The evil man in Eriann Day's life disappeared after she moved into her new apartment.
He had a beard, carried a knife and had followed the 22-year-old Vancouver woman for months. She was homeless, but he would tap on the window of wherever she was staying. He wanted in.
"It is hard for me to understand it as a hallucination. No one else could see him, but for me, he is as real as the toes on my feet," said Ms. Day, a 22-year-old with schizophrenia, a serious psychiatric illness.
She hasn't felt threatened by him since the spring of 2011, when she became a participant in At Home, a large, federally funded research and demonstration project that has put about 950 homeless people with mental illness into apartments in Vancouver, Winnipeg, Toronto, Montreal and Moncton, and given them the support they need to build better lives for themselves.
The study's preliminary findings, to be released Friday by the Mental Health Commission of Canada, show the potential value of an approach known as "housing first."
The idea is that a safe and comfortable place to live is an essential first step to coping with mental illness, and that giving people the support they need in their communities will reduce the burden on shelters, hospitals and the criminal justice system.
The research project looked at how participants like Ms. Day fared compared with others in a control group, who were treated in the more conventional way – in systems that rely on shelters and temporary housing and require homeless people to meet a number of conditions, like getting treatment for addictions, before becoming eligible for help with an apartment.
The housing-first approach does not expect people to prove they are ready for a home of their own. It offers them choices of where they would like to live, usually apartments rented from private-sector landlords, and then helps them deal with their problems.
"The message is having a place of your own and having the support to stay there isn't a magic bullet. But, boy, it is a basic necessity at having a chance to turn your life around," says Paula Goering, a researcher at the University of Toronto and the Centre for Addiction and Mental Health who led the Mental Health Commission study.
The five-year, $110-million project – the largest study on the housing approach that has ever been done – is now wrapping up. The findings could reshape the way governments and social services agencies in Canada and around the world tackle the challenge of helping some of the most vulnerable citizens: Dr. Goering is helping researchers in France launch a similar study.
The Canadian research team found that in the first 12 months of the program, the participants spent an average of 73 per cent of their time in stable housing. That compares with only 30 per cent for the control group.
Many who were housed began to train to do volunteer work, found part-time or full-time jobs or reunited with family members, including their children. They also shifted away from extensive and sometimes inappropriate use of hospitals and had fewer police detentions.
For the 10 per cent of participants who used the most services before the study began, the program saved government $9,390 a year. So for every dollar spent on these participants through the program, $1.54 was saved. The numbers were less dramatic for the group as a whole. But for every dollar spent on each participant, there was a 54-cent saving on shelters and other services.
But these are short-term benefits. It is difficult to know how much the housing-first approach could save in the long-term by keeping people with mental-health challenges off the streets, out of jail and out of hospital. Many of them suffer from other serious health problems as well as from a mental illness.
Dr. Goering said a crucial question at the outset was whether landlords would rent units to participants.
"It is a shift away from government providing all the housing to the market getting involved. It was one of the big questions about it. Would it work? Could you really get landlords who are willing to work with these support teams and give us enough units to house these individuals? That is one of the really remarkable stories.It did work, even in Vancouver, pre-Olympics, we found enough housing for our individuals."
Michael Arruda, one of the landlords, is a Montreal police officer who often deals with the homeless and the mentally ill. He also owns a small apartment building.
His was keen to rent one of his units to a participant in the research project.
"I do a lot of work trying to demystify mental illness," Mr. Arruda said. "I felt this was more than talking about the problem, it was doing something.
His tenant, Johnny Nowra, had a drinking problem.
"The first few months were rough. There were a lot of complaints from the other tenants. There were a lot of parties. One night, he ran around the building with knife," Mr Arruda recalled.
He met with his other tenants and told them they could call him any time if there were problems. He also helped create a social network for the new arrival.
"He stopped drinking a year and a half ago and is off drugs. He is an amazing person. He comes to my family for Christmas and on outings with us.''
The project has also changed Ms. Day's life, and she said she now believes that a more peaceful and productive life is within her grasp.
The young woman had been sleeping on friends' couches or in transition houses, and the stress of not having a home was making her illness worse.
She moved into a one-bedroom apartment in Vancouver in April and her new psychiatrist prescribed a combination of medication that has made her symptoms more manageable. She has not been hospitalized since the move, and hasn't seen the man who haunted her. A support worker checks on her once a week, but she can call if she needs extra help.
She has started training so she can be a peer support worker and help others. She also sings with a semi-professional choir.
"It is hard to take control of anything without stable housing," she said.
Breaking down mental illness
All participants in the study, including the control group, had at least one mental illness.
52 per cent
13 per cent
29 per cent
post-traumatic stress disorder
24 per cent
35 per cent
16 per cent
mood disorder with psychotic features
35 per cent
were dependent on alcohol
46 per cent
dependent on drugs
96 per cent
of participants are currently single, divorced or widowed
4 per cent
are veterans with wartime service for Canada or an ally
56 per cent
did not complete high school
93 per cent
were unemployed at the start of the study
65 per cent
worked steadily in the past
64 per cent
were emotionally abused in childhood
55 per cent
physically abused as children
40 per cent
have a learning problem or disability
51 per cent
have a history of head injuries involving unconsciousness
40 per cent
or more have a serious chronic respiratory illness
35 per cent
were involved with the criminal justice system in the six months prior to the study
7 to 10 per cent
have other serious health problems like HIV/AIDS, epilepsy, diabetes or heart disease