• icon-cloud
  • icon-facebook
  • icon-feed
  • icon-feed
  • icon-feed

Psychotherapy (and psychotherapist) outcomes are good but largely stagnant

I have been asked to write a chapter on the importance of obtaining regular feedback on client progress in a book on psychotherapist self-practice & self-reflection.  This initial section (see below) of a draft of the chapter comments on the current state of psychotherapy itself:

(Note the ideas in this blog are explored in more detail in the chapter "Client feedback: an essential input to therapist reflection" in the forthcoming Haarhoff, B. and Thwaites, R. (2016) "Reflection in CBT: Increasing your effectiveness as a therapist, supervisor and trainer." London: SAGE Publications Ltd.)

Psychotherapy outcomes are pretty good, but largely stagnant:  Psychotherapy is effective.  When treating people suffering from psychological disorders, psychotherapy achieves similar results to psychiatric drugs but typically with outcomes that are better maintained (De Maat, Dekker et al. 2006; Cuijpers, Sijbrandij et al. 2013; Huhn, Tardy et al. 2014).  To put this into a broader context, psychiatric drugs themselves achieve results that are similar to outcomes produced by pharmacotherapy for general medical conditions – a major comparison of key meta-analyses found average effect sizes of 0.49 for psychiatric medications and 0.45 for general medical medications (Leucht, Hierl et al. 2012; Moran 2012).  When contrasting diverse diseases, one should treat simple effect size comparisons with considerable caution.  The overall picture provided by these reviews is however still useful – the benefits achieved through psychotherapy are worthwhile and roughly comparable to treatment outcomes achieved across the whole field of medicine.  In fact, hundreds of psychotherapy meta-analyses have shown effect sizes of approximately 0.6 (0.4 to 0.8) (Lambert 2013).  In the social sciences this is considered a ‘medium’ to ‘strong’ result (for Cohen’s d).  And increasingly, economic analyses are demonstrating how cost-effective these psychotherapy outcomes can be (Smith and Williams 2013).

Robust results, encouraging reductions of relapse, cost effectiveness – so what’s not to like about these benefits of psychotherapy?  Well since Smith, Glass & Miller’s famous early estimate of an 0.85 effect size from a meta-analysis of 475 psychotherapy studies (Smith, Glass et al. 1980), research has failed to show any general improvement in psychotherapy outcomes over the subsequent 30 plus years (Lambert 2013).  When one considers the vast amount of effort that has been put into trying to make therapy more effective, this lack of improvement seems hard to credit (although interestingly one could make similar pessimistic claims for all the research money poured into trying to find more effective antidepressant medications).  An example of this very disappointing lack of progress in psychotherapy is illustrated by Lars-Goran Ost’s 2008 paper – he analysed the outcomes reported in 364 research studies of CBT treatment for anxiety disorders published over the last 40 years (Ost 2008).  He concluded “The results showed that in most instances there were no significant change in ES (effect size) across time”.  In fact, when we look at the psychotherapy field in general, improved statistical methods suggest we have been over-estimating our effectiveness (Cuijpers, van Straten et al. 2009; Lambert 2013) and far from seeing progress, our results are actually worse than we were claiming three or four decades ago.

The same is true for results obtained by individual therapists, where training & years of experience seem to add little to effectiveness:  Depressingly, psychotherapy’s lack of progress in producing better outcomes over the decades is mirrored by similar findings for individual psychotherapists themselves.  Research study after research study has shown that most therapists do not seem to get better results as they gain increasing professional experience.  Outcomes achieved are typically unrelated to the therapist’s qualifications, training, type of therapy, age and gender (Baldwin and Imel 2013).  Possibly experience with particular types of client difficulty may boost the helpfulness of therapists working in specific domains or with specific problems (Kraus, Castonguay et al. 2011), but even this common sense finding waits for adequate replication.  At least the position of psychotherapy may not be as dire as in medicine more generally, where a systematic review of 62 studies, examining the link between physician experience and patient outcome, found that in 73% of cases increasing experience was associated with decreasing performance for some (21%) or all outcomes (52%) that were assessed (Choudhry, Fletcher et al. 2005).  The review found that there seems to be an inverse relationship between years of experience and the quality of care that is provided – and in some studies an inverse relationship with resultant death rates.  It’s easy to reassure ourselves as psychotherapists that at least with talking therapies we are not liable to produce such toxic damage if we are not working optimally.  This reassurance is hollow though and only defensible if we are unaware of the increased disability and mortality associated with persisting psychological distress.  There is so much research demonstrating this (Russ, Stamatakis et al. 2012; Collaborators 2013; Whiteford, Degenhardt et al. 2013; Saint Onge, Krueger et al. 2014).  And better management of mental disorders can change subsequent risk – see, for example, a recent study on improved management of depression in older adults associated with a 24% decrease in death rates over the 8 years of follow-up (Gallo, Morales et al. 2013).  It is likely that psychotherapists are ignorant of the mortality increases associated with poor treatment largely because – in contrast to poor surgery – patients do not typically die in front of them.

In the next posts in this sequence, I will discuss the encouraging finding that actually psychotherapists themselves vary very considerably in the outcomes they achieve ... and that feedback about this can help us improve our effectiveness.


Share this