Pat McGorry - Symposium 2003 - BMA Home -

The early intervention paradigm in psychiatry: a gateway to neurobiological understanding and improved outcomes.

Pat McGorry
ORYGEN Research Centre

Thanks very much Chris and first I would like to thank very much Max…..for inviting me to be part of this fantastic day. It has been a really very stimulating, very interesting day. I have thoroughly enjoyed it and I guess I am the token clinician here and so I would like to try and link some of the things you have been hearing about to the treatment of patients, but I also think as Chris was hinting at the approach of focussing on the early stages of a variety of illnesses not just schizophrenia and psychoses has a lot to offer the neuroscience field and I think some things that you have already heard would fit into that today. But to give you a bit of a break, you've been very good hanging around all day until the last session so I am not going to show you a whole lot of science right off, I'll show you a very short video.

The film goes on and you actually see what happens next and what is involved in treatment. I just thought I would show you that because it make it a little bit more sensible what I am going to say next.

So just to try and link in with what happened earlier today. This is something we are all very keen to find the answers for obviously, especially Max and this is what we really think the general public should be demanding actually, in relation to these very serious and potentially serious disorders and at the moment unfortunately they are not getting it despite a lot of rhetoric and policy support, and if you are going to have early intervention you have to have it early, it has to be something that happens quickly.

Now just before I get into a bit of the data I want to spend a little bit of time talking about some things which Assen I think primed us for this morning. The idea that our phenotypes may not be that crash hot and we might be able to turn this to our advantage in a sense by applying a concept that is used in other parts of medicine. The idea of staging or phase specific intervention and that is really what we have been working on the last ten years or so. Trying to develop more phase specific treatments. Originally for schizophrenia and psychoses but now moving more into a broader spectrum of disorders in young people. It is a clinical staging obviously because we don't have the luxury yet of doing the pathological stagings you do have in some of the other medical disorders. Unfortunately particularly with schizophrenia we have been focussing very much on the late stage of the illness, and in fact the whole concept, I think it was mentioned this morning, was defined on the group of patients who had very severe chronic illnesses which didn't recover and that concept serves us quite poorly when we start to move it forward into the early stages of illness as we certainly found clinically many years ago. And really I would say it is a conceptual flaw, it is like trying to for example find the cause of, to dissect out the different cause of renal disease by taking a group of patients with chronic renal failure as a spectrum of the disorder here. We have done it in psychosis by defining schizophrenia as the treatment resistant subset almost. We haven't really done that with depression so we have had a much more flexible attitude for depression. If you look at young people like Tom in the video, he obviously wasn't going to march into the surgery or into the clinic with a very clear diagnosis and you know, people do not come in with DSM4 check lists ticked off for you. On the other hand we have got be sure that we are not labelling normal behaviour, we have got to be very careful with young people, adolescents and young adults to make sure the way you started to present, you probably thought, oh well you are just an irritable teenager but then things unfold a little more to it so it is quite a difficult diagnostic task, especially when we are struggling with diagnostic concepts that we have actually got at the moment. I think a number of people were critical of these diagnostic concepts during the day. This is what the patients and families will say, they will sort of say this. It is amazing how many times you actually will hear a family member saying, well I just knew that something was not quite right but I could not put my finger on it and if a clinician sees a patient at this phase of illness, this early prodrome or phase, they also will be a bit mystified. The patient themselves, the young person themself will have difficulty articulating what's going on, what's wrong. The subjective changes will be quite ineffable in a way and hard to put into words. This is a slide from a British GP, David ?Chyer, whose daughter developed schizophrenia and he made the point, having struggled greatly to actually get help for his daughter in the early stages, that patients will present with normal terms for what's actually happening, the changes that are occurring and it's our job somehow to make sense of these in terms of recognising an emerging mental disorder and I think that this issue of the overlap between symptomatology and need for care and need for carers is a very critical thing, there have been some very important epidemiological studies done in the last few years by Jim ?Varnos and colleagues which have shown that the spectrum of psychotic symptoms, even frank psychotic symptoms in the general public is much wider than was previously thought but it does not necessarily equate with need for care and this is something that we have to be very conscious about when we are looking at young people presenting or potentially presenting with symptoms and being at risk in the way of being diagnosed with a mental disorder so it is something I think we have to research fairly carefully.

One of the things we have got working against is that young people tend not to believe that anything is wrong until the last possible moment and so there is a sense of invulnerability and denial that's operating against you and there are a whole lot of reasons for that of course, some of which we might be able to address better through better health literacy and de-stigmatising and creating much more youth friendly clinical settings for them to come to, and I will come back to this at the end but adolescence and young adulthood is the peak period for the onsets of mental disorders of all types pretty much. Mood disorders, bipolar disorder predominantly, substance abuse, psychoses, personality disorders and typically, as I think Phillip was mentioning you get a comorbid mess of all the above. And we haven't really researched the onset of disorder in a very professional way. We still, particularly around the onset how do we actually determine when an onset has actually occurred, what's the border with normality, how do you operationalise it? Do you pick up a DSM 4? It doesn't really have much to say about this. What's the potential benefit versus the harm for actually being diagnosed? I mentioned some of the other issues on the slides and the person who has probably done the most clear thinking on this is an American epidemiologist by the name of William Eton at John Hopkins and he has written some very clear things about how you actually might acquire a disorder in the first place.

So just coming back to the staging idea, I think it is something which we can probably only do in a crude way at this stage but it is something worth thinking about and I'll just put up some examples of the diseases where staging is applied and I think we have actually in recent years come up with a relative crude staging of psychotic disorders which hopefully I can show you in the next bit of the talk. Staging in psychiatry has not really been looked at. I think the key principal is that treatment needs differ by phase of illness and certainly in schizophrenia until the last ten years it was always assumed that treatment of someone who has had schizophrenia for twenty years was exactly the same as the treatment of a young person presenting with their very first episode and that's obviously completely wrong in terms of their psycho-social needs and also in terms of the biological treatments that are appropriate so I think we are starting to show that this is clearly a useful principal. The other thing I think is that the treatment might be more benign and more effective in the early stages. We certainly see that the treatment of first episode psychosis is much more effective than the treatment of late stage schizophrenia and also that treatments applied at that point can be more benign. For example much lower doses of medication are effective. Maybe a different range of medications will turn out to be effective at this stage and if we can develop and use neuro-protective agents such as were sort of tantalised by this morning. We could actually try those kinds of more benign and specific and subtle treatments in the early stages. They are more likely to work at that stage and more likely to be effective and benign. So I think that is the logic and that is the argument.

If you do start to do early intervention and try to find patients at earlier stages of illness you may be actually doing two things. You may be relating two group present. Group 2 is the obvious group, they're people who would otherwise progress to more severe forms of illness but you are getting them earlier and hopefully at a more receptive treatment phase but you may also be including a group of patients such as the ……?…….group who actually have got a more benign kind of condition and maybe there is a higher rate of self limiting illness and I certainly think that would be possible with depression but it may also be true with psychotic illness as well so there is a lot to think about in this staging issue and many of these ideas have come from general medicine where screening programs have been used for various forms of cancer and these arguments have been developed, particularly by an American epidemiologist by the name of Feinstein.

So just coming back to mental health, what I am arguing for and what I am sort of going to describe Chris has already talked about this morning and it is the application of a concept called indicator prevention in psychiatry and that is becoming probably the best bet in terms of where the preventive efforts ought to go at this stage. If you have seen this model before you would be aware that universal prevention is offered for the general population as a whole in terms of a population wide strategy. Selective prevention is for asymptomatic patients who are basically a high risk group on an increased risk of disorder but have no symptoms at this point. Indicated prevention involves patients with significant symptoms but not yet at the threshold of full disorder so sub-threshold disorder or they could be defined by the kind of ?endothema types that you have heard Pat and others talk about during the day. You can define a group for either category by the presence of endothema type as well so this is an important sort of potential category. A recent paper in the American Journal of Psychiatry by ?Kypers points out that the kinds of sample size that you need to assess the impact of universal strategies are basically prohibitive even probably for high incidence disorders. They are in the millions and certainly in the tens of thousands, very very expensive studies. That doesn't mean to say those sorts of activities are not worth doing. They may well be effective nevertheless but it is very difficult to show in research terms that they actually are effective and there has been a bit of a retreat I would say from these large scale types of interventions by the people who started off working in these areas. They seem to be moving more towards the more targeted or indicated sorts of approaches For obvious reasons the power that you need is much reduced, or the power and the sample size to get the power that you need is much reduced. So that is a paper well worth a look I think.

What we are trying to do obviously then is pick up people with early symptoms like Tom in the video who haven't yet made a big mess of their lives basically by being just allowed to progress and become more severely ill before help is offered and so we have got to pick the right point to do that, and that's what indicative prevention is doing. It is trying to prevent the symptoms developing into a coherent and disabling syndrome. Obviously another intervention focus could be picking first episode of a major psychiatric disorder like depression in bipolar or first episode of psychosis and trying to prevent that becoming a chronic and persistent relapsing sort of condition. So that is also a preventive thing to do and certainly something we could do today. In psychotic disorders, I am using this as the paradigm here, you can define three phases really for early intervention. One is this prepsychotic or prodromal phase. The second is trying to reduce the period of untreated psychosis in people who have reached the full threshold and the final thing is much more specific and effective and probably intensive treatment of the first episode in the first few years after the onset and these are things that we are really in a position to do already and as Chris hinted at they have already taken off around the world as a major focus in psychiatry developing programs to do this in young people with early psychosis. Looking at the pre-psychotic phase there are plenty advantages or potential advantages of doing this. We know that it is a prolonged phase, usually at least two years on average where the young person is becoming increasingly disabled but the symptomatology is not sufficiently clear to make a clear cut diagnosis of a psychotic disorder. The patients are accessible as you could see with the patient in the video. You can form a therapeutic relationship at least with a subset of these patients and this means that even if a person does progress on to a more clear cut psychotic illness they probably will accept your recommendations for treatment because they have got some help, they have engaged with you and formed a relationship and that certainly has been our experience that these young patients never need hospitalisation even if they do become more clearly psychotic, you could treat them solely in an outpatient setting which works very well. So there are cost indications in that as well.

Finally for today's audience I think we have certainly, it's made the study of the biology and transition and also the clinical epidemiology of onset, what kind of factors, what sort of cause and risk factors might be affecting this process of resolution or worsening. Certainly we have done a lot of work in that area and starting to report that more. This is what I am just trying to convey, the idea that the disability in the patient is greater than the actual prominence of the symptoms in this phase so we will see a young person who is failing at school, becoming more withdrawn from their peers but you can't really elicit the diagnostically specific symptoms the easily at that point and so they progress on, they develop their first episode of psychosis, you treat that effectively because we have got very effective treatments for first episode psychosis these days. 90% of patients will get a good remission of the positive symptoms but you are left with this disability which has gradually accumulated in the years before they actually presented with their first episode and it is very hard to shift once your peer group has moved on and you have dropped out, you've become dependent on cannabis, your family is stressed. All of those negative things have actually occurred. They are very hard to put back together, it is a real Humpty Dumpty kind of situation. Now there are problems with this obviously as has been struck in other areas of medicine as well. The issue of false positives, does our, a clinical screening approach, does it actually, how well does it predict. We can decrease the rate of false positives but then we reduce our sensitivity and we target a smaller group of patients, how generalisable and applicable is the issue for the inaccessible false positive issue and can we find, do we have the structures, do we have clinical programs to find them and this is a good thing to do. Give them the quality of standard care and schizophrenia under resourced mental health system in Australia. Is it a good thing to do to try to bring these young patients into those kinds of clinical settings? I would say no in many cases and we really need to set up much more supportive and affective youth friendly sort of environments. This is the screening problem. As you can see on the top row there. If a prevalence or I suppose more appropriately the annual incidence of this disorder is in the very low level the even if you have a perfect diagnostic test or screen on the second row there you can see that the positive predictor value is very very low. So you get a very high percentage of false positives which is not good so what that implies is you need a two stage process. You need to enrich a sample up to a true base rate on at least, I would say, 30%+ before a test, in this case a set of operational criteria that are going to predict more accurately transition to the first episode state. So this is a general law and you can see it is from David Sackett's book on clinical epidemiology and it is something that is a structural property of screening ……..So really we are looking for a needles in a haystack here, we are looking…………disorder. So how do we do this, how do we recognise the individuals of high risk for psychosis as Chris was describing? Many of the prodromal features are non-specific and they have low predictor power. So what we try to do and it is based on an idea in a paper by Richard Bell in 1992 which he called the close in strategy and gets round many of the problems of earlier studies, early high risk studies. You are taking people close to the onset and you are actually using the subthreshold symptomatology as a risk factor for conversion to the disorder and Bill Eton was the first person to actually do this. You are combining it with other features such as the age range of the patients and family history and so on but there actually aren't that many risk factors that are useful for this purpose in psychiatry so the list actually turns out to be enough as I will show you. Clinically how do you actually enrich the sample. Well we have tried to do it by creating clinical structures that allow access to these patients. Originally with the ethic program when we saw a high volume of first episode, first psychosis patients each year which meant that there was a subset of patients who were sub-threshold who actually came into our clinical environment and it was a very proactive attempt to engage these patients and to structure the whole direction in the way that suited them. That seemed to work and more recently we developed a much broader spectrum youth mental health model which hopefully is working in an even more potent way. For example we see 2,000, we get 2,000 referrals into our triage and entry system each year now so I think that is the kind of effort that you have go to if you are going to get the enriched samples. One point to make, that the base rate can fluctuate depending on how you do sample in this way, if it is very much an ad hoc or poorly controlled method of sampling you are you are not going to get a communicative sample with screening you are actually are creating access through various means so you might actually get fluctuations in the transition rate of the sampling depending on where the referral sources are coming from. Something we have seen actually in practice. So what we are doing instead of looking for needles, looking for anything that is sharp and metal basically. That is the concept and we tried to use this magnetic strategy. You can see how concrete I am can't you. The threshold for transition is actually arbitrary and who says that in a week or two weeks of confluent symptoms is sufficient for onset diagnosis and for commencing, say antipsychotic medication so different clinics like this which have developed around the world now have slightly different thresholds for this and it is an arbitrary thing and comes back to the whole notion of how do you define when someone has across the threshold categorically for the disorder because are really dealing with dimensional features, dimensional phenomena. I won't go into detail, they have been operationalised using variety of rating scales but we did …..humanise, operationalise what we thought was this at risk mental state and, this ultra high risk mental state using three separate definitions, one of which involved family history plus reduced functioning but the others were already plus reduced functioning but the others were already focussed around attenuated or some threshold psychotic symptoms and the transition rate turned out with these criteria in that particularly setting to be about 35% of the sample. This is published earlier this year in Schizophrenia Research. You can see that they are impaired, their gaff score is 60, they have got very significant levels of negative symptoms in particular. A san score of 18 there and we can rack up the prediction power with some very simple clinical variables listed there, any one of those four variables on that slide will increase the positive predictor value from 35% up to 80% so you are getting much more potent prediction. On the other hand you are reducing the number of actual cases in that sample so you are probably are being more selective so it's a bit of a trade off as you can see. Chris talked about some of the neuro-cognition work this morning, well ?Shona France and Warwick Brewer have been involved in this along with Steven Wood and Chris Nixon for smell, as probably the most clear predictor. Most of the other neuro-cognitions are disappointing. We thought that it would be quite annotive in relation to clinical variables but apart from a few very specific clues that Chris already mentioned it did not really add much to the prediction which was surprising. We think that is because all the other patients who didn't convert or didn't make the transition to first episode also had neuro-cognitive impairments as well. They had the same kind of phenotype if you like and they certainly had a lot of psychiatric symptoms but they had a similar, I guess a similar pattern of impairment. Chris already showed you that and I what I would like to show you next is, what do we do about this, we have been this for about eight or nine years and we have been watching the patients despite fairly heavy psycho-social input, continue to progress to psychosis in about 30 to 40% of cases so we felt it was appropriate to try something a little bit more specific in terms of trying to reduce this transition rate. We published a trial at the end of last year with more specific treatments as well as needs based intervention which were supportive interventions, we added some low dose Risperidone and cognitive behaviour therapy to the basic psycho-social treatment and were able to show that the transition rate could be reduced quite significantly during those six month periods so it is possible to at least delay the progress of these patients into their first episode of psychosis. If you follow them up for another six months after the treatment is withdrawn you see there is a catch-up. The first six months of treatment is on the left and the next six months of follow-up without specific treatment is on the right and interestingly over the far right hand side you see a group of fourteen patients who actually took the Risperidone on a very regular basis, they had good compliance with medication and they actually got some carry over protection even when they ceased taking the drug so there was some continued benefit for that group of patients, so in a way you could say they had a course of medication which gave them some protection for at least twelve months. Now these are very small numbers so you wouldn't want to get too carried away with this and you'd certainly need replication but they're, I would say, are indicative of the capacity to at least slow the progression of these illnesses and maybe there are even better treatments in the wings that we could try for these purposes. Certainly the psychological treatment was also effective as well. The number needed to treat was about four so it is a potent intervention in terms of reduction of risk. It compares with treatment of moderate hypertension for stroke with an NNT of 13. There is centre in the US at Yale headed by Tom McGlassen which is replicating these findings now with alanzapine and showing that these symptoms respond also to another a typical antipsychotic.

I am just going to finish the last couple of minutes to try to broaden the focus a little bit to show you I hope that this approach, this idea of focussing on the onset both for treatment purposes and also for more basic research about the process involved has got wider applicability. Some data from an epidemiological surveys in Australia and overseas showing that the onset phase of most of the major psychiatric disorders is actually in this age group, this adolescent young adult age group. You can see that the methodologies are different but there is a jump in the prevalence of disorder between the adolescent years and the young adult phase and 75% of symptomatology commences of the adult type disorders before age 24. So if we are going to look at incidence of an onset we have to focus on young people predominantly. This is supported by Victorian burden of disease study showing the same mean in different sort of groups, pretty much clustering around late teens and the early 20s in terms of incidence. In fact it is the only game in town in this age group. If you are talking about young people's health it is their mental health and substance use that is the issue. 70% of ?Dowie, 60% of the disability and adjusted life years in this age group is made up of mental disorders and substance use disorders so it's a very important health issue for young people. It is actually increasing according to the British psychiatrist, Michael ?Lud… who made an exhaustive study of the phenomenon some years ago and since the second world war and you probably can tell from just picking up the newspaper every morning, drugs, suicide, mental health problems have really become a very significant issue and they seem to be increasing as far as we can tell so it does need better research. There are a number of tasks obviously, I tried to list them on this line but I think one of the key things is to provide better access and better environment for young people to actually receive treatment. The primary care environment is not accessible to them particularly. The number of ………….is showing that now. It is not geared for them. Just like the specialist mental health system is not geared for them despite the fact that they should be really the prime client and there is tremendous potential for reducing the health burden and focussing on this group and you can see there is a big health issue for this age group and as someone has said this morning if you become disabled for whatever reason in this age range then it's got major human and social costs, economic costs. And I am just going to finish by giving a sense of what my colleagues are sort of involved in. We broadened our early psychosis focus in the last few years to a broader youth mental health focus. We see a standard case load in our service now is about 750 young people between the age of 15 and 24, half of those have got psychotic disorders, the other half non-psychotic now. There is a population base of about 1,000,000. There would be tens of thousands of young people in that catchment area who could potentially benefit from some kind of mental health intervention so you can see now how under-resourced even within a small focus the mental service actually is. We tried to integrate a new research centre with this clinical service in a very tight way and Chris is obviously part of that and he is very integrated within that and so there are other collaborations involved as well.

The key objectives of the research program coming back to the… of what I think was highlighted this morning is we really think we have to develop some kind of more practical clinical diagnostic strategy to tackle these problems. It may not be the same as the research diagnostic criteria, you may need a different set of research diagnostic criteria to actually understand more about the process ………..phenotypes I suppose but certainly in a clinical sense we need to have a much useful sort of clinical approach that means obviously a lot of clinical research. We need to understand how to develop a much more sort of receptive and integrated, I suppose stigma free service model for young people, and particularly so with a primary care centre which needs to be developed as well. We new treatments, both psycho-social and biological and I think there is tremendous scope for both of those. There are some disorders, very serious ones that we have virtually no effective treatments for. The treatment for anorexia nervosa has not progressed in the last several decades, a very very serious disorder. Borderline personality disorder, a major public health problem, very little research evidence in relation to its treatment and finally for today obviously the contributory causes, and I'm also talking about the causes that are going to be useful for us in terms of treatment, not just understanding the genetic basis but actually understanding risk factors that are going to help us do something positive for the mental health of young people in the short term as well. So I've got a number of colleagues working in each of these as a leader for each of these fields and obviously significant collaborations especially around neuro-?detection focus and we would like to develop more and more links with people and it was great to be part of today for that reason.

So in conclusion I have certainly felt that we need to learn a lot more about lessons of early intervention strategies and prevention from the rest of the medical field. Especially not having to re-invent the wheel about some of these strategies. We certainly need to develop a method for ?onset cases and maybe we can actually develop a staging model, it might depend on the particular syndrome but it is one window, it is not the only one obviously but it has been particularly useful from our point of view and collaboration in Melbourne looking at this onset phase and I think there is a crying need for better interventions, better treatment but even the treatments that we currently have, particularly for psychoses, they have never been better but we are not delivering them in the real world in an effective way so that is something we could start working on today really and I think it might be bipolar disorder as well. So there is the old adage an ounce of prevention is better than a pound of cure. Well at the moment in Australia we have got a ton of standard care in psychiatry and it is not necessarily working as well as it could so we think, and hopefully try persuade people to put a few little kilos of of early intervention around the place, especially focussing on young people and we might actually get some more benefits and also research might benefit as well. So I will stop there, thanks very much.

© 2004, Brain and Mind Australia Inc. - Copyright Notice -