The RAISE Early Treatment Study – Dr. Nina Schooler


Nina Schooler:
I was not the principal investigator in the study. That was John Kane. And I want to extend his regrets at not being
able to be here because he’s been very active in presenting information about the study
and, of course, in the conduct of it. And he would, I know, have greatly enjoyed
the opportunity to interact with you all. So I’m the stand-in. With that, the NIMH program was called and
is called RAISE, Recovery After an Initial Schizophrenia Episode. And our part of it, there were two sections
of the RAISE program, the NIMH program. One was called RAISE Connection, and ours
was called RAISE Early Treatment Program, RAISE ETP, and that’s what I’m going to be
talking about this morning. And I should tell you is that this was planned
as a very data-dense presentation, lots and lots of stuff. And as I’ve been listening this morning and
listening to the comments that you all made, you’re going to see me advancing rather quickly
through some of the slides, which I think are not precisely what you may want to hear
in detail. I know that everybody has been given a copy
of the slide set earlier. And here is our group. That’s John at the top. That’s me. And as you can see, our large group — this
is not working for me — our large group that was involved in the program, and this is not
even the full extent of the number of people. And I might stop and comment that one of the
things that has happened in the world that we live in now is that so much of what we
do is actually team science. There are very, very few things — we’re not
Wilbur and Orville Wright anymore. And so, one has to think about teams, and
it can be a very exciting prospect to be part of a team that’s integrated. And these are our NIMH collaborators. And I especially want to acknowledge Bob Heinssen,
to whom enormous credit must be given for moving this project forward and for sustaining
it. And of course, we have 34 sites. These are our clinicians. These are our sites. And of course, we had many participants in
the program. So what we were mandated to do was to think
about the first episode of psychosis and how we could treat that in order to maximize recovery
— our goals are not maintenance — to maximize recovery and, hopefully, change the trajectory
of illness over time. And so, what I’m going to do now is, first,
I’m going to describe the program that we developed to do this. And then the second mandate that the NIMH
gave us was they wanted us to evaluate the program, not just describe it, but to evaluate
it in a rigorous, randomized clinical trial. So, first, I’ll tell you what the program
was about because I think that’s especially important to give you an idea of how we were
thinking first episode psychosis, defined fairly broadly, ought to be treated. And then what we’ll do is talk about how we
evaluated it, and then finally what we found, which, as you’ve already heard, is now in
press in the American Journal — American Journal of Psychiatry. So, as I say, the overall goal is recovery. Not only are we doing team science, but we’re
talking about a team-based intervention. This is not looking for a magic bullet that
there’s one thing we can do that’s going to fix this problem. We’re seeing it as a team-based approach. And the model you see here, a group of people
who represent the team at the clinical site that provides the integrated treatment. And the fact of this, an important part of
this is we’re not trying to design something in a test tube, again a mandate from NIMH,
but to see services that can be — generally more or less work within the current mental
health reimbursement system. So that the treatments that were provided
were, with very small exceptions that I’ll talk about, had to be paid for by the ongoing
system. At the same time, what the RAISE initiative
paid for was the research to develop the program and then to evaluate the program. And what we had was a general model for how
treatment ought to be delivered, which is actually quite a modern model if you think
about treatment when I first came into this field, which is a doctor proposes, patient
accepts. So, first of all, the term we used in our
study and our program was “client,” and our overall model was one of shared decision-making,
which involved the team and the person and, very importantly, the families of people who
were part of the program. We developed manuals, rather elaborate manuals,
if I do say so, and provided training and ongoing consultation. All team members, and as I’ve said, there
was a project director, a psychiatrist or nurse practitioner prescriber, an individual
resiliency training person who provided the psychotherapeutic portion, and a family psychoeducation
person. So we had a four-person team, and all of them
were expected to have this set of skills. And what the skills are, they all reflect
the notion of team, team building, and integration. This meant that in order to do this, there
had to be team meetings, and I should mention — so that was a requirement that there were
weekly team meetings that were held that allowed all the members of the team to interact and
to communicate about the person who was being cared for. And that’s particularly important because
I think, as we all know, often treatments can be delivered in silos so that nobody — one
group doesn’t know what the other part is doing, and that’s not the best way for treatment
to be delivered. The components were, as I said, a psychopharm
— psychopharmacologic treatment delivered by a prescriber, family treatment, individual
resiliency training, which was our psychotherapy, and supported employment and education. I forgot to mention the SEE specialist, and
this was a particularly important component. If you think about recovery, people want to
get back to their real lives. And particularly at the time of a first episode
of psychosis, many people don’t think of this as an illness, and in fact, it’s not an illness,
it’s an experience that people are having. And these experiences are not things they
necessarily are seeking treatment for. What they’re seeking help with is getting
back on track and back into life. The family program was actually designed to
be relatively brief, was designed as a 12-session program, designed to begin soon after the
person entered treatment and to include the client, relatives, and other important support
people that might be in their lives. And it was coordinated with the individual
resiliency training that I’ll talk about in a few moments. The goals were to try to assess the situation
that the person and the family found themselves in, provide education about the illness, and
provide some skills in terms of how to deal with a person who’s living with a first episode
psychosis. And to provide families the skills to support
and protect themselves at the same time. It’s designed to be relatively brief, 12 to
15 sessions, and then to provide what we called monthly check-ins so that there would be ongoing
support throughout the course of the treatment program. Treatment program was designed — is not designed
with a fixed period. It’s not 15 weeks. It’s not 20 weeks. It’s not 12 months. It’s not a year and a half. The minimum exposure that we looked for was
2 years, but the expectation was that the intensity of treatment would fade as people
reached milestones, made recoveries, and moved on to their lives. Our main goal here was not to engage people
in treatment so that they became clients in perpetuity, but to foster recovery, as I’ve
said before. Individual resiliency training was, again,
goal oriented and specifically designed to build on strengths that individuals had. And as you’ve already heard from Andrew Solomon,
he used the term “resiliency” in his talk, and I’m using it here again, an important
feature. And I saw someone in the hallway earlier from
NIAAA, the National Institute on Alcohol, and he said they were holding a workshop on
resiliency today as well. So this is an important concept that we all
use. We had a series of modules, which were expected
that everyone would experience, which included things like goal setting, developing resiliency
especially, and then what we called individualized modules. And these were things that people could experience
if they needed the particular issues. So there was an issue — if people had substance
use issues, we could focus on those. And these were not designed in a sequence
that you had to do module one before you did module two, before you did module three. They were available to the clinician in order
to be able to provide them to people in the best way possible. Supported employment and education was one
of the key features of the program, and this is a very, very valuable treatment, well,
well, documented in the literature. And one of the things that’s important about
supported employment and education is that it’s not designed on a vocational rehabilitation
preparation for work. In other words, before we can let you go out
into the community, take a job, you have to prove that you can come in on time. You have to be properly dressed. You have to do a wide range of pre-vocational
activities. The goal of supported employment/education
is to provide supports that enable people to move into competitive employment or competitive
education. In other words, back to school if they’d been
students, back to work if they’d been working. Of if they’d not been working, putting them
into a work situation. It’s a very, very important distinction, and
it really resonates extraordinarily well with the young people who came into our program
because these were the things that they saw as their goals. The goals that we heard most often is “I want
to move out of the house.” “I want to get a job.” “I want to get a car,” and “I want to get
a girlfriend.” Those were the things that we heard most frequently
from people when they entered. Pharmacologic treatment. And as our group has said very much, first
episode treatment for psychosis with medication, using the same medications that are used in
schizophrenia and other serious mental illnesses more broadly, same antipsychotic medications. Treatment is the same, but different. And these are some of the principles that
are key here. One, two most important ones — that doses
need to be much lower, and the second one is that side effects, even with much lower
doses, are much more common. And so, therefore, the ability to deal with
these two issues is very important. What we were doing here because, as I said,
the goal of the NIMH RAISE program was to provide the treatment in community settings
that already existed. We were not setting up specialized first episode
programs. And so, therefore, the prescribers in these
settings had extensive experience in using the medications that we were going to be using,
but relatively limited experience with patients with first episode psychosis. And therefore, we developed a manual and also
a support system that would enable us to convey the information to clinicians, and this was
a tool we called COMPASS. It seemed that NAVIGATE and COMPASS went together,
and COMPASS is an online decision support system that facilitates the decision-making
that’s required here. It’s a Web-based application, and it was available
through the program. It could be done on a laptop, on a desktop. And then in the middle of the program, we
were able to provide our prescribers with iPads that they could — that they could use
as well. And I’m not going to go through the details
of this, but the program includes — the decision support system includes two very important
components. The first is a patient self report, and the
purpose of this is to provide the individual with an opportunity to say how they’re doing
and how they’re feeling, and that then feeds in seamlessly to the prescriber, who then
conducts an interview, which is based on what the person has said. And I’ll just give you one example. If a person says, “I’ve been feeling depressed,”
then the clinician will ask, “You’ve been feeling depressed. Can you tell me a little bit more about that?” If the person says, “I haven’t been feeling
depressed,” the prescriber will say, “You said you haven’t been feeling depressed, but
I wonder have you been feeling sad, out of things, having a hard time getting started?” So the person is always being heard, but all
of the questions are being asked. And then, finally, there is a selection of
medications, which is recorded. What’s really exciting for us is that we completed
3,939 of these assessments, which provide an enormous amount of data about prescriber-client/patient
interactions, and we’ve only begun to look at those. I’ll just give you a couple of examples. So one of the questions that we ask people
as part of the decision support is “Do you want to change your medication?” And of those 3,939, about a quarter of the
time people said yes. And in fact, that changed over time so that
they said less later in the course of the program than they did earlier, which suggests
that what was happening is prescribers were starting to get it right or getting it “righter.” Not a real word. So let me now go on to say this is the program,
and the question is we developed it. We thought we knew how to deliver it, and
the question now is does it work? All right? Is this going to be better for people than
what they’re normally getting? And so, what we did was to conduct a randomized
controlled trial to compare NAVIGATE to community care. By community care, what we meant was the care
that was offered in clinics in the United States today. At the community care sites, we did no interventions,
all right? And what we did was we did what’s called cluster
or site randomization. What that means is rather than having a study
at a number of sites and in each study individuals are randomly assigned to receive one treatment
or another, what we decided to do instead was to randomize the sites. That has a number of very important advantages
that I’ll talk about in a minute. What we did was we set this up for a minimum
of 2 years of assessment, but people who entered early had a longer period of time in the trial. And we did both onsite and remote assessment
in order to protect the masked or blinded nature of the study. And so, here we are in 21 States with 34 sites,
the groups that I identified earlier when I thanked them. And here’s what we did. As I said, random assignment of sites, 34
sites — 17 received NAVIGATE, 17 received community care — and those are the numbers
of patients/clients who entered the program. Ultimately, we had 404 people in the program. And here is what they had to be. These are the criteria. It had to be a first episode of psychosis. You’ll notice there’s a broad range of diagnostic
criteria. We were bound by the diagnostic criteria,
but the broad range gives us an opportunity to look at a — not just look at schizophrenia,
per se. And in order to be sure that we were early
in the course of treatment, no more than 6 months of antipsychotic medication. Now the problem when you’ve done this kind
of randomization and people are at a site and the site knows what it’s doing, and indeed,
that was exactly what we wanted, we needed remote masked assessors. And the people that we were working with actually
not only were masked, blinded to which site had which treatment, they didn’t even know
very much about the — they didn’t know the nature of the study. All they knew was they were being asked to
diagnose people with any of the psychosis disorders, and they were being asked to use
assessments that are commonly used in schizophrenia. So they knew that much. But beyond that, they didn’t know what our
goals were, what our outcomes of interest were, and so forth. So the summary of this is we think it’s a
relatively novel clinical trial model using site or cluster randomization and very, very
well suited to treatment — to multimodal treatments that really can’t be blinded well. And also, the kind of treatment we were delivering
leads to something that can be called “bleed.” In other words, if you’re a clinician at a
site and you’ve learned how to provide a very interesting treatment to somebody that you’re
seeing, it’s really hard to say when the next person comes in through your door, “Oh, no. No, no, no. This guy doesn’t get that. That’s only for that person.” And this avoids that problem, and it also
avoids another really critical problem in randomized trials and very important in first
episode psychosis, and that is people are often very reluctant to agree to treatment
by toss of a coin. In other words, we have two treatments, and
you’re either going to get one or the other. They very, very often walk away from it. In this study, they were told there are two
kinds of sites. At your site, this is what you’re doing — we’re
doing, and we’d like you to join. At the other sites, they were told, this is
what we’re doing, and we’d like you to join. So no patient had to agree to be randomized. The only people who had to agree to be randomized
were the 34 sites, all of whom had said they would be willing to do either community care
or NAVIGATE if they were assigned to it. And I have to tell you, we held our breath
when we revealed the condition to which sites had been randomized because we were afraid
that some sites would then say, “Oh, NAVIGATE? That’s too much work. I don’t want to do that. I was hoping to get the other.” And some would say, “Community care? That means I’m not going to learn how to do
this terrific new treatment.” But as it turned out, this is very important
to me, no site refused to participate once they had been given their random assignment. So, so what did we learn? All right, that was the first. I’ve told you what we did, and now I’m going
to tell you what we learned. First, who were these people? They were, indeed, young. The mean age was 23. And in many clinical trials today that are
being done for new medications or for psychosocial treatments, we’re seeing patients who are
40 years old who’ve had their first illness when they were 23, but who’ve now been ill
for 15 years. They were more white, but indeed, there were
more African Americans in the study population than we would expect in the U.S. population. And that has to do with the fact of our going
to public community mental health centers. In terms of working and school, what we had
was that some proportion of patients, relatively low proportion, were working at the time they
came into the study. And the vast majority had been hospitalized
before coming to these community mental health centers. In terms of diagnosis, although we allow this
full range of diagnoses, by the time people came to these community mental health centers,
they were, in fact, more than 50 percent could meet a diagnosis of schizophrenia, which means
that the illness had been ongoing for more than 6 months. So you will remember that I told you we had
these four components to our treatment, and we had sites that did it and sites that didn’t. So our first question was what did people
think they got? So we asked questions every month about a
variety of things. But among the questions we asked were, “Have
you had individual sessions with a mental health provider who helps you work on your
goals and look positively toward the future?” That’s our code for individual — that’s translating
what we thought of as individual resiliency training. And in all of the slides you’re going to see,
red is NAVIGATE, and blue is community care. And what we see here is looking over the 24
months, clearly there is always a significant advantage for the NAVIGATE group in that they
believe they’re getting more of this than do the people who are receiving community
care. But this also shows the thing I talked about
earlier, which is these treatments are not etched in stone, and they decline over time
as patients go through, as clients go through the 2-year period. So that’s individual resiliency training. And this is the family one. And here what you see is, again, a significant
difference between community care and NAVIGATE. But one of the most important things here
is that while in the previous slide, virtually every client experienced the individual resiliency
training, only a portion of clients, about 50 percent, experienced the family intervention. And the reason for that, I think, is that
even this early in the illness, many clients have already become distanced from their families,
or their families have issues that make it difficult for them to engage with the client
and with the team in treatment. And then the third was asking about “met with
somebody who is helping you get a job in the community.” And here — get a job or go back to school
— we see the same kind of story. And then, finally, the piece about the pharmacologic
intervention was, “Were you asked to record your symptoms before you came in to see your
prescriber?” And you see here a very, very high percentage
of people who experienced that through the 2 years. And I should mention that our model for pharmacologic
treatment was we did not expect that to fade over time. So in the last 10 minutes that I have, I want
to tell you about what we found. So, so far I’ve told you about what our treatment
was like, how we designed our study, evidence that indeed we did manipulate the treatment
difference between the two sets of sites, and now the question is what did we learn? What are our answers? So the first point is that people who are
receiving NAVIGATE stayed in treatment longer than people who were in community care. And we know, one thing we all know from anything
that we’ve seen in the past is that if you’re not there, you can’t get the help. And people walk away. That’s a bad thing. So we did a better job of keeping people in
treatment in the NAVIGATE sites than we did in the community care sites. The second thing is quality of life, the quality
of life scale, which is a scale that addresses things like self perception, getting on with
friends and family, work and school, and the ability to participate in normal community
activities. It’s a composite scale. This was completed every 6 months, as you
can see here. And what we see is that the NAVIGATE group
started a little bit lower than the community care group, and over time, by the end of the
treatment period at 2 years, the clients who received NAVIGATE were doing significantly
better than those who received community care. This was our primary outcome, and we were
— to say the least, we were thrilled to see this finding. The second issue has to do with work and school,
and this is a slightly more complicated picture. But again, if we look at the interaction,
by the end of the 2 years, the NAVIGATE group is actually doing just a little bit better
than the community care group. This is not as dramatic a finding as the one
we had with quality of life. And when we look at the symptoms score, this
is an instrument called the PANSS, which is very widely used, what we see is that at 6
months, there is a significant advantage for the NAVIGATE group. What this is telling us is that symptoms improve
more rapidly, which is a very important kind of baseline to set in order to see where you’re
going to go with future quality of life development. But over time, the community care group does
about as well. In terms of depression, we see advantages
for the NAVIGATE group. I should say that in this case, the Calgary
depression rating score is a lower score is better. So the lower level is better. And that’s where we see our advantage for
NAVIGATE. Now we come to hospitalization. And we look at this, and I don’t have to tell
anybody in this room, regardless of your statistical expertise, that those two lines are not different. There is no difference in hospitalization. And we — we pondered over this very long
and very hard, and I want to show you the next slide, which is designed to put this
into context. So here is our study over 24 months, and these
are our treatments. This is our intervention, and this is the
treatment as usual. And these are all multimodal studies that
were done in the field. In other words, studies that looked at a range
of outcomes — at a range of treatments. So they were not just providing a pharmacologic
treatment or a single psychosocial treatment. And what we see is 34, 37 over 2 years. STEP, which is a recent study conducted in
New Haven, you see is a 12-month study. They’ve got 23 percent in their first year
and 44 percent in the community group, in the group that didn’t receive the treatment. OPUS does it year by year. This is a Danish study, and they’ve got 59
percent in the first year, 71 percent in the second. In the second year, it’s down about to where
we are, but that doesn’t include what happened in the first year. LEO is again comparable to us in 15 months,
but with about 50 percent in their comparison, and RAISE Connection, which is the other RAISE
program, looking at their data for their first 12 months, what you see is 32 percent, comparable
to ours, and there wasn’t a comparison group. That was an implementation demonstration project. So we had to think about this, and what we’ve
come to the conclusion is that our community care sites are not treatment as usual. Our community care sites were 17 sites that
had said at the outset that if randomly assigned to deliver NAVIGATE, that they could and would
deliver that. So these were a group of sites that we’d selected
very carefully and that also believed in the idea of an early intervention. We don’t have details on what they did, but
we believe that they did better than would a random selection of community sites in the
United States to their and our credit, and we’re very grateful to them. So the other question is nobody is average. As Garrison Keillor has often said, the women
are strong, the men are good looking, and all the children are above average. Indeed, all the children are not above average. But what we were looking for was to examine
important predictors of outcome, and we’ll be doing much more work in this area. But what we did first was to look at the duration
of untreated psychosis. And in this community clinic, these community
clinics — and Tom actually already mentioned the number — the median duration of untreated
psychosis was 74 weeks. That means half below, half above. Seventy-four weeks before people came to any
treatment, and that was not necessarily our treatment. So they’d had 74 weeks of psychosis untreated
before they came to treatment, median split. And what we did was to see whether looking
at those who were better than the median and those who were worse made a difference. And here is the finding. And what I want you to look at is this is
the quality of life, and this is for those who had less than 74 weeks. And what you see is a huge effect for that
group, and all the other three groups, the two community care groups and the group with
NAVIGATE with a longer duration of untreated psychosis, does not have as good an outcome. So this becomes a really important marker
for us in many ways. We have a similar finding for the psychopathology
score. You will remember I showed you that the symptoms
only showed an effect at 6 months. When you look at it with duration of untreated
psychosis, it goes throughout the entire 24-month period. So last minute for a few conclusions, and
I really, really am looking forward to your comments. First, the recipients of NAVIGATE were more
likely to remain in treatment and received — had greater improvement in our primary
outcome measure, which was quality of life. They were more likely to be working or going
to school, and they had a significant degree of symptom improvement in the first 3 months. Duration of untreated psychosis is an important
moderator/predictor of NAVIGATE success. And this is a very important finding for us
because when we think about things that we might be able to change, perhaps at Google
in the future, think of things that we might be able to change, duration of untreated psychosis
is certainly something that we can strive for. So what we’ve shown is that a coordinated
specialty care program can be delivered in the community, that it can be successfully
— well, it can be successfully delivered. It can be successfully evaluated, and it does
have an effect. Have we changed the trajectory of illness
for people with first episode psychosis? We don’t know. We’ve seen an effect at 2 years. We’re currently engaged in a follow-up, which
will take these clients out to 5 years after initial engagement, and I know that John or
I or someone else on our team will be very happy to come and speak to the Alliance and
tell them what we know at that point. Thank you for your attention. (Applause.)