Hello and welcome. I'm Dr. Amy Cohen, Program Director for SMI Advisor and a clinical psychologist. I am pleased that you're joining us for today's SMI Advisor webinar, Supported Housing for Homeless Adults with Serious Mental Illness. How does it work? SMI Advisor, also known as the Clinical Support System for Serious Mental Illness, is an APA and SAMHSA initiative devoted to helping clinicians implement evidence-based care for those living with serious mental illness. Working with experts from across the SMI clinician community, our interdisciplinary effort has been designed to help you get the answers you need to care for your patients. And now I'd like to introduce you to the faculty for today's webinar, Dr. Robert Rosenheck. Dr. Rosenheck is Professor of Psychiatry and Public Health at Yale Medical School and also Director of the Division of Mental Health Services and Outcomes Research in the Department of Psychiatry. He is an internationally known mental health services researcher and a leader in cost-effectiveness studies of behavioral health interventions and in implementing and monitoring quality of care and other aspects of the performance of large healthcare systems. He's received awards recognizing his work from the American Public Health Association and the American Psychiatric Association. Dr. Rosenheck, thank you so much for leading today's webinar. All right. So supported housing seems simple, getting people houses, but in fact, it's a good deal more complex. And of course, when I started as a psychiatrist 40 or so years ago, housing was not any part of my training or my professional identity. I was involved in the development of VA homeless programs and started working on the problem of housing 30 years ago when it was a peculiarity that a mental health system would be concerned with housing, although housing was obviously an issue for homeless people. It was not something that was part of the medical bag of therapies. It's ironic or perhaps telling that we're back to talking about housing and the whole world, as all of you know, has changed in the past six months and we're facing a housing crisis, not just of homeless people, not just of seriously mentally ill people, but of low income people and middle-class people around the country. And so the issue of whether housing is a right, which nobody would have taken seriously in past years, is suddenly not yet on the agenda, but it clearly is what we are confronting. Now a simple view of supported housing is that you have a house, you have a caseworker, and you have a bunch of other supports and you link them together. In fact, it's vastly more complicated than that. And the clinician is one part of a much larger system. And rather than start with a clinical perspective of what we do, I think it's worth starting with a housing perspective. And in the beginning of this century, it was recognized that housing in cities was a major problem, that housing was for many low-income people inhumane. And there was a great deal of effort to develop standards and improve housing. Part of the problem of that was that much housing was improved, putting it out of reach of low-income people. And by the 1980s, there was a major shortage of low-income housing. The central dogma of this whole area of service is that the following ingredients lead to housing. Income, if people have adequate income, whether from disability or public support or employment, they can get housed. If rents are low, if there are decent housing available at a low rent, low-income or disabled people will be able to get housing. If there are rent subsidies, if the government provides subsidies for low-income housing, that will help. And the fourth ingredient is case management, especially for people with serious mental illness, negotiating the complex bureaucracies to get income, to get housing is very complex. And often such people are not taken seriously. And having a case manager is sort of like a legitimized guide through the bureaucracy. Now all of these ingredients are in short supply and have been in declining supply since the 1980s. Public support income has declined. Low-income rents have disappeared as cities have gentrified. Rent subsidies, I'll discuss, are limited. And case management increasingly is provided by the mental health system, but in limited supply. So let's go back and do a little history. At the end of World War II, there was a limited problem of homelessness, largely confined to Skid Row or the Bowery areas of New York, largely white alcoholic men. And in the 1980s, there emerged what was called the new homeless, who were largely minority, had alcohol and drug problems, and were much younger, and was a new problem of which posed a new problem of combined serious mental illness and substance abuse. But this was the time when the Reagan reaction against the war on poverty was dominating the cultural landscape. And frankly, for those of you who remember that time, poverty was not something that could be spoken of. It could not be thought of as a government or a social goal. It was despised. But homelessness was undeniable because wealthy people going to and from work were tripping over homeless people in Grand Central Station and other major transportation hubs around the country. But it could not be addressed as a housing problem or a poverty problem because of the culture. And so it came to be addressed as a medical problem, and particularly as a psychiatric problem. Now, in fact, only 25% of homeless people have mental illness, and even fewer have serious mental illness. But homelessness became identified as a medical problem because that was what was possible where you could legitimately talk about it in public. And I got involved as a psychiatrist in the VA in developing programs for homeless people. Well, the world has changed a good deal now. And it came to be recognized largely through the work on homelessness that housing was a major problem of people with serious mental illness, many of whom were homeless. And the field of social determinants of mental health has now emerged and is coming into its own even before the pandemic began. And I just would mention three perspectives on this because each of you will play a different role. One is social risk-informed care. Increasingly, clinicians doing standard psychotherapy or pharmacotherapy need to know something about the socioeconomic status of their patients to understand what they're facing. A second level is social risk-targeted care, in which clinicians directly help patients to reduce social or economic adversity, either through direct counseling, direct advice addressing these issues, or potentially through developing a program. But the third and the most complex element here, which some of you are in a position to be engaged in and some may not be, is social program development. And this involves engaging with political and or bureaucratic systems to develop programs and resources to provide supported housing, to provide housing and clinical services together. And this is my little map just so when you feel if you're working in this area and anyone who's working in mental health is working in this area broadly, this is the system in which you are an aunt. And I'll just briefly say we have at the huge level of society calling the main tune is policy. And we have low-income housing policy. To what extent do we provide subsidies for poor people? To what extent do we build low-income housing? All of that comes from the government, even as it uses private sector developers to build the bricks and mortar. There's income support and disability policy, the transfer of funds from the wealthy to the less wealthy, which has declined dramatically in recent decades. And then there's our area, which is mental health and social service policy. And when I was in training, we thought of psychiatry as a separate discipline. And in the 70s, for those of you who remember, psychiatry wanted to get in on national health insurance, which people thought was coming, and medicalized itself and more or less abandoned its earlier interest in social situation of its patients. Not entirely, but there was a sea chain. And so all these area of policies get filtered through bureaucracies. And bureaucracies are often condemned and damned for political reasons. It's always a good move to criticize the bureaucracy. But in fact, bureaucracies are the way policy benefits are delivered to citizens. And one cannot avoid negotiating, dealing with bureaucracies, learning how to use bureaucracies, which are well-meaning, well-intended, but invariably complex, bewildering, and seem off-putting. At the next level is the local level of bureaucracies, which is the local public housing authority, social security and VA disability offices and city welfare, and the mental health agencies. And even these, which are more local and which are more knowable, can be formidable. And at the lowest level, at the bottom of this huge chain of being, are rent subsidies, cash payments, entitlements, and working relationships with case managements or service brokers. Now, I have over here on the side supported employment, which is part of the mental health system, but has a unique role because it bears the hope that people could return to work and earn their own benefits and pay for their own housing. So this is the context. For those of us who have engaged in this, this is the immensely complex situation in which we are trying to work. So let me talk a little bit about this context. It can be difficult, and there are good reasons why it is difficult. And just to say a little bit, we have and many of us benefit from the mortgage interest deduction, which in fact is a welfare payment. It's a tax payment. You don't have to pay taxes. You get a deduction if you have a home mortgage. But that is in fact the government paying you. It's called the tax payment. And it amounts to about $66 billion a year, 84% of which goes to households with incomes over $100,000. Then there are housing subsidies, which are tenant-based rental assistance, public housing, project-based assistance, the kinds of housing we're trying to get for our clients, whether homeless or just with serious mental illness. And the total outlay for that is $39 billion. Now that should aggravate people, that we spend $66 billion in handouts to people with incomes, mostly with incomes over $100,000, and just over half that for people who are homeless and have low income. So why is there this huge difference? Well, it's the difference between an entitlement, of which there's no limit, no matter what it is, the government will pay it, versus a budgeted item. And the ironic thing is we have a totally, our government has a totally open checkbook when it comes to the mortgage interest deduction. Nobody ever gets denied their mortgage interest deduction because we've run out of money. All the funds are used up. That goes on forever, whereas Section 8 housing vouchers are limited. And as many of you know, there are very long waiting lists because the supply is limited and it's become more limited in recent years for somewhat complicated reasons. So let me just give you a little data on who we're trying to help, a dimension of their lives that I was never taught about or directed to think about when I was in training. And these are some data from our HUD-BASH program. A lot of my experience has been with the VA's HUD-BASH program, which now has 80,000 homeless veterans, formerly homeless veterans, housed, or at least it did before the pandemic. It's not clear what the status is now. But this is data which I'm sure is quite familiar to you. The total income of the average client per month is $433, half of which is disability or welfare. Some of it is earnings, but these are very poor people. And here's some data from two of our most prominent studies of schizophrenia, the Cady study in which the total monthly income was $1,000 a month and $700 of which was disability. And the RAISE program, which was the largest study of first episode treatment of schizophrenia, and those clients had incomes of $700 a month. And even those people very early in their experience living with schizophrenia, 42% of their income was disability. And I highlight these simple facts from some of the most prominent programs and studies in our field, because there's very little attention to this fact from us as professionals, while this is in fact maybe the most serious problem faced by the people we're trying to help. And just to nail this one notch further, the average payment for SSI, which is what many of our clients are on, is $231 a month. The minimum wage, if they were working at minimum wage, would be $377. Their income at poverty would be $644. The average renter income is $900 a month. And the average one bedroom fair market rent in 2019 was $970. For a family, for a two bedroom, which might support a family, it's $1,100, which is to say that systematically in this country in which we live, the people we're trying to help, even if they're working, have less money, cannot afford their housing. And that is the situation that we and you face when you try to address this issue. So what's in the name? So let's turn now to supportive or housing first. These are all names that are used. To me, they're fairly interchangeable. Housing first, which was championed by Sam Simberis, particularly emphasizes getting people right into housing. There was a tradition of saying, you have to be on your medication. You have to be going to the clinic. You have to have your symptoms stabilized before we'll get in your housing. And what Simberis did was he said, let's get people, no matter what shape they're in, let's get them into housing. That's a decent, humane right that they should have. And he showed success in getting people into housing. He also emphasized, and we'll talk about this a little later, the priority of choice. That when you have people with serious mental illness and you find them housing and put them in housing, they should be treated like any other tenant. They should not be treated like patients in a hospital, but rather offered case management, offered treatment. But essentially, he insisted that they be treated as free adults deciding what they wanted to do. Now, there are complexities with that, which we can talk about, but that was a useful principle to have out there as an ideal. So what is it? What do you do? What are the benefits and what are the costs? Well, the easiest way for me to talk about this is to describe a randomized clinical trial that we did in the VA when we started the HUD VA supported housing program. And we randomly assigned homeless veterans with serious mental illness or serious addiction problems to either get case management and a voucher. That's the key ingredient is those two elements, or just case management alone, or standard HCHV is healthcare for homeless veterans. And this is our data on what happened in the first three months. And you can see they were much more like the first column is the HUD-VASH group. The second column is the case management alone. And you can see that vastly more veterans got a voucher, got access to funds for housing in the first three months, but it was only 55%. One would have fantasized that if you give people, if the government makes money available for housing, everybody would use it. Well, it's more complicated than that. And both because veterans and patients are reluctant for some aspects of this, and clinicians are reluctant for some aspects of this, and the bureaucracy intrinsically brings friction to the situation. The clinicians helped the veteran locate an apartment. That was more often when they had a voucher, but it was also true of case managers who were focused on working with the veterans. How many apartments did they visit? They visited on average two apartments and 71% of them met with a landlord, a crucial interaction because many landlords are reluctant to take a person with serious mental illness, but are much more comfortable if there is a case manager there who says, I will be available to you if there are problems. If you have apprehensions, you can give me a call at any time. This of course opens the whole complex area of transparency. And to what extent are clients forced to disclose that they have a mental illness? Well, in the case of the VA, you don't have to do that. You can say this is a veteran and we're providing a service. So there's maybe more flexibility there, but that's an issue that always needs to be addressed. A major challenge is furnishing the apartment, which at least a third of the veterans had their case manager helping them. Sometimes they had to borrow a truck to move furniture around. And if you see the proportion who terminated, far few people terminated. Now another measure was the therapeutic alliance. We tend to emphasize the case management relationship and how important it is. Well, these data show that the therapeutic alliance is strongest for these people when you're offering them a real benefit in the form of a subsidized housing situation. And I know we often think, and it's not something I would want to criticize, but the case managers often feel as if the critical thing is the relationship. But as I was saying before, we shouldn't minimize the issues of poverty and the importance of having a way of paying your rent. So in terms of the duration of participation, if you gave them a voucher or a case manager, they stayed involved in the program much longer, they lived in the community much longer, and a greater proportion participated for more than two years. So both case management and the voucher sustained engagement. And these are the housing results, and the black line is, this is over three years, is the proportion of people who were housed. And what you can see is significantly higher proportions were housed, about 80% after about one year. The downside of it is it took a year to get to that number, and many of the others got housed as well. So the voucher and the program, they made a statistically significant difference. They made a moderately big difference, but it was not 100 to zero. And it's important to understand that there are other ways people have of getting housing in addition to the voucher. This is data from a large program conducted at five cities in Canada, and shows fairly similar results, where the group with the black circles got housed fairly quickly, but the others caught up. So that by the end, in this case, the study was 24 months. In Toronto in particular, there wasn't a difference in the proportion who were housed. And this is from our study, the proportion who were homeless, which declined substantially. And this is data on psychological distress. These programs do not have a major impact on psychiatric symptoms. They do have an impact on substance abuse. And this is a tricky thing to study, because people who are using substances often drop out of the programs. So in a research study, that makes it look like it didn't have any benefit, because the people who were having the biggest problems with substance abuse are not available for assessments. We used a method that addresses that problem, and we're able to show that getting people housed in this VA program did in fact reduce alcohol and drug use. And this is the sort of somewhat grim situation. This is a study of what changed over the first 12 months of an integrated program. And you can see that red line that's shooting up to the top is the impact on housing. But when you look at the impact on community participation, civic activities, religious involvement, social support, and mental health support, none of that changed. And how could that be? We all hoped and thought that homelessness was such a terrible situation. If you got people out of it, much else would be improved. Well, that turns out not to be the case. Some things are improved in some cases, but on average, if you give people housing, what gets better is housing. And these are functioning and quality of life from the Canadian study, showing modest but significant improvements, somewhat different from the slide I just showed, a little more hopeful. But this does not turn people's lives around on average. It gets them housed. So then there's an ongoing debate about substance abuse. And clinicians often are reluctant to give someone a resource, like a precious resource, like a housing subsidy, if they're using. So we decided to test that in some of our programs, and we compared what happened to people who were abstainers, who when they entered the program had no days of alcohol or drug use, and people who were high-frequency users had over 15 days of alcohol or drug use in the previous 30. So these are two extremes. And remarkably, there was absolutely no difference in the number of days housed. So substance use, now, on the one hand, I've talked about the fact that people who use substances often drop out of the programs. In this program, where they had access to housing, they stayed in the program just as much as everyone else. So substance use was not a reason to exclude people if you wanted to get them housed. And if we look at their ongoing substance use, we see that it didn't change very much. It went down. That's the red line. It went down a little bit and then up a little bit. And the abstainers, they used a little bit more. But basically, the difference was maintained. And if we look at psychiatric symptoms, the people who were using substances had more psychiatric symptoms, but it stayed about the same. So this data and a number of other studies we've done demonstrate that there is no need to exclude people from housing because of substance abuse if you provide some kind of supportive ongoing assistance. So one of my focuses is on cost. And so I don't want to overload you with this. These are expensive people. This is their costs over three years, all their costs to society. And you can see that the intervention, that's the bar on the left, does have increased costs. We would love the idea of saying, if we just house the homeless, the costs, there would be huge savings. And there are some studies which purport to show this. But the most rigorous studies do not. And if you look at, particularly on the left, psychiatric and substance abuse, people who get good services use more of all kinds of services. And the three bars on the right show the increased costs of the case management. And this, just to show you intuitively, many people would say, but we know that when people get housed, they have far lower costs. And this shows that when you randomly assign people, so the people are virtually the same, they all show a decline in costs. And why is that? So you can see the same black line, costs peak just before they enter the program and then decline. And this is a phenomenon I've seen in every cost study I've done. It's not that all our treatments are miraculous. It's that people come to treatment when they're at their worst and when they're using a lot of services. And they do, the statistical name for this is regression to the mean, and they do naturally experience a decline in their costs. I'm not going to go into the details of this, but I will again show you the cost data from the Canadian study, which showed a pretty similar phenomenon. You can see the green bars are generally at 0 to 24 months are higher than the blue bars. And this was again, a very rigorously conducted randomized trial. Doing good things sometimes costs money. We love the idea that we can save money and help people as well. But sometimes you have to invest additional funds. And these are some details. Now, this is a study I do want to highlight because while we did a randomized trial that showed the benefit of this supported housing approach, Bill Evans, who's an economist, did a study, a population study in which he looked at the level, the impact of the expansion of this huge VA program, which is the gray line going up to the right. As there are more and more vouchers become available, the black and the dotted lines are the number of homeless people. And the total homeless is the black line, which is coming down. And the dotted line is homeless veterans, which is coming down in proportion to the availability of the vouchers. And this study actually showed at a population level that for each additional voucher that the VA and HUD put into the system, there was actually one fewer homeless veteran. So what about employment? Just very briefly, we did a study of, can we get people housed by helping them get to work? Totally apart from the vouchers. And we did a study of this type where we provided them with individual placement and support therapy. And this shows that they worked more. The program is called Therapeutic Employment and Support. And that's the dark blue diamonds. And they, over 24 months, the homeless veterans that received that service worked more. And to a small extent, they had more days independently housed. It was only 38% as compared to 33%. And so the principle is for a problem like housing needs to be solved with housing. A problem like employment needs to be addressed with employment services, that you don't get miraculous cross-improvements. If you fix the housing, everything else will fall into place. I believe that's a grim fact of how hard this work is. OK. In this last section, I want to talk about the medical dimension and the question of does putting people in supported housing save lives? As I mentioned earlier, when this whole area of intervention started, homelessness was seen as a medical or psychiatric problem. Increasingly, it's been seen as a socioeconomic problem. And recently, the National Academy of Sciences did a big study. And the question they wanted to address was, if we provide homeless people with housing, will we improve their health sufficiently to save enough money to pay for the housing? Which given all the many, many studies, these many studies, all of which have showed that homeless people have worse health, that homelessness leads to worse health, gave people great hope that if they provide housing, they could improve health and therefore justify an expansion of Medicaid to pay for housing. Well, it turned out it wasn't so. And I reviewed this report at an earlier phase, and it was clear they were straining to be able to say that supported housing improved health. But in the end, the committee found that there was no substantial published evidence, including some from us, that demonstrated that permanent supported housing improves health. Remaining house, it was said, should improve the health of these individuals because it alleviates a number of negative conditions. It just turned out, in fact, it didn't. Now it may be, and this is the way I would think of it, is that by the time you've been homeless, and in many of these programs, people have been homeless for a total of, as many as eight years on average, a few months of housing or a few years of housing is not gonna suddenly undo all the damage of a lifetime of homelessness and various other kinds of abuse and neglect. And so it's not that if we provided housing to people, we would have a healthier, better society. It's just that once the damage is done, it can't be easily reversed. And these are some data on mortality, just to hammer the point home. The gray bars are life expectancy for non-homeless people. No, no, the gray bars are the number of years of life lost by homeless people, and the black bars are by non-homeless people. And you can see that the homeless people lose many more years of life than you would expect. And this is a study, I won't go into the details, where we actually compared homeless veterans in the VA to non-homeless veterans who had serious mental illness and the mortality for the homeless. The many lines represent different age groups, but for each age group, especially the elderly, people who are homeless were at greater risk of dying. Now, this is a figure which is not easily understood, but basically it also found in the housing first model of mortality was not reduced. In fact, it was increased for reasons that aren't clear. This lists the reasons that, and again, I won't go into the detail. There was no clear explanation. So what would I conclude from this rapid review of principles and data? Supported housing does indeed move individuals out of homelessness, but it requires a system of redistribution. It's not just a simple clinical intervention of a health professional and a client. And it involves everything from politics and policy to case management and needs to be understood in that broad system. There is little evidence at present of improvement in physical health and mortality. So that's not the justification. It increases costs and we have to ask, is it worth it? Is it a solution to the overall homeless problem? Well, I think what we can say is it's the best solution that we have thus far short of major redesign of how we provide housing in the society as a whole. And I wanted to say, as I started to prepare this, COVID hit and I suddenly felt that I was, everything I knew about this program was irrelevant because face-to-face case management is limited or prohibited. What we're doing now in this program is telehealth, although most homeless people do not have smartphones and we are, at least in our system, we're connecting to them with telephones. Interactions with landlords and public housing authorities are more difficult because you can't have face-to-face interactions. And in all these cases, they all depend on building a trusting relationship, which turns out from talking to case managers is indeed much harder to do so far through the technology we have with this population. The one good thing is that the public housing authorities evidently dropped a lot of their restrictions and requirements for inspections, which made it easier once you had a person to put into supported housing, it was easier to get them in. But whether this will change the effectiveness, cost satisfaction is unknown, but from having talked to a number of the case managers, the work is much harder. I've done this, I've had these discussions with both ACT teams and with supported housing teams and the ACT teams have a much better time of it because they have ongoing relationships with their clients. They've known them for many years and while there's a lot of strain, they have a lot of relationship capital to build on. With the homeless, there's much less because you're largely starting with new people. So the final questions to ask is what do we value and for whom? And in the end, this is a mixed picture. This is a good thing to do, it benefits people, but it doesn't solve major problems of disadvantaged lives. And so the question at the bottom we have to ask is should all Americans be housed? Is it a right? What are the moral choices? Do we choose equality or something approximating equality versus freedom? People can be free to do what they want or are they obligated to others? What would we sacrifice for this goal? And we've learned in the past six months a complicated series of lessons about who is willing to sacrifice for whom and that's a mixed story. Could a war on homelessness, if there was ever to be such a public policy be won and we can ask in the end, why has no one ever asked this question? And let me stop there and welcome your comments and questions. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you.

Dr. Amy Cohen

Program Director

SMI Advisor

Dr. Robert Rosenheck

Supported Housing

Homeless Adults

Serious Mental Illness

Housing Crisis

Social Determinant of Mental Health