Reflections on the debate with David Gillanders about Acceptance and Commitment Therapy at the British Pain Society, Glasgow, September 15, 2017
David Gillanders and I held our debate “ACT: best thing since sliced bread or the Emperor’s new clothes?” at the British Pain Society meeting on Thursday, September 15, 2017 in Glasgow. We will eventually make our slides and a digital recording of the debate available.
I enjoyed hanging out with David Gillanders. He is a great guy who talks the talk, but also walks the walk. He lives ACT as a life philosophy. He was an ACT trainer speaking before a sympathetic audience, many who had been trained by him.
Some reflections from a few days later.
I was surprised how much Acceptance and Commitment Therapy (along with #mindfulness) has taken over UK pain services. A pre-debste poll showed most of the audience came convinced that indeed, ACT was the best thing since sliced bread.
I was confident that my skepticism was firmly rooted in the evidence. I don’t think there is debate about that. David Gillanders agreed that higher quality studies were needed.
But in the end, even I did not convert many, I came away quite pleased with the debate.
Standards for evaluating the evidence for ACT for pain
I recently wrote that ACT may have moved into a post-evidence phase, with its chief proponents switching from citing evidence to making claims about love, suffering, and the meaning of life. Seriously.
Steve Hayes prompted me on Twitter to take a closer look at the most recent evidence for ACT. As reported in an earlier blog, I took a close look. I was not impressed that proponents of ACT are making much progress in developing evidence in any way as strong as their claims. We need a lot less ACT research that doesn’t add any quality evidence despite ACT being promoted enthusiastically as if it does. We need more sobriety from the promoters of ACT, particularly those in academia, like Steve Hayes and Kelly Wilson who know something about how to evaluate evidence. They should not patronize workshop goers with fanciful claims.
David Gillanders talked a lot about the philosophy and values that are expressed in ACT, but he also made claims about its research base, echoing the claims made by Steve Hayes and other prominent ACT promoters.
Standards for evaluating research exist independent of any discussion of ACT
There are standards for interpreting clinical trials and integration of their results in meta analysis that exist independent of the ACT literature. It is not a good idea to challenge these standards in the context of defending ACT against unfavorable evaluations, although that is exactly how Hayes and his colleagues often respond. I will get around to blogging about the most recent example of this.
Atkins PW, Ciarrochi J, Gaudiano BA, Bricker JB, Donald J, Rovner G, Smout M, Livheim F, Lundgren T, Hayes SC. Departing from the essential features of a high quality systematic review of psychotherapy: A response to Öst (2014) and recommendations for improvement. Behaviour Research and Therapy. 2017 May 29.
Within-group (pre-post) differences in outcome. David Gillanders echoed Hayes in using within-group effects sizes to describe the effectiveness of ACT. Results presented in this way are better and may look impressive, but they are exaggerated when compared to results obtained between groups. I am not making that up. Changes within the group of patients who received ACT reflect the specific effects of ACT plus whatever nonspecific factors were operating. That is why we need an appropriate comparison-control group to examine between-group differences, which are always more modest than just looking at the within-group effects.
Compared to what? Most randomized trials of ACT involve a wait list, no-treatment, or ill-described standard care (which often represents no treatment). Such comparisons are methodologically weak, especially when patients and providers know what is going on-called an unblinded trial– and when outcomes are subjective self-report measures.
A clever study in New England Journal of Medicine showed that with such subjective self-report measures, one cannot distinguish between a proven effective inhaled medication for asthma, an inert substance simply inhaled, and sham acupuncture. In contrast, objective measures of breathing clearly distinguish the medication from the comparison-control conditions.
So, it is not an exaggeration to say that most evaluations of ACT are conducted under circumstances that even sham acupuncture or homeopathy would look effective.
Not superior to other treatments. There are no trials comparing ACT to a credible active treatment in which ACT proves superior, either for pain or other clinical problems. So, we are left saying ACT is better than doing nothing, at least in trials where any nonspecific effects are concentrated among the patients receiving ACT.
Rampant investigator bias. A lot of trials of ACT are conducted by researchers having an investment in showing that ACT is effective. That is a conflict of interest. Sometimes it is called investigator allegiance, or a promoter or originator bias.
Regardless, when drugs are being evaluated in a clinical trial, it is recognized that there will be a bias toward the drug favored by the manufacturer conducting the trial. It is increasingly recognized that meta analyses conducted by promoters should also be viewed with extra skepticism. And that trials conducted with researchers having such conflicts of interest should be considered separately to see if they produced exaggerated.
ACT desperately needs randomized trials conducted by researchers who don’t have a dog in the fight, who lack the motivation to torture findings to give positive results when they are simply not present. There’s a strong confirmation bias in current ACT trials, with promoter/researchers embarrassing themselves in their maneuvers to show strong, positive effects when their only weak or null findings available. I have documented [ 1, 2 ] how this trend started with Steve Hayes dropping two patients from his study of effects of brief ACT on re-hospitalization of inpatients with Patricia Bach. One patient had died by suicide and another was in jail and so they couldn’t be rehospitalized and were drop from the analyses. The deed could only be noticed by comparing the published paper with Patricia Bach’s dissertation. It allowed an otherwise nonsignificant finding a small trial significant.
Trials that are too small to matter. A lot of ACT trials have too few patients to produce a reliable, generalizable effect size. Lots of us in situations far removed from ACT trials have shown justification for the rule of thumb that we should consider effect sizes from trials having less than 35 patients per treatment of comparison cell. Even this standard is quite liberal. Even if a moderate effect would be significantly larger trial, there is less than a 50% probability it be detected the trial this small. To be significant with such a small sample size, differences between treatments have to be large, and there probably either due to chance or something dodgy that the investigators did.
Many claims for the effectiveness of ACT for particular clinical problems come from trials too small to generate a reliable effect sizes. I invite readers to undertake the simple exercise of looking at the sample sizes in a study cited has support of the effectiveness of ACT. If you exclude such small studies, there is not much research left to talk about.
Too much flexibility in what researchers report in publications. Many trials of ACT involve researchers administering a whole battery of outcome measures and then emphasizing those that make ACT look best and either downplaying or not mentioning further the rest. Similarly, many trials of ACT deemphasize whether the time X treatment interaction is significant in and simply ignore it if it is not all focus on the within-group differences. I know, we’re getting a big tactical here. But this is another way of saying is that many trials of ACT gives researchers too much latitude in choosing what variables to report and what statistics are used to evaluate them.
Under similar circumstances, showed that listening to the Beatles song When I’m 64 left undergraduates 18 months younger than when they listen to the song Karamba. Of course, the researchers knew damn well that the Beatles song didn’t have this effect, but they indicated they were doing what lots of investigators due to get significant results, what they call p-hacking.
Many randomized trials of ACT are conducted with the same researcher flexibility that would allow a demonstration that listening to a Beatles song drops the age of undergraduates 18 months.
Many of the problems with ACT research could be avoided if researchers were required to publish ahead of time their primary outcome variables and plans for analyzing them. Such preregistration is increasingly recognized as best research practices, including by NIMH. There is no excuse not to do it.
My take away message?
ACT gurus have been able to dodge the need to develop quality data to support their claims that their treatment is effective (and their sometime claim it is more effective than other approaches). A number of them are university-based academics and have ample resources to develop better quality evidence.
Workshop and weekend retreat attendees are convinced that ACT works on the strength of experiential learning and a lot of theoretical mumbo jumbo.
But the ACT promoters also make a lot of dodgy claims that there is strong evidence that the specific ingredients of ACT, techniques and values, account for the power of ACT. But some of the ACT gurus, Steve Hayes and Kelly Wilson at least, are academics and should limit their claims of being ‘evidence-based” to what is supported by strong, quality evidence. They don’t. I think they are being irresponsible in throwing in “evidence-based’ with all the
What should I do as an evidence-based skeptic wanting to improve the conversation about ACT?
Earlier in my career, I spent six years in live supervision in some world-renowned therapists behind the one-way mirror including John Weakland, Paul Watzlawick, and Dick Fisch. I gave workshops world wide on how to do brief strategic therapies with individuals, couples, and families. I chose not to continue because (1) I didn’t like the pressure for drama and exciting interventions when I interviewed patients in front of large groups; (2) Even when there was a logic and appearance of effectiveness to what I did, I didn’t believe it could be manualized; and (3) My group didn’t have the resources to conduct proper outcome studies.
But I got it that workshop attendees like drama, exciting interventions, and emotional experiences. They go to trainings expecting to be entertained, as much as informed. I don’t think I can change that.
Many therapists have not had the training to evaluate claims about research, even if they accept that being backed by research findings is important. They depend on presenters to tell them about research and tend to trust what they say. Even therapist to know something about research, tennis and critical judgment when caught up in emotionality provided by some training experiences. Experiential learning can be powerful, even when it is used to promote interventions that are not supported by evidence.
I can’t change the training of therapists nor the culture of workshops and training experience. But I can reach out to therapist who want to develop skills to evaluate research for themselves. I think some of the things that point out in this blog post are quite teachable as things to look for.
I hope I can connect with therapists who want to become citizen scientists who are skeptical about what they hear and want to become equipped to think for themselves and look for effective resources when they don’t know how to interpret claims.
This is certainly not all therapists and may only be a minority. But such opinion leaders can be champions for the others in facilitating intelligent discussions of research concerning the effectiveness of psychotherapies. And they can prepare their colleagues to appreciate that most change in psychotherapy is not as dramatic or immediate as seen in therapy workshops.
I will soon be offering e-books providing skeptical looks at positive psychology and mindfulness, as well as scientific writing courses on the web as I have been doing face-to-face for almost a decade.
Sign up at my website to get advance notice of the forthcoming e-books and web courses, as well as upcoming blog posts at this and other blog sites. Get advance notice of forthcoming e-books and web courses. Lots to see at CoyneoftheRealm.com.