Skip to content

Monthly Archives: May 2013

David Clark on Improving Access for Psychological Therapy (IAPT)

Five years ago, the British government launched a mental health initiative called Improving Access for Psychological Therapy (IAPT), which hugely expanded the provision of talking therapies within the National Health Service, with the aim of getting therapy for depression and anxiety to just under one million adults a year. It is the biggest expansion of mental health services anywhere in the world, ever – and arguably the only instance of a government providing free talking therapy on a mass scale. IAPT was the brain-child of Britain’s leading expert in Cognitive Behavioural Therapy, Professor David Clark. Here’s an interview I did with him, which I used for an article I wrote on IAPT for Aeon magazine, published this week (for which I interviewed several other people involved with or using IAPT).

How did IAPT come about?

I think the first thing is NICE in 2004 starting publishing guidelines on the treatment of different mental health problems, and pointing out that for anxiety and depression there was good evidence for some therapies being effective. And then a number of people noted that the British public wasn’t getting much access to these treatments. Richard Layard and I met at around that time. And we formed a partnership to try and put forward an argument, based on the fact that most people weren’t getting access to the NICE-recommended evidence-based psychological therapies, and that if they did get access to them, and they were properly implemented, then this would be a programme that would not only provide people with treatments that were helpful, but it would also be economically viable. Richard’s ability to put a cost-effectiveness to the argument was very helpful.

How did you and Richard Layard meet?

Completely by chance. We were both being elected fellows of the British Academy one day. We were standing next to each other having a cup of tea, and we introduced ourselves. Richard explained he was an economist but writing a book on happiness, and was writing a chapter on mental health, and did I know anything about mental health, and I said, well, I’ve spent most of my life developing psychological treatments, so we had a lot to talk about.

We put together a paper which went to the government, an internal briefing paper, laying out the case for what became IAPT. The Cabinet Office organised the seminar. Richard laid out the broad case and I had to cover the evidence base for psychological therapies. That, alongside other arguments at the time, led to the 2005 Labour election manifesto commitment to increasing access to NICE-recommended psychological treatments.

To what extent was IAPT a step forward?

The exact way IAPT framed came about from lots of discussion after the election between lots of people. The Department of Health set up an expert reference group to map it out. But it has a lot of pretty radical elements to it. The first is the very high level of outcome monitoring. Prior to IAPT, I dont think there was a single service anywhere in the country, where you could go long, if you were suffering from anxiety or depression, and say ‘if I go to you, what’s my chance of recovery?’

There were quite a lot of services that were doing their best to collect outcome data, but they might give a questionnaire at the beginning of the treatment, and then maybe at the end, but there would be a lot of variability on when people ended their treatment, so there was a lot of missing data. On average, those services trying to monitor outcomes would get data on 30% of their patients. IAPT has changed that by adopting a session-by-session monitoring system, so that if someone finishes therapy a bit earlier than you anticipated, you still have data on how they’ve done, up to that session.

That was based on something we developed in Northern Ireland, following the Omagh bomb in 1998. They set up a walk-in community service, and of course you didn’t know if people would be coming for lots of sessions or not, so we developed a system for measuring outcomes each time. That turned out to be very helpful because it meant we could show the Northern Ireland office the results of the first 100 or so people that we saw. And they did very well. They improved as much as in randomised controlled trials in universities, although this was a community service that helped everyone. That led to the government of the day funding a treatment centre called the Northern Ireland Centre for Trauma and Transformation, which then made these treatments available to victims of other terrorist attacks.

So we built on that, and thought we should be able to adopt the same principles, and get data on most people. That has been really crucial in several ways. The first is, it allows for continued political support in the programme, because the government knows what it is getting for its money. Prior to that we didn’t really know. You might reduce waiting lists but we didn’t know: are people getting better, and are more people getting better? Now we know.

It also has produced extraordinary transparency in mental health. We’ve been very keen that the results are published every quarter – the outcome data from every IAPT service in the country. This sort of information has never been available to users of mental health services. And I think going forward it’s likely to be a real driver for quality improvements, because it allows services to bench-mark themselves against other services. There is variability of course, but once you know about it, you can ask, how do we move the services that are not doing so well closer to those that are – just as Bruce Keogh did for cardio-vascular surgery.

I see this as one of the big achievements. There were a lot of professionals who were against this sort of outcome-monitoring. They thought it was too much of a hassle for patients, and maybe they wouldn’t like it. But our experience has been, that actually patients really love it, and they find it really helpful.

To what extent was IAPT an expansion of services?

The obvious problem was we didn’t have enough therapists. So the heart of the programme was training new therapists in evidence-based therapies. Then the big decision point was, obviously you can only train people at a certain rate. Train some people one year and another cohort the next year. Should we take the first year of trainees and distribute them in small numbers to existing services? In which case there’s a good chance they’ll just get lost in the system. Or should we try and create a small number of new services more or less at full capacity, and get them to open their doors and properly function. We decided that’s the way we’ll go, because the training is more rigorous in terms of following evidence-based protocol, everyone in the services would do outcome monitoring, and not everyone in routine services was doing that. And we needed for everyone to have good supervision in the services. So we thought we’d create new services, small numbers in the first year, and then spread them around.

Why do you think there was the political will suddenly to substantially increase government resources into mental health services?

The economic argument was very influential. We argued that the programme would largely pay for itself. When people are less depressed and anxious they’re more productive at work. They’re more likely to get into work. They also cost the NHS quite a lot less in terms of unnecessary physical investigations. So someone with panic disorder will get lots of physical investigations, they worry about their heart or whatever. I think the government was persuaded by the argument that this is really an invest to save programme, it’s not a costing. It’s also having the benefit that a lot of people are getting better. That was the argument that worked. It probably wouldn’t have got off the ground if it wasn’t a coalition between three different groups: obviously academic clinicians like myself, who are arguing that these are effective treatments which people should be able to access; secondly the economists, above all Richard Layard; and then many voluntary sector groups like MIND and Re:Think, who were very aware that lots of patients were complaining that all they got was drugs, and a lot of them would like psychological therapy.

Five years on, how successful has IAPT been?

Nationally, it’s more or less on track. We had targets for the number of people seen and the clinical outcomes at this stage. And we’re more or less on those targets. We’ve now got an outcome monitoring system in place. And commissioners are realising that you commission services for whether or not they get better, rather than just waiting lists. That has changed the way commissioners think of mental health, and that will have an enduring impact not just on IAPT but throughout mental health.

But we’re not finished, in terms of the number of people who are meant to be seen. We’re aiming for 15% of those with depression and anxiety, and we’re operating at 10% at the moment, so there’s a big step up to be done there. We’ve only got about two thirds of the people trained so far. The other thing is this is all happening in the context of reforms and changes to the health service. And some of those changes slow things down, because people who might be commissioning a service are different from one week to the next. So there’s a lot of uncertainty.

And the data that’s publicly reported at the moment is fairly simplistic – it’s just some index about whether people have dropped below some clinical cut-off for recovery. We collect much more sophisticated information about how much people have improved, across a lot of different domains, and also what sort of people they are, ethnic groups, disabilities and things. We want to make all of that information available, which will happen soon. When that happens, people will discover lots of things they would like to improve. I suspect some services are much better than others at giving different parts of the community access to the services. And also with a more detailed data we’ll get a much better idea of who benefits and who doesn’t. And that information all needs to be fed back into the services. I see the service as a continued improvement initiative. They’re up and running, but there’s a lot more that needs to be done.

Does the data show the recovery rates of different disorders?

It will do, in about three months time. Up until recently, we’ve relied on the commissioners of the services sending headline figures – the number of people recovered, the number of people seen. But about eight months ago, we shifted to a system where individual patient level information goes to the NHS information centre. It becomes anonymous, but all the data – ethnicity, disability, what type of problem they had, what kind of treatment they had, how much they improved – all that flows now, nationally. In about three months time you’ll get very detailed reports. Nowhere in the world is that kind of detail available.

So the data at the moment shows recovery rates of about 40%?

The current rate is 46%.

Is that for people who complete the course of therapy?

It’s for people who’ve been seen at least twice. In psychotherapy research there was a period where people would record ‘completer analysis’, on people who had got through a full course of therapy, and rather ignore the people who dropped out, and that’s completely wrong. You don’t get a fair idea of how a service is doing unless you basically take more or less everyone, which is what you call an ‘intention to treat’. And in IAPT we say, if you’ve come along for at least two treatments, then you’ve engaged with the treatment and we should be reporting the outcome. Now a lot of the therapists might say, the course of the treatment might be eight sessions and they only came for three, they dropped out. But in our national reporting we ignore that, we still give their data. Wherever they finished is where they got to.

How reliable is the data? If patients are filling it out and handing it in to the therapist, they might not want to offend the therapist. Does that create a bias in the data?

There are risks of that sort. They probably operate at different levels. If you take the high intensity therapy – face-to-face CBT – the patient fills out the forms in the waiting room, not in front of the therapist. It’s rather like what you do in a randomised controlled trial. And we know, there’s a large literature from those trials, when people fill in measures that way, but they’re also seen by an independent assessor, who doesn’t know what treatment they’ve had, and you seem to get similar results, from independent non-biased assessment as from patients’ self-reported assessment collected that way. So I feel reasonably confident that the self-report data that you get from traditional high intensity therapy is pretty valid.

But we also have low intensity work, where people are having guided self-help, and quite a lot of that is done on the phone. And at the moment the therapist asks the outcome measurement questions on the phone and the patient answers. It seems to me there is more potential for a demand effect there. And it’s not an ideal situation. We should be moving to a more automated situation which goes automatically into the IT system. It also frees up more therapy time.

To what extent are referrals and applications for therapy going up?

It’s continuing to increase. A key feature for IAPT, another revolution I suppose, is self-referral. When the NHS was created in the 1940s, everyone was concerned about the cost. And so partly for that reason, we created a universal GP system. So the GP is the gate-keeper for the costs, so you don’t normally get specialist treatment unless your GP refers you. That was certainly true of specialist treatments for anxiety and depression.

But when we did two pilot projects for IAPT, in Newham and Doncaster, we discussed with the Department of Health that it’s possible in mental health that the GP referral only system is excluding systematically people that really need help. That could be for reasons of stigma, they may be unhappy about talking to their GP about the problem, if they know the GP isn’t going to do the treatment; and some groups may be more averse to coming along to primary care than others. So we asked, will you let us experiment with self-referral. And some people thought, this is a dangerous way to go, because you’ll be flooded by very mild cases who perhaps don’t need so much attention.

But what we found was really the opposite. If you compare the self-referrers and the GP-referred, the self-referrers were as severe, but tended to have the problem longer, and were more representative of the community in general. This was particularly clear in Newham, which is a very ethnically diverse borough. We found that the rate at which different ethnic groups came into the service pretty much tracked their prevalence in the community in the self-referred route, but in the GP route, people from black and ethnic groups were seriously under-represented.

Does that mean they were more likely to be prescribed chemical treatments?

It might be, we don’t know. That led Alan Johnson, when he announced the IAPT initiative, to be really radical in NHS terms, and say, this new initiative will be open to self-referral everywhere. I think that’s helped and improved fairness of access. Interestingly, if you look at the data, those people who self-refer are just as likely to recover as GP-referrals, which is good, but they recover with less therapy sessions. We think that’s probably because, if you go through the process of self-referral you might look at the website of a service, and see a description of the problems that they treat, and you think through ‘is this right for me?’. You’re almost socialised into the process when you come along.

Has IAPT had any impact on anti-depressant prescriptions?

We don’t have in the same database anti-depressant prescriptions so we don’t know. What we have looked at is people who were on anti-depressants at the start of the treatment and what happened to them, and you had more people coming off than staying on. But we don’t have a large connected up database.

What about relapse rates, can we know?

This is a weakness of the current IAPT system, and one we’d like to see changing. Most IAPT services will not do a systematic follow-up. Commissioners have been very keen that services see a large number of people and haven’t been particularly keen to pay for follow up. And I think that’s a false economy, particularly because some of the treatments like CBT provide good evidence that they provide durable effect, but they have built in to the treatments relapse prevention programmes that you do for the last couple of sessions. Really what you’d want to do is see if they’re working for everyone, and quite a lot of these relapse prevention programmes have things like, if you’re noticing a bit of a set back, look at your notes, and then ring up your therapist and come in for a booster session. At the moment only a few services do that, so as things move forward we’d like that to happen. Research-wise, in the Newham and Doncaster pilot sites, we did a nine month follow up. What we found was in Newham people were as well after nine months as they were at the end of treatment. In Doncaster, there was a small but significant drop back but they were still a lot better than when they started.

It seems that the drop out rate is quite high.

Well, if you say that people finish treatment in less time than was expected, that’s true. But that’s not what is taken into account in the outcomes. So people who dropped out are still reported in the outcomes.

But is there concern that a lot of people are just dropping out and deciding this treatment is not for them?

I don’t think so. Where the issue of uncertainty is is somewhere different. In the original model, there was the idea that some people would actually benefit a lot from having a good assessment, helping see what their problem is, and getting some sort of simpler advice, and maybe being sign-posted to somewhere else, like debt counselling for example. So they might only have one session, but it might be very satisfactory. And then there were other people where they definitely need psychological treatment, so they’d come in for regular therapy. The weakness of the system at the moment, in terms of national reporting, is we only have outcome reporting for people who come in for therapy. We have a lot of people who have this one session – it’s about 40%. There is no reporting of whether than one session was a satisfactory one or an unsatisfactory one. There might be people who have that one session, who are offered therapy and decide they don’t want it, because they don’t like this service, which is a bad outcome.

Or maybe not been offered therapy, and needed it.

Yeah. So I think it’s been a serious weakness of the reporting so far, that we just don’t know and can’t distinguish between those possibilities. We need to know exactly what happened. And it’s probably very variable between different services. The Department of Health has recently agreed a coding for these one-off sessions which would allow people to be classified as ‘reasonable outcome’ or not. Were they discharged after mutual consent  with advice and sign posting, or were they offered therapy, because they thought they needed it, and they said ‘piss off, I’m not interested’. That’s what we need to get the numbers on.

OK. In IAPT, a lot of the ‘guided self-help’ work is done by PWPs (Psychological Well-Being Practitioners). I’ve spoken to some PWPs, and they express some concern that sometimes they’re seeing cases they’re not trained to handle, after a year’s training. And of course, there’s quite a lot of stress, big workload, and they didn’t always feel the promotion avenues are that open. Is that a concern?

These are all serious concerns. The PWP role is the most novel bit from a clinical viewpoint. It exists because there were a reasonable number of trials supporting guided self-help. But creating a workforce that delivers it and that operates within its capabilities, and with reasonable career progression is a completely new beast. And a lot of learning is coming out of where we have go to so far. And there’s no doubt that in services, PWPs are seeing cases that are more complicated than would be appropriate for the training they have.

In terms of career progression, it’s true there isn’t much career progression. But some services are working hard at that, in terms of creating a new position of senior PWPs, and some people going into training PWPs as well as doing senior clinical work and supervising other PWPs. But the turnover of people in PWP posts is much higher than in high intensity posts. And that raises questions about IAPT. If the turnover is high because they’re staying in IAPT but going to work in a neighbouring service which gives them a senior role rather than a junior role, that’s good. If however they are being lost to the system or they’re all doing high intensity training, then our estimate of how much that workforce costs are wrong, because we’re having to do constant training of new PWPs. Obviously the big argument for PWP work is its an economic way of providing treatment for mild to moderate cases, but if it turns out it costs twice as much as we think, because of the training costs, then that part of the argument might not hold up. One would have to re-think the whole role.

IAPT’s Stepped Care approach

The original idea was that IAPT pays for itself. Has it done that?

If you look at the number of people who move from being unemployed to employed or part-employed, the number is in line with the projections in Richard’s analysis.

Is there a risk that CBT could become overhyped, and seen as a silver bullet by politicians?

IAPT isn’t just about CBT, it’s about implementing NICE guidence. We are actually supporting four other non-CBT therapies in IAPT, and a third of the IAPT workforce can deliver these non-CBT therapies. Interpersonal Psychotherapy, Couples Therapy, Counselling, and Behavioural Activation (though I suppose that’s also CBT), and Brief Psychodynamic Therapy. The view we’ve taken is, if NICE recommends several therapies for a condition, then patient choice should operate, because people are more likely to get better from something they’ve chosen. If NICE only recommends one treatment for a disorder, then that’s all we’ll provide in the service. The idea is the offer will change as NICE guidance changes. In depression, the offer is already beyond CBT.

Do you think there will be more choice in IAPT as we go forward?

I think so yes, for a lot of reasons. One is because as it matures we’re able to put more emphasis on training people in other therapies. The other thing is, a crucial message has gone out to the therapy community, which is: if you record outcomes and results, there’s going to be investment. They’ve never known that before. Previously, you do your best to argue with commissioners, that we need some more therapy because a lot of people are handicapped, but if you’re not presenting outcome data, commissioners have been hesitant to invest. Now we’ve seen that you can get real investment, if you can show that patients get better with measurements that people are happy with. That’s meant there’s been a big increase in interest in people doing controlled evaluations of a whole range of psychological therapies. There’s now much more outcome research being done now. The consequence of that is we’re going to learn a lot more about a range of therapies that work. So we will be able to support more therapies as time goes on, because of the sea change in attitudes to evaluation that has come from the IAPT programme.

I was helped by CBT in my early twenties, but I do also see studies which suggest the Dodo theory – lots of other therapies do just as well. Could be the therapeutic alliance or perhaps we don’t know. What do you think of the Dodo theory?

The Dodo theory – named after the race in Alice in Wonderland in which ‘all shall have prizes’ – suggests that all therapies work equally well.

It’s not a false theory but it also doesn’t quite say what it seems to. NICE doesn’t endorse the Dodo theory. There aren’t NICE guidelines saying ‘just do any psychological therapy’, which is what you’d think from the Dodo account. And the reason is no NICE guidelines say that is, if you look at specific conditions and you look at all the randomised controlled trials, you get strong support for some therapies, less strong for others, and also evidence that some don’t work. So that is the position within conditions. But of course in these RCTs, you’re almost always using highly trained therapists. But in the databases that are used for the Dodo account, you might just use anyone who is giving therapy in large services, and ask ‘what therapies are you giving?’, and then you get less marked differences, but you also don’t quite know what therapies they’re giving. If you take a large service of therapists, some of whom are not highly trained, then some of these differences wash out. But if you’re dealing with highly trained therapists, then there’s clear evidence that certain ways of doing therapy work better than others.

Now regarding the allegiance bias, it’s a perfectly reasonable hypothesis. But most of the data advanced to support it is post-hoc. So the argument is advanced is this: if you’re involved in developing a treatment, you’re going to be very enthusiastic about it. You’re going to give a really good go to your treatment, but be a bit half-hearted about the rival treatment. A lot of the evidence for the allegiance bias comes from researchers looking back at trials, and saying, looking at the address of those authors, I think they were in favour of IPT, or in favour of CBT, and then let’s code them that way, and then let’s look at the data and whether it fits with allegiance. But that is very post-hoc, and it’s non-blind rating by people who have a view – they believe in allegiance. What you need to do is prospective experimental test. The classic way to do that is to take two therapy centres, use them both in a trial, and use one of them because it’s expert in one therapy, and the other because it’s an expert in the other therapy, then you train therapists in both centres in both treatments. Now you have a proper design and you can distinguish between the allegiance hypothesis versus the procedures. This has been done many times, and the answers is, it’s not the allegiance. Unfortunately people who advocate the allegiance hypothesis forget to mention these studies.

IAPT focuses people very much on training people up to the standard you’d expect in trials, using national curricula, with an emphasis very much on quality.

Some psychodynamic therapists say their services are being cut while IAPT funding is being protected. And IAPT therapists say they’re seeing patients with things like bipolar disorder etc, which is putting a strain on them as well. Is that happening?

This is a very serious issue. If that’s generally true that’s a big problem, as the whole point of IAPT is additionality – it’s not meant to be removing existing services, it’s meant to be creating a whole new set of services which create additional capacity and improved access. If instead commissioners are saying ‘we have IAPT so we’ll close down the other stuff’, that’s undermining the whole point. Because of this worry, the Department of Health asked all PCTs last year to return information on how much they were spending on talking therapies that were not IAPT and on IAPT. They have this data going back to 2004. The spend on non-IAPT services nationally has held up  – there’s not a reduction. As a consequence, before IAPT, the NHS was spending 3% of total mental health budget on talking therapies. It’s now 6.6%. The national picture is this isn’t happening. That doesn’t mean that in some areas there are services that have been going for some time, which commissioners have now chosen to decommission, which people feel aggrieved about.

Are some people being squeezed into IAPT services with problems that IAPT people haven’t been trained to treat?

That may be happening, and it shouldn’t be.

A broader criticism of CBT is that it focuses too much on individual thinking errors and not enough what might be genuine environmental adversities. Are we trying to treat cognitively things that might be economic or environmental?

IAPT was never envisaged to be solely a psychological treatment initiative. Right from the start we built into all the services employment advisors and debt counsellors, because depression and anxiety occurs in a social context and there are some things that the services can do to help that. That’s why right from the start we said it isn’t just CBT or any other therapy, it’s also some help with social problems. The other thing is that, yes, it’s true that on an actuarial basis you get higher rates of mental health problems in more deprived areas. But it’s also true that if you equip people to be more robust in the face of adversity, they’re less likely to suffer in a protracted way. So it’s not an either / or. Try and help as much as you can with the social adversity, and also equip people with the mental skills to manage that adversity.

Another critique of CBT is that it’s too individualistic, very much focused on the social. But it seems there is more group CBT beginning to be able in IAPT. Is that the case?

Yes quite a lot of the services have groups run. And NICE recommends group CBT for depression. I think one of the things which is a misunderstanding of CBT is to say ‘CBT is CBT is CBT’. It isn’t. When done properly it focuses on your particular concerns, your social circumstances, your behaviours. While there are broad themes covered, it should feel very different for different patients. It’s not like giving a drug, you don’t do the same thing with everyone.

Yes, it can feel quite cookie-cutter – you go to see a PWP and come away with a list of thinking errors to watch out for.

Yes, though what you’re describing is more psycho-education than face-to-face CBT, which would look very different. The idea in the IAPT services is if people don’t find psycho-education helpful, they should step up to face-to-face therapy. Step up rates vary – we did a study from the pilot schemes, and we found several features of services that predicted better overall outcomes. And one of them was having a high step-up rate. So if you are a service where if someone doesn’t recover at low intensity, you are very likely to step them up to high intensity, then your service is likely to have higher recovery rates. There were some services with a very low step up rate, and those services had much poorer outcomes.

The risk is the PWP might see it as a failure if they have’t cured the person.

That could happen, but obviously that’s not the model.

At the end of the 10 weeks…

10 weeks?

Isn’t that how long the treatment is typically?

It varies from service to service. Some are much more flexible in the number of sessions. And one of the other predictors of overall success is a higher average number of sessions per patient. NICE doesn’t just recommend treatments, it says there’s a dose of them. For depression it’s nearer 20 sessions. So we would say if you’re fully NICE compliant you should offer up to those numbers.

What are the options for a service user after that course, if they want to keep practicing? Are there options in community groups etc?

There are some IAPT services that run continuing groups. As services mature they need to focus more on this longer-term monitoring and follow up and continuation.

Is mindfulness CBT a growing part of IAPT?

NICE recommends mindfulness CBT for prevention of depression recurrence. So the evidence for it is restricted to those who have had at least three episodes of depression. They are at a much higher risk of recurrence. There are two therapies NICE recommend to reduce that risk. Drugs don’t reduce it, but high intensity CBT and mindfulness CBT both have good evidence that they halve that risk. We are encouraging IAPT services to offer one or other of these treatments. Mindfulness CBT is offered when people are recovered – it’s like a class people go to.

Can you tell me about Any Qualified Provider and what it means for mental health.

It’s just starting, so we don’t quite know how it will pan out. The idea is that the government would like to open up the provision of a lot of healthcare to groups that are suitably qualified and produce a more competitive market. In IAPT we do already have a multitude of providers. For example, in some IAPT services, low intensity help is managed by a voluntary group like MIND or Re:Think. There are many examples of that working very well and being reasonably economical.

But as the market opens up more, there are big risks, and the risk is people use too simplistic a method for organising payment. AQP could mean just payment by result. If you say ‘we’re going to pay you if someone reaches clinical recovery’, which could be a temptation for a commissioner, then you’d be providing a perverse incentive to services to only see mild cases, because they’re closer to the cut off so you don’t need to improve so much to get the money. It would be a travesty if we started commissioning services that way.

Or to duke the figures.

Yes. These are all risks. So if the benefits of competition are not to be outweighed by the perverse incentives, you do need a much more sophisticated way of assessing outcomes, so you can avoid these perverse incentives.

Like improvement?

Yes. On the IAPT website there is a document on AQP on perverse incentives, and it explains these issues. The DH has got 20 pilots running which are collecting data for payment by results system, to work out what would be a fair system. It’s certainly not just going to recovery, it’s also improvement, fairness of access, and delivering NICE recommended treatments, so people don’t end up skimping, and patients suffering.

How do you see IAPT developing and what other countries are doing?

Peter Fonagy, managing the roll-out of IAPT in child services

In terms of how it’s developing in England, the coalition has made several commitments to expand it. The first one, which is really important, is to develop a children and young person’s IAPT. It’s based in CAMHS. What IAPT does is train up new people, and this time place them in CAMHS. It brings in universal outcome monitoring. And it also brings in some general service improvement initiatives. So try to bring everyone in the services to bring in evidence-based practice and monitoring outcomes.

Will that be a culture clash as child psychology is more psychoanalytic?

It seems to be going well, the child IAPT, perhaps because people have seen the adult IAPT. The national advisor is Peter Fonagy, the Freud memorial professor of psychoanalysis. It’s not just CBT, it’s a range of therapies.

Will it mean more money?

Yes, but it’s started small. It’s more like proof of concept at the moment. Just like the adult one, new training has been developed with national curricula linked to agreed competencies.

The other things happening in England is focusing on people with long-term physical disease and mental health problems, like cardiovascular illness and depression. And these people tend not to have been seen so much in traditional therapy services in the past, which is a shame because actually if you can help them with depression, it’s much easier to manage their other conditions. There’s also a start of looking at IAPT-like services for psychosis and personality disorders. What that isn’t, is saying ‘let’s get all the people with psychosis treated in existing IAPT services’. It’s not that at all. What it is, is trying to build some of the ideas of IAPT into these secondary services, like the outcome monitoring, and training people in evidence-based protocols. And there are a number of pilot sites doing that around the country. We want a joined-up system where, whatever your illness and wherever you’re seen, there are some basic standards: the people who treat you will be fully trained, services look at how they’re doing and adjust their behaviour.

IAPT is an English initiative. The Scots are increasing access to psychological therapies, but not on the same scale. In Northern Ireland they are looking at implementing clinical guidelines. Outside of the UK, IAPT has been watched very closely. The Norwegians have opened up 14 IAPT-like services, which are very closely modeled on UK example, and they’ve been using the material that you can download from the IAPT website, and they’re using our outcome monitoring level. And there’s discussion about using IAPT-type services throughout Norway’s mental health services. That’s interesting, as Norway in a sense have more funding for this than we have, but still aren’t sure what they’re buying. So they want to move towards more evidence-based therapies.

What about Sweden?

I’ve just come back from Sweden, and they haven’t really moved into any national programme. The government is interested in evidence-based interventions. They’ve made some investment into back-to-work programmes….

Which haven’t worked that well?

That’s correct.

Is that a concern for IAPT?

Well, IAPT is not just a back-to-work programme.

What about the US and Canada?

It’s difficult to do it in the US. The Canadians have a health commission, but haven’t progressed as far as the IAPT system.

Is there a meaning gap to CBT?

CBT doesn’t really focus on meaning. If people wanted to explore that, then other therapies would be more suitable. But if that’s a limitation on the outcomes people get, that’s an empirical question. CBT doesn’t work for everyone. In an ideal world, perhaps we could judge who would be likely to respond to different approaches, but we haven’t got there yet.

Is the future bleak for existential and psychoanalytic therapies?

I certainly don’t think so, quite the opposite. We’re supporting training for brief psychoanalytic therapy for depression, and the people involved in that have agreed to do a randomised-controlled trial for that. People are doing evaluations of more psychoanalytic work. The psychoanalytic treatments are changing, and becoming more focused and differentiated for different conditions, which is good.

Are the days of dream analysis gone?

It doesn’t figure very prominently.

Have we lost something there?

There’s not a lot of evidence that it helps people get better to do it.

Here’s the transcript of the interview I did with Richard Layard for the Aeon piece.

A brief history of IAPT: the mass provision of CBT on the NHS

I’ve a long article in Aeon magazine this week, looking at Improving Access for Psychological Therapy (IAPT), which is the first ever provision of talking therapy on a mass scale by a government. Before IAPT, the NHS spent just 3% of its mental health budget on talking therapy. IAPT has tripled that budget, and aims to train 6,000 new therapists in CBT by 2014, who will treat 900,000 people for depression and anxiety annually in England and Wales. It is, as one therapist put it, ‘the biggest expansion of mental health services anywhere in the world, ever’. Quite a feat.

In the piece, I tell the story of how IAPT occurred because of a chance meeting at a British Academy tea party:

In 2003, Lord Richard Layard was made a fellow of the British Academy. He’d made his reputation as an unemployment economist at the London School of Economics, but he’d always had an interest in depression and happiness. He inherited this interest, perhaps, from his father, the anthropologist John Layard, who suffered from depression, shot himself in the head, survived, was analysed by Carl Jung, and then re-trained as a Jungian psychologist. Layard junior was more interested in hard data than the collective unconscious, but he’d become interested in a new field in economics that tried to measure individuals’ happiness, and use the data to guide public policy. Layard wondered: what if governments started to take happiness data as seriously as they took unemployment or inflation? He tells me: ‘The most obvious policy implication was for mental health services.’

At the British Academy tea party, Layard struck up a conversation with the man standing next to him, who was called David Clark. ‘It was a fortuitous meeting’, Layard tells me. Synchronicity, his father might have said. Layard asked Clark if he happened to know anything about mental health. Clark replied that he did. He was, in fact, the leading British practitioner of CBT. He had helped to set up a trauma centre in Omagh after the Provisional IRA bombing of that town in 1998. The centre treated Omagh citizens for post-traumatic stress disorder, and kept careful measurements of the outcomes. The data showed that front-line provision of CBT in the field showed comparable recovery results as in clinical trials: roughly 50% of people recovered. Clark explained to Layard that trials of CBT showed similar results for depression, anxiety and other emotional disorders. He also explained that there was very little CBT (or any other talking therapy) available on the NHS for common problems like depression. Layard, who is nothing if not a doer, decided he wanted to ‘get something done about mental health’. So, at the age of 70, that is what he did.

With Clark’s help, Layard assembled a powerful argument for the British government to increase its spending on CBT. Depression and anxiety affect one in six of the population. Besides causing a lot of human suffering, this costs the economy around £4 billion a year in lost productivity and incapacity benefits. This problem has a solution, Layard argued: CBT. The government’s own National Institute for Health and Care Excellence (NICE), which evaluates evidence to guide NHS spending, recommended CBT for depression and anxiety in 2004. Yet for some reason, the NHS just £80 million a year on talking therapies, out of a total NHS annual budget of £100 billion. Layard and Clark recommended doubling the budget, so that 15% of adults with depression and anxiety would get access to psychological therapy. Some of them would get off incapacity benefits in the process, it was argued, so the service would pay for itself.

Layard and Clark presented their recommendations at a seminar at 10 Downing Street in January 2005. They managed to get IAPT into New Labour’s manifesto for the 2005 election, and were then faced with the task of turning it into a reality following Labour’s election victory. Clark designed the service. Firstly, and radically for the NHS, it allowed for self-referrals. Secondly, the service would have a ‘stepped-care’ approach: for mild cases of depression and anxiety, people would be treated by ‘Psychological Well-Being Practitioners’, who had a year’s training in CBT, and who provide ‘psycho-education’ and guided self-help, often over the phone. If that wasn’t adequate, people were encouraged to ‘step up’ to more intensive face-to-face therapy for a longer period of time, with a fully-trained therapist. Thirdly, IAPT would only offer NICE-recommended evidence-based therapies, which meant mainly CBT. Finally, IAPT centres would measure outcomes at every therapy session, and make this data available online, so both patients and politicians could see the results.

The reason Layard and Clark convinced politicians to put serious money into talking therapies is that CBT had built up a big evidence base to show it worked. I look at the origins of this evidence – the invention of the ‘Beck Depression Inventory’:

Beck developed Cognitive Behavioural Therapy in the early 1960s. He tells me: “I was also influenced by the Stoics, who stated that it was the meaning of events rather than the events themselves that affected people. When this was articulated by Ellis, everything clicked into place.” While Ellis was content to be a free-wheeling rebel, Beck was more of an institution man. He wanted to transform clinical psychotherapy from within, by building up an empirical evidence base for cognitive therapy.

Before Beck, evidence for psychotherapy mainly consisted of therapists’ case studies. The reputation of psychoanalysis, for example, was built on a handful of canonical case studies written by Sigmund Freud, like ‘the Wolf-man’, ‘Dora’, and ‘Anna O’. The problem with that approach was the evidence was anecdotal, non-replicable, and relied strongly on the therapist’s own account of a patient’s progress. The therapist might exaggerate the success of a treatment, as Freud arguably did in the foundational case of Anna O.

Beck’s radical innovation was to develop a questionnaire which asked patients how they felt on a four-point scale. In 1961, he created the Beck Depression Inventory, a 21-question survey which measured a person’s beliefs and emotional state through questions like:

0 I do not feel like a failure.
1 I feel I have failed more than the average person.
2 As I look back on my life, all I can see is a lot of failures.
3 I feel I am a complete failure as a person.

By measuring the intensity of a person’s negative beliefs and feelings, Beck discovered a way to quantify emotions and turn them into data. Using the BDI, he could quantify how a person felt before a course of CBT, and after it. According to the BDI, after 10-20 weeks of CBT, around 50% of people with depression no longer met the diagnostic criteria for major depressive disorder. And, crucially, this result was replicable in randomised controlled trials by other therapists. CBT showed similar recovery rates for anxiety disorders like social anxiety and post-traumatic stress disorder.

Beck launched the era of ‘evidence-based therapy’. In doing so, however, he made some drastic alterations to the ancient philosophy that inspired him. He pruned out anything that was not scientifically measurable – including any mention of God or the Logos, virtue or vice, the good society, or our ethical obligations to other people. I once asked Beck if he agreed with Plato that certain forms of society encouraged particular emotional disorders. He replied: ‘I am loath to toss out an opinion that is not based on empirical evidence.’ There is much about which CBT is silent. It teaches you how to steer the self, but does not tell you where you should steer it to, nor what form of society might encourage us to flourish.

I wax lyrical about the place of IAPT in the history of ideas:

IAPT is an interesting moment not just in the history of psychotherapy, but in the history of philosophy. It is an attempt to teach Stoic – or ‘Stoic-lite’ – self-governance techniques to millions of people, an exercise in adult education as much as healthcare. The scale of it is beyond the dreams of the ancient Stoics, teaching on the street corners of Athens. Although the early Stoics wrote political works, they were all lost in antiquity, and later Roman Stoics viewed Stoicism more as a sort of individual self-help for the elite. Marcus Aurelius, the Stoic emperor of Rome, was in a position to spread Stoicism to the entire empire if he so wished, but he had a pessimistic sense of the limit of politics. ‘I must not expect Plato’s commonwealth’, he told himself. ‘[For] who can hope to alter men’s convictions, and without change of conviction what can there be but grudging subjection and feigned assent’.

Stoicism’s therapy of the emotions remained popular with intellectuals, but few believed it could be taught by the state to the masses. David Hume wrote that the majority of humanity is ‘effectually excluded from all pretensions of philosophy, and the medicine of the mind, so much boasted…The empire of philosophy extends over a few, and with regard to these, too, her authority is very weak and limited.’

The early results of IAPT have been better than Hume might have predicted, with recovery rates of 44.4%. IAPT is now being rolled out into child services, into the treatment of chronic physical conditions which have an emotional toll, and into the treatment of unexplained conditions like Chronic Fatigue Syndrome. An IAPT-style programme is also being piloted in Norway.

And finally I consider whether the state has any business providing therapy for our emotions. My position is basically that I’m all for the provision of CBT because it doesn’t try to tell people what ‘flourishing’ or the meaning of life is. But I’m wary of state support for Positive Psychology precisely because it does try to tell people what flourishing ‘is’. In place of Positive Psychology, I’d like to see something else – call it Positive Philosophy – which is more open-ended and Socratic when it comes to discussing the good life.