So, How Effective is Cognitive analytic therapy?

In this blog, Stephen Kellett and Mel Simmonds-Buckley describe their recent (2021) meta-analysis of research into cognitive analytic therapy.

This was an attempt to review all of the outcome studies completed to date looking at clinical outcomes for CAT.  In order to do this, we looked at CAT’s effectiveness in routine practice, how it compared to other treatments and therapies, and how acceptable it is to people using it as a therapy.

Both those who go into psychotherapy, and NHS commissioners, need to know that psychotherapies like cognitive analytic therapy (CAT) are acceptable, effective and efficacious. Otherwise they may decide not to use them or to fund them, so that they are available in the NHS.

  • Acceptable means that clients turn up for the therapy and stay in treatment until it’s completed.
  • Effective means that clients’ symptoms reduce over treatment time. 
  • Efficacy means that the therapy improves a client’s mental health compared to doing something else, such as another psychological treatment or medication.  These are often the ‘control’ groups in randomised controlled trials. Researchers allocate participants either to CAT or another approach, and then see what is most useful.  

What we wanted to achieve with this project was a ‘meta-analysis’ of the CAT research studies that have looked at effectiveness and efficacy. This was in  order to arrive at some summary statements about CAT and whether it works.  

What we did

Carrying out this meta-analysis meant finding all the relevant studies as a first step. We did this using what is called a systematic review. Then we rated the quality of these studies according to a set of standards or criteria. Next we created a spreadsheet of outcomes and also the features of each study. 

Once we had all the information in a spreadsheet, we did a narrative synthesis of the features of the studies. We also applied statistical techniques to combine the findings from all the CAT studies. This produced a more accurate estimate of what is likely to be the ‘true’ effect of CAT.

We combine the studies by ‘weighting’ the findings in each of them. For example we can have more confidence in the results of larger studies on the impact of CAT. Or more precise studies can show where CAT had more influence on the overall treatment effect.

We could not combine the studies in this way when looking at the evidence on acceptability. Instead we worked out the average ‘dropout’ rate for CAT. This took into account the proportion of clients who left therapy before completing it.

We wanted to explore the impact of CAT on three areas:

  • depression (i.e. how down or low in mood people feel)
  • global functioning (i.e. how people are psychologically getting on generally in their lives), and
  • interpersonal problems (i.e. how well people’s relationships are going) 

So, what did we find?  

There were twenty-eight studies and their quality was okay.  Ten of these studies (36%) were randomised control trials, where CAT was compared to something else. The remaining eighteen (64%) tracked changes over time. However in these studies there was no comparison with another treatment.  

In the meta-analysis we could use twenty-five studies. (These studies reported outcomes in a way we could cut and paste into the spreadsheet.) 

The studies we looked at showed that CAT is offered to people with complex problems such as long-standing and complex trauma . Therapists tended to deliver it on a one-to-one basis in public health settings, typically in the NHS.  There were however, good CAT studies from other countries, such as Australia.


At the end of treatment people who had CAT tend to have significantly reduced depression, better interpersonal functioning and reduced global distress.  The effect was larger for global distress and reduced depression than the improvement in interpersonal functioning.  It didn’t make a difference whether the clients were male or female, or how old they were. It did make a positive difference when CAT was provided by a qualified CAT therapist (rather than for example a trainee CAT therapist or trainee clinical psychologist).

In the studies that followed clients up after therapy finished, clients had held onto the gains that they had made during CAT.  The follow up periods were normally at around 6-months after the end of therapy.  At follow-up, there was also a significant improvement in clients’ interpersonal difficulties.  This means that the effectiveness of CAT is long-standing and not a ‘flash in the pan’ treatment.  The interpersonal focus of CAT seems to do its work over time.  


In studies where CAT was compared to other approaches, CAT was more effective than these other approaches. The comparisons included ‘treatment as usual’ in a service, being on a waiting list or having another psychotherapy

This shows that CAT has treatment efficacy.


The dropout rate for CAT was 16%. This compares very well against other psychotherapies.  CAT therapists often remark on how people seem to stay in CAT therapy, and that the model really engages clients.    We need to be careful not to assume that this lower dropout rate will apply in all situations, as what we looked at was just the study average, The best way to work this out more accurately would be to do another meta-analysis.    We are on this!

You can find the original study here:

Stephen Kellett & Mel Simmonds-Buckley 

Dr Stephen Kellett is a Consultant Clinical Psychologist & CAT Psychotherapist. He is IAPT Programme Director at the University of Sheffield’s Department of Psychology, an active member of the PEARLS Research Lab, and contributes to ACAT-accredited CAT Practitioner training.

Dr Mel Simmonds-Buckley is a Post-Doctoral Research Associate in the Clinical Psychology Unit at the University of Sheffield and a member of the PEARLS Research Lab.

If you would like to see a listing of recent research into CAT, visit our page on the CAT Evidence Base.

So, How Effective is Cognitive Analytic Therapy? by Dr Stephen Kellett & Dr Mel Simmonds-Buckley is licensed under CC BY-SA 4.0