Interview: Talking therapies

What led you to clinical psychology?
It was probably my mother, who was profoundly deaf and had to raise me on her own from the age of two because my dad died of leukaemia. We were quite poor and it was a difficult situation, but she always seemed to be the person that other people came to with their problems. That got me interested in trying to help people. At school I really enjoyed chemistry and I thought drug treatments were probably going to be the big advance so I went up to Oxford to study chemistry, but by the time I got to Oxford I was much more interested in psychology as a more precise type of intervention for people with mental health problems.

Who were the main influences on you at Oxford?
My tutor, Alan Cowey, was a physiological psychologist who worked on perception; he was a great experimentalist and had an enormous influence on me. As did Dick Passingham – who later became a very well-known imaging person in mental health –  and Anne Treisman and Pat Rabbitt. The course was very experimental, but it was also good at conveying the idea that the way we perceive things can be quite different from what appears to be objective reality and that beliefs influence perceptions. That is such a central idea of cognitive therapy; the fact that that could be analysed so systematically was really fascinating to me.

What are the origins of your interest in cognitive therapy?
I was lucky to come to the Institute of Psychiatry and had Jack Rachman as one of my main tutors, along with Dave Hemsley. My training was a behavioural one, but Jack always encouraged us to look at those people who weren't responding to current treatments, and it became clear to me that some of the limitations of our behavioural approaches seemed to do with our not tackling beliefs. There was a patient on the unit who wasn't responding to a behavioural programme for her severe phobias. I reviewed her treatment and thought that the exposure therapy hadn't been done optimally and that we should try a more intensive flooding experience, so Dave Hemsley and I spent all day with her in the hope that she would habituate in a long session – but she didn't! It was extraordinary for me to see someone being anxious for so long without any habituation and I started to think 'how could this be?' I noticed that she was breathing very quickly during her panic attacks and so in a break I asked one of the psychiatrists 'What impact would that have?' He said 'That's hyperventilation – why don't you try it?' I thought it was quite pleasant but it had big physiological effects. I got the idea that maybe the hyperventilation was maintaining the anxiety, but it couldn't be just that because while I thought it was quite pleasant she thought it was dreadful, so it must be something about the way we were thinking about these sensations. That was what led me to the panic model, and the panic model got me involved with Tim Beck. I met him at a conference and went out to Philadelphia to learn about cognitive therapy.

When the Royal Institution held a debate in 2007 you promoted Beck as the greatest mind who has changed other minds. Why do you rank him as highly as you obviously do?
He has enormous strengths. He lives cognitive therapy in his life in the way that he does with his patients, and so after even only a short conversation with him you get a clear idea of cognitive therapy. What I found so fascinating about his approach was a combination of scientific rigour – you have to test out your ideas with experiments and trials – and the notion that your own views aren't so important; the first step has to be to get into the patient's head and see the world from their perspective. It's only when you've done that that you have a chance of working out a way to help them move to a more fulfilling perspective. Other people's heads are just always the most fascinating place to be. Everyone is different, and he taught always to start with the patient's own perceptions and move from there – but do it systematically and scientifically.

CBT is sometimes caricatured as the rote application of technique, whereas your emphasis seems very client-focused.
It's partly because there is a strong emphasis on writing manuals in order to ensure that therapies are well delivered and widely disseminated, and these sometimes read like cookbooks. Over the years there has been an unfortunate trend for people to be a bit lazy in the way they publish them – publishing manuals developed for a clinical trial without thinking about how the approach needs to be adapted to fit with the full range of people that one sees in normal practice. It's also because people haven't had the chance to observe the masters of these therapies. I was lucky to be able to watch Beck, and it was very noticeable that he would drop a technique very quickly if he felt it wasn't having any impact – even if it seemed to be the obvious thing to do given the model and the manual. When you asked him about it he'd say, 'Well what's the point in persisting with something that doesn't work?' He has a mental road map of the problematic beliefs that are linked to affect and he's using that to drive his selection of what he does in the session. So what he does might seem a bit surprising to someone who has just read the manual, but it's because he is following a clear map – and the thing is that it's an affective map.

One dangerously seductive aspect of CBT is that it looks like an assembly of clever techniques; what you're emphasising is the importance of holding on to a model of the mind out of which techniques emerge.
Almost anyone who is severely depressed, and many people who are anxious, will have a lot of negative thoughts. The temptation of a novice is to jump in on a thought because they think 'I know how to challenge that' without really checking out 'is that thought driving the affect?' What you see Beck doing is standing back for a while; it looks as though he is asking a series of random questions, moving around the houses. But he's not just looking for negative thoughts; he's looking for those where the affect is. So you don't quite know where he's going and then you suddenly see the person's facial expression change. They mention a belief, he says it back to them and suddenly there's a tear in their eye or their face goes very tense. And then you can start the stopwatch – the whole apparatus of cognitive therapy can come out at that point.

CBT practitioners are sometimes seen as engaging in a sort of triumphalism which puts to one side the accumulated clinical wisdom of other approaches. Do you have any thoughts about that?
Within the CBT movement there has always been a feeling of dissatisfaction; what drives it is not the idea that CBT is effective for everyone, but the absolute fact that it's not. We don't have a single disorder in which it works for everyone, and so it's a fascinating area for people to work in because they feel there are still challenges and there is a method which helps them pursue them. One of the criticisms of the Improving Access to Psychological Therapies (IAPT) programme is to say 'CBT is not a panacea', as though CBT researchers have been saying it is. Most of the people who are really involved in treatment development obsess about those patients who didn't get well – that's what you're interested in because that's the next advance.

One thing which isn't highlighted enough is that quite a lot of the innovation in cognitive therapy actually comes from observing the enormous advances made in other therapies and then incorporating the most appealing aspects into the cognitive therapy intervention. People often say that our work in anxiety disorders has a lot of gestalt influence on it, or that it takes quite a lot from hypnosis, and they are right – there are techniques in both areas which have proved very helpful and so become incorporated in the programme. But in doing so they are all deployed in order to hit the target of the cognitive model. Beck is very explicit about this when you talk to him. He trained as an analyst and many of the things he learned about the therapeutic relationship from within that perspective are central to cognitive therapy. I see cognitive therapy as a bit like a magpie looking around at other therapies and when it spots some of their shining jewels it hops over and borrows them!

Therapists of other persuasions often characterise CBT as neglecting the therapeutic relationship, in contrast to the emphasis they place on this as a mutative factor.
Maybe CBT researchers themselves are partly to blame for that perception because they talk more about the particular techniques that are being developed rather than the core of CBT, which is collaborative empiricism and Socratic questioning. Those are actually the inalienable core of the therapy and those are the things which tend to produce a good therapeutic alliance.

One of the things that I emphasise when teaching people how to treat anxiety disorders is that you do need to change your own interpersonal behaviour in therapy sessions as a function of the different anxiety problems you are treating. An obvious contrast is social phobia with PTSD. In PTSD when people go over their traumas they are reactivating all of their beliefs and experiences from the traumatic event, and so they are very prone to see themselves as being victimised again. So the therapist has to be very explicit in being warm, empathic and non-judgemental and creating a safe environment for the person. That has to be done very explicitly because when you activate the trauma memory it produces a completely different schema, which has to be counteracted in therapy. But if you are talking to a social phobic and you are very warm and empathic and have lots of eye contact, it's the exact opposite of what they want; it will make them feel much more self-conscious and the therapy doesn't progress well. So therapists need to be a little bit more distant in order to get the therapy progressing well, whereas in PTSD it's the absolute opposite.

I also think that there are challenges to our ideas about the alliance that come from novel ways of delivering therapies. If you look at the success of computerised CBT it makes you wonder, because computer programs are not very strong on warmth and empathy. What they are strong on is that they are very unlikely to be misperceived. When people come into therapy with their own traumatic history they are very vulnerable to interpreting what we do in a way which makes them feel demeaned, or victimised or humiliated. It's very easy for something that the therapist does to be misinterpreted in a way which is seen by the patient as critical or demeaning – it's very difficult for patients to see a computer program as being that. Maybe the defining feature of a good therapist is the ability to behave in a way which isn't open to misinterpretation, especially by individuals whose history makes them prone to do just that.

People sometimes react to research evidence by saying 'well that's very impressive but the patients in these trials are highly selected; they're not like the clients that most people see in the NHS'. Do you have any thoughts about that?
These are issues which have to be addressed in a scientific way. It is true that, historically, randomised trials have selected the patients they see to produce a homogeneous group. But whether the results in those trials then generalise to less selected populations is an empirical question; that's the important thing to recognise, rather than say 'Oh, because this isn't necessarily an identical population to the people I see in my clinic then it's bound not to work'. In the outside world people think you are excluding the more severe cases, but the most common exclusion in the trials I know about is the milder ones.

For me, the most important opportunity to look at generalisation was the work we did in Northern Ireland following the Omagh bomb. We were asked to train up clinicians, who weren't originally CBT therapists, to offer cognitive therapy to everyone who had PTSD and wanted treatment. So a completely unrestricted sample, and the really wonderful therapists in Northern Ireland got as good results as we did in our randomised trials with more restricted populations. We didn't know that's what would happen, but it did.

While there's been a lot of progress in relation to anxiety disorders, current approaches to depression show more modest efficacy. Do you have a sense of where the field is heading?
Beckian cognitive therapy for depression has not changed much, and that's a problem because only 50–60 per cent of people will recover. After the NIMH trial in 1985 (where CBT didn't fare as well as many people expected it to, especially for people who were more seriously depressed) a lot of time and effort was devoted to checking out whether its results were a quirk, because the outcomes were somewhat discrepant from other trials for cognitive therapy. From Steve Hollon and Rob DeRubeis' work we now know that cognitive therapy does pretty well in severe depression, but the distraction held up the field for a long time. But we're over that now, and there are lots of talented people doing new things which will take things forward. For example, Ed Watkins' work on rumination focused cognitive therapy; Adrian Wells' metacognitive work; Chris Brewin's focus on intrusive memories from the past. And of course behavioural work has come back in a much more systematic way, and with a new slant, through behavioural activation. So we have at least four rather promising ways forward. What they are doing now is similar to work on anxiety disorders in the last 15 years: it's focusing on particular features which we aren't good at changing in the existing treatment and asking what we can do to develop something which has a better traction on that phenomenon.

Where is your own research heading over the next five years?
One thing we are trying to do is make treatments more broadly available. We are working on an internet version of our social phobia treatment, which we hope will be as effective as the therapist-delivered treatment. If that is the case we will be aiming to make it available free – computerised CBT can greatly increase availability of treatments, but if it's linked to very restrictive licences then it rather undermines its point.

Our work on PTSD has been aimed at making treatments more appealing, and an exciting development is a one-week treatment which concentrates the intervention. That's quite radical because people thought you couldn't deal with such severe and disabling traumas in such a short period of time, but our latest trials suggest that one-week treatment does as well as treatment spread over three months.

Another study is looking at a version of cognitive therapy where people work on self-study modules between sessions. We were able to get twice as much improvement per hour of therapy than we had previously done, so we could get 80 per cent of people recovered with seven sessions rather than fourteen. But some people didn't improve, and we have a 'post mortem', trying to work out if there is some phenomenon that we didn't get good traction on and what can we develop to improve that.

That is an aspect of the work that people don't always spot – a focus on failure rather than success.
That's why the monitoring system in IAPT is so important, because it allows us to compare traditional monitoring systems (where you only collect data before therapy starts and after it finishes) with session-by-session monitoring. What you find is that the people who have missing data using the traditional system tend to be those who have done less well. That's a problem for two reasons: it means you overestimate how good your service is, but also the people who do less well are the people you want to focus on in order to improve things for the future. If you don't get to spot them, then the field dies; there is no possibility of advance.

But I?would guess focusing on treatment failure in the field can be problematic, if therapists experience this as a threat?
Yes, there is a danger that in an NHS with targets outcome monitoring will be seen as something critical rather than a treasure house for new ideas. It's up to people involved in managing services to cherish their workforce and use the data in an inquisitive way to get people to reflect and improve their practice.

Why is IAPT such an important programme from your point of view?
In the last 20 years there have been enormous advances in developing effective psychological treatments for anxiety disorders and depression but it has been a persistent sadness that the majority of people don't get these treatments. IAPT tries to change that by vastly increasing access to the treatments and so it's something that I passionately support.

I also think that we are at a crucial moment in history where several important themes are coming together. Society is increasingly valuing subjective reality. Political leaders throughout the Western world are making speeches to the effect that they are no longer judged simply on financial success, but also in terms of whether the population feels that it's having a satisfying and happy life. That creates a great opportunity for psychology; it means that good experimental work and good science by psychologists can influence policyin a way that hasn't been possible before.

The investment that the government is making in IAPT is so substantial – it's very important that we make the best of it. If we don't, people might be disillusioned about the value of psychological therapies and reluctant to do anything further in the future to improve public access to it. So it's very important that we do it well and we do it in a way that is responsive to what the public wants.

When he announced IAPT Alan Johnson said that it would be judged by how many people recover, not just what the waitlists are; it's a welcome change for the NHS to be talking about people getting better rather than how long they wait for something. But obviously this is a sharp sword to be judged by and we need to step up to the plate.

And as you know, some people are very opposed to IAPT and see it as a potentially destructive programme. How you understand that?
Some criticism of IAPT has been very understandable, because in order to make progress much of the initial emphasis has been on developing training in CBT; because it's not focusing on other approaches, people can feel that this carries a longer-term message. But there are good reasons for this emphasis. Although other evidence-based treatments are recommended for depression by NICE, they are not for anxiety disorders. As CBT is a treatment that covers the full range of cases, there's clearly a need to start there. But as IAPT goes forward it needs to bring in other sorts of therapist, and there are plans to do that – for example IPT, which is as strongly advocated by NICE in depression as CBT is. And it's good that it's going that way.

When you develop new services and new training programmes there are important questions about how they intersect with existing systems, and these need to be worked through carefully. There aren't straightforward answers and the psychological community has to come together to try to solve the problems. We need coherent care pathways that not only deal with people who are going into IAPT but also with those people who are not – and we need a seamless system. We don't want to create IAPT as a sort of ghetto, separate from the rest of the mental health services. There is also a tension between commissioners on the one side and clinicians on the other, and commissioners for understandable reasons often want to try and do things more cheaply. As clinicians we are rightly concerned that this might degrade quality and there needs to be an ongoing debate to try and get to good value without reducing this.

I think all the discussions IAPT is producing should result in a more level playing field for psychotherapy in general. Its emphasis means that for those therapies that haven't enjoyed the opportunity to be systematically examined in controlled trials, routine monitoring will automatically constitute pilot work and create a strong argument for approaching funders and getting support for randomised trials. The creation of this monitoring system will inevitably mean that the range of therapies recommended by NICE in five, ten years' time on the basis of good evidence will be much wider than now and much less CBT dominated – and that's a really positive thing.