_DSC1611Therapists do a lot of good. People who have been through therapy are much better off.

But the question of whether psychotherapy can hurt people is not one we see much research on.

A new study by Mike Crawford in England (Crawford, et al., 2016) gives us some very good information about that. He surveyed 14,270 individuals who had been through therapy, and of those, 763 agreed “slightly” or “strongly” that they were worse off after therapy than before. So in this large study, about one person in twenty ends up feeling worse.

Now a 5% deterioration rate isn’t good. It isn’t bad. Surely there are lots of drugs that have more than a 5% adverse effect rate. But still, we want to improve our outcomes, and reducing the adverse outcomes is a good way to do that.

What can be done?

In this study, the most common orientation was cognitive therapy. But a substantial number received solution-focused and psychodynamic therapy. And some people didn’t know what kind of therapy approach their counselor used.

There was no relationship between the type of therapy and adverse outcomes, but there was a significant relationship between not knowing what approach was being used and adverse outcomes. That is, people whose therapists discussed the type of approach they were likely to use seemed to benefit more from therapy. In other words, people benefit from knowing more about how the therapy will be proceeding. Specifically, if a client didn’t understand the type of therapy approach, he or she was 150% more likely to have a negative outcome. Conversely, with a sufficient discussion at the beginning, one has 65% less chance of an outcome that is adverse.

Another finding from Crawford was that ethnic or sexual minorities do worse in therapy. This suggests that possibly more cultural competence training can help, although I am not sure about that. There is actually very little evidence that it does very much good (Lie, et al, 2011, Griner & Smith, 2006). Griner and Smith reported a positive effect size of .45, which isn’t much, but at least I can find no evidence that it is harmful.

What then can we do? Cultural competence seems good. But the most robust answer is to cultivate a culture of feedback. That is, we ask people consistently how they are doing and how they feel about the alliance with the therapist. Lambert (2007) argues that consistently giving a good feedback instrument at the beginning of each session will reliably improve outcomes. Lambert and Burlingame developed the OQ-45, the instrument I started using in 1994, and have used ever since. Mike has shown that you can cut the deterioration effect in half, simply by giving a good outcome measure each session. That is, the outcome measure shows the therapist that things are not going well, and invites her / him to discuss that deterioration with the patient.

I also developed my own alliance measure, and I invite you to download a version of it here. My measure was originally 10 items long, but I have simplified it some and it is down to 7 items. Items 4 and 5 tap into whether the therapy you are offering is right for this particular patient, and items 6 and 7 are the best predictors of good outcomes. I believe that 4 and 5 can directly address the cultural competence aspect of therapy, at least if you do two things.

First, you have to instruct the patient to be rather blunt and honest, since it is in the patient’s best interests for you to know these things.

Second, you discuss items where you do not get a “4” rating. I usually get “4” and “3” ratings, and I make a point of carefully questioning the client. “What would have to happen today so I can raise that rating to a 4?” Do that in the context of “the patient is always right” and being grateful for the courage and commitment it takes for the patient to give that feedback.

I discussed my measure with Mike Lambert, and he thought it was an excellent idea but said that therapists had been quite resistant to that idea, at least back in the early 1990s when he and Gary Burlingame were developing the OQ-45. Maybe the atmosphere has changed and now people are more willing to measure the alliance. Bear in mind that Mike and other researchers have consistently found two things.

Number one, Mike and other researchers have convincingly shown that the model of therapy makes no difference. Cognitive therapy is a good approach, but it is not any better, in some very large studies, than psychoanalytic oriented therapy. (That finding shocks people trained int he last thirty years.)

Number two, the alliance is the best predictor of the factors a therapist can influence in the sessions. So it makes good sense to consistently use an alliance measure. There are several short ones out there. When I first worked out my version in the early 1990s, the measures were too long to use in therapy. But now we have several four – ten item measures that are shown to predict outcome, as mine does.

(Why is my photo on this post? Simply because I can’t think of how to illustrate it, and anyway, it is kind of about me.)


Crawford, M., Thana, L., Farquharson, L., Palmer, L., Hancock, E., Bassett, P., Clarke, J., & Parry, G. (2016). Patient experience of negative effects of psychological treatment: results of a national survey The British Journal of Psychiatry, 208 (3), 260-265 DOI:10.1192/bjp.bp.114.162628  The article is paywalled, but you can get a good review here. 

Griner, D. & Smith, T. B. (2006). Culturally adapted mental health intervention: A meta-analytic review.Psychotherapy: Theory, Research, Practice, Training, 43(4), 531-548.http://dx.doi.org/10.1037/0033-3204.43.4.531

Lambert, M. (2007). What we have learned from a decade of research aimed at improving psychotherapy outcome in routine care. Psychotherapy Research, 17, 1–14. To read this see: http://bit.ly/1SkYrr1

Lie, D.A., Lee-Rey, E, Gomez, A., Bereknyei, S., & Braddock, C. H. (2011). Does Cultural Competency Training of Health Professionals Improve Patient Otucomes? A systematic review and proposed algorithm for future research. Archives of General Internal Medicine, 26(3): 317–325. http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3043186/