10 invested in HF services resulted in an average savings of $8.27 for high-need participants and $7.19 for moderate-need participants.

In short, the study has shown that it is feasible in Montreal to house many homeless people with mental illness in subsidized apartments of their choice and greatly increase the quality of their lives, at minimal cost to society.

Researchers involved in the project

Jean-Pierre Bonin, Ph.D. Associate Professor at the Faculty of Housing, Université de Montréal.
Sarah Boucher Guevremont, MSW. Université du Québec à Montréal.
Anne Crocker, Ph.D. Associate Professor, Department of Psychiatry, McGill University; researcher, Douglas Mental Health University Institute.
Henri Dorvil, Ph.D. Professor at the School of Social Work, Université du Québec à Montréal.
Marie-Josée Fleury, Ph,D. Researcher, Douglas Institute. Associate Professor, Department of Psychiatry, McGill University.
Baptise Godrie, Ph.D. candidate in sociology. Université de Montréal.
Guy Grenier, Ph,D. Research Associate, Douglas Institute.
Roch Hurtubise, Ph.D. Director and Professor at the School of Social Work, Université de Sherbrooke.
Eric Latimer, Ph.D. Researcher, Douglas Institute. Professor, Department of Psychiatry, McGill University. Associate Member, Department of Epidemiology, Biostatistics, and Occupational Health, McGill University.
Pierre-Luc Lupien, MA in Sociology. Université du Québec à Montréal.
Nancy Mayo, Ph.D. Professor at the School of Physical & Occupational Therapy, McGill University.
Christopher McAll, Ph.D. Scientific Director at CSSS Jeanne-Mance. Director and Professor at the Department of Sociology, Université de Montréal.
Erica Moodie, Ph.D. Associate Professor, Biostatistics, Biostatistics Program Director, Department of Epidemiology, Biostatistics & Occupational Health, McGill University
Daniel Poremski, Ph.D. Postdoctoral student at St Michael’s Hospital and University of Toronto.
Laurence Roy, Ph.D. Postdoctoral student at the School of Physical & Occupational Therapy, McGill University.
Catherine Vallée, Ph.D. Assistant Professor, Department of rehabilitation, Université Laval.
Robert Whitley, Ph.D. Researcher, Douglas Institute; Professeur adjoint, Department of psychiatry, McGill University.


Economic analyses

Eric Latimer
A cost analysis was performed to evaluate the impact of the HF programs, considering all costs incurred by participants. On average, the intervention cost $22,482 per person per year for high need participants and $14,029 for moderate need participants. This cost is almost completely offset by savings in costs of other services, notably hospitalizations, shelters, and other types of housing. Over the two-year period following participants’ entry into the study, every $10 invested in HF services resulted in an average savings of $8.27 for high need participants and $7.19 for moderate need participants. Please consult our methods document (pdf link) to see the unit costs (costs per service, per bed-night, etc.) and their sources.

Ongoing analyses being carried out in Montreal aim to: (1) estimate costs per person of being homeless, in each of the five sites, and the personal characteristics associated with greater costs; (2) evaluate the overall cost-effectiveness of Housing First for people with high and with moderate needs, and the factors associated with greater or lesser cost-effectiveness; and (3) identify the individual-level and site-level factors associated with greater or lesser offsets in costs of resources such as shelters, psychiatric hospitalizations and court appearances.

Planning, development and implementation of the project in Montréal

The role of advocacy coalitions in a project implementation process: The example of the planning phase of the At Home/Chez Soi project dealing with homelessness in Montreal. Evaluation and Program Planning. 2014 Aug;45:42-9. doi: 10.1016/j.evalprogplan.2014.03.008.

Marie-Josée Fleury
Guy Grenier
Catherine Vallée
Roch Hurtubise
Paul-André Lévesque
This study analyzed the planning process (summer 2008 to fall 2009) of a Montreal project that offers housing and community follow-up to homeless people with mental disorders, with or without substance abuse disorders. With the help of the Advocacy Coalition Framework (ACF), advocacy groups that were able to navigate a complex intervention implementation process were identified. In all, 25 people involved in the Montreal At Home/Chez Soi project were surveyed through interviews (n = 18) and a discussion group (n = 7). Participant observations and documentation (minutes and correspondence) were also used for the analysis. The start-up phase of the At Home/Chez may be broken down into three separate periods qualified respectively as “honeymoon;” “clash of cultures;” and “acceptance & commitment”. In each of the planning phases of the At Home/Chez Soi project in Montreal, at least two advocacy coalitions were in confrontation about their specific belief systems concerning solutions to address the recurring homelessness social problem, while a third, more moderate one contributed in rallying most key actors under specified secondary aspects. The study confirms the importance of policy brokers in achieving compromises acceptable to all advocacy coalitions.

Mapping of existing services for homeless people in Montreal

Profiles of agencies serving the homeless or those at risk of homelessness in the region of Montreal. Santé mentale au Québec. Vol 38, numéro 1, spring 2013, p. 119-141, doi : 10.7202/1019189ar

Guy Grenier
Marie-Josée Fleury
Armelle Imboua
André Ngui
This article presents a study of organizations serving people who are homeless or at risk of becoming homeless (PHRH) in Montreal, as well as the determinants of their inter-organizational relationships. The study shows that greater inter-organizational collaboration is needed, particularly within the network of health and social services (NHSS), to deal with the concomitant problems faced by PHRH. Among determinants that have an impact on the extent of inter-organizational relationships are the number of services offered, the appreciation of the relationships between organizations within the NHSS, and the ratio of Anglophones among the homeless and of individuals with gambling problems.

Narrative interviews

Christopher McAll
The impact of the Montreal At-Home project after 18 months, from the participants' point of view.
McAll, C., Lupien P-L., Gutiérrez M., Fleury A., Robert A., & Rode A. (2013). CREMIS.

A set of narrative interviews was conducted on 45 randomly selected for a qualitative interview, 18 months after study entry, in order to understand possible differences between the two groups over the course of the study. In contrast with the quantitative measures, the qualitative data suggest more substantial differences between the HF and TAU groups.

Wellbeing, defined as a sense of “peace,” “security,” and the ability to live at their own “rhythm,” is a feeling that participants in the HF groups were more likely to mention in the 18-month interviews. Experimental groups were also more likely to say in interviews that they felt less stressed (15 % of HF vs none of the TAU participants sampled). Participants from the HF group were three times as likely as those in TAU to mention in interviews that their mental health had improved over the 18 months, four times as likely to say that they had reduced their consumption of prescription medicines, five times as likely to say that they had made new friendships, and six times more likely to say they had reconnected with family members.

Four factors stand out among those to which participants attribute the success of the HF intervention: rapid access to stable housing, intensive long-term support by ACT and ICM teams, a non-judgmental attitude on the part of case managers, and regaining responsibility for decision making.


Eric Latimer
Robert Whitley
Daniel Poremski
The objective of the Individual Placement and Support (IPS) sub-study was to determine the best way to help homeless people with mental health problems fulfill their career goals. The team compared the impact of the IPS approach with usual vocational rehabilitation services. The IPS program aimed to help people with mental health disorders find employment on the regular job market. While employment rates were greater in the IPS group (34% vs. 22% in the comparison group) the percentage of people who found work was lower than what could be expected from an IPS program. The short period of observation, as well as experiences of homelessness, may have contributed to this finding.


Henri Dorvil
Sarah Boucher Guèvremont
This sub-study is looking at how participants adapt to their new housing and at the rapport they develop with their landlords/superintendents and caretakers, as well as the perception that the latter group has of their At Home/Chez Soi tenants.


Jean-Pierre Bonin
This paper explores the experience of families supporting a homeless relative with mental illness. The authors investigate the types of support offered by families; they also examine the relational and personal dimensions which can support a redefining of the role and consequently a modulation of the level of support. Using qualitative research, semi-structured interviews were realized with 14 family members. The results indicate that housing and the financial support are the types of support most frequently stopped while the emotional and the social support tend to be pursued. The control in the relation, the emotionalism and the perception of future appear as important element for the initiation of a role redefining process.

Victimization and Criminal Justice System Involvement

Anne Crocker
Eric Latimer
Laurence Roy
This sub-study will examine the extent of victimization and Criminal Justice System (CJS) involvement among At Home project participants and compare the profiles of their CJS/victimization experiences. The study will evaluate whether the Housing First intervention is more effective at reducing the frequency of such experiences than treatment as usual. An economic analysis component will compare the costs associated with different CJS profiles and the factors behind them.

Legal trajectories, victimization and impulsivity

Anne Crocker
Laurence Roy
This sub-study evaluates the extent to which the approaches of the At Home project influence participants' experiences with the legal system and degree of victimization. Participants' impulsivity is measured to determine whether this aspect affects individuals' behaviour in housing or jeopardizes their success in a Housing First-type program.

Peer participation

Christopher McAll
Baptise Godrie
This sub-study is assessing the integration and influence of peer support workers on decisions and perceptions of different clinical teams of the At Home/Chez Soi project.

Theatre forums

Christopher McAll
Pierre-Luc Lupien
With the collaboration of the Mise au jeu theatre company, theatre forums are a way to present research results to a mixed audience of users, case managers, researchers, and administrators to communicate, validate and discuss the project findings.

This type of activity promotes the participation of marginalized individuals—in this case homeless people who use mental health services — with dramatizations of key events from their experiences.


For all questions related to the At Home / Chez Soi project, please contact: 
Eric Latimer, Ph. D.
Lead investigator, Montreal At Home project

McGill Logo

CIUSSS ouest montreal logo


  • Douglas Hospital
  • Dobell Pavillion
  • Brain imaging centre