Trial By Error: Yet Another Meta-Analysis Purporting to Prove that CBT Is Effective for “Chronic Fatigue Syndrome”

By David Tuller, DrPH

I just wrote about how PACE was favorably cited in an article in Nature Reviews Cardiology. Last month, that piece of crap was also included as part of yet another meta-analysis that mushed together the findings from a load of bad papers and concluded that, collectively, they prove something or other. This new paper—“Efficacy and Acceptance of Cognitive Behavioral Therapy in Adults with Chronic Fatigue Syndrome: A Meta‐analysis”–was published by the International Journal of Behavioral Medicine and written by investigators from the clinical psychology department at the University of Wuppertal in Wuppertal, Germany.

Was the situation really calling out for yet another major review of trials of psycho-behavioral treatments for what the investigators insist on calling CFS? According to this crew, yes.

The strategy here appears to consist of throwing review after review at the issue, with each new iteration making claims that CBT and/or GET are effective, and arguing soberly that these new findings add further weight to the body of evidence—without pointing out that every review includes some grouping of the same universe of stupid and inadequate studies. When you’re a kid, you learn pretty quickly that if you add a pile of mud to another pile of mud, you just get a bigger mud pile. I guess some people skip that lesson.

In this case, the investigators accept without question the validity of findings from unblinded studies relying on subjective outcomes—the standard design for research in this domain. Unfortunately, this study design is a recipe for generating unknown amounts of bias. Of course, combining biased studies with other biased studies does nothing to eliminate the bias. For unexplained reasons, the meta-analysis ignored null or clinically insignificant results from objective measures included in the body of research—such as, in the PACE trial, the six-minute walking test, step-test for fitness, employment status, and social welfare status. Overlooking or ignoring outright objective outcomes that contradict subjective ones is unacceptable in scientific reporting.

Most of all, the new meta-analysis seems explicitly designed to counter the 2021 ME/CFS guidelines issued by the UK’s National Institute for Health and Care Excellence. The guidelines repudiated the use of CBT as a curative treatment while allowing for it as supportive care. In developing the guideline,s NICE downgraded studies in which not enough patients were shown to have experienced post-exertional malaise (PEM), which is now considered a core characteristic of the illness. Those studies were generally based on the flawed 1994 Fukuda criteria from the US Centers for Disease Control, or on the even worse 1991 so-called “Oxford criteria.”.

The Oxford criteria required only six months of unexplained fatigue. Fukuda required the fatigue plus four of eight additional symptoms; PEM was one of the eight, so it was not necessary for patients to receive a diagnosis. The studies using these definitions are the ones generally favored by members of the CBT/GET ideological brigades. But in the world outside that bubble, these case definitions were considered to have been rendered obsolete by more recent ones that require the presence of some form of PEM.

It is therefore not surprising that most of the 15 studies included in the meta-analysis identified participants based on either the Fukuda or Oxford criteria, and they were mostly from the UK and the Netherlands—two hotbeds of the CBT/GET propaganda machine. The results of the meta-analysis were what would be expected: “The results indicate that CBT for CFS is effective in reducing fatigue, fatigue related impairment, and severity of depression and anxiety.” (Given the overall meaningless of the findings, there’s no point in bothering with the actual numbers presented. Life’s too short.)

Of course, the authors recommended further research to “focus on examining the maintenance of effects”–as if the modest reported effects weren’t just artifacts of bias. And patients themselves were presented as obstacles to the possible usefulness of the intervention–not limitations in the therapy. As the meta-analysis authors wrote: “Patients that suffer from CFS often have their own explanations about the causes and nature of their illness. When these subjective explanatory models are primarily physical and not related to psychological processes…they may be hesitant to engage in a therapy that emphasizes these aspects.”

According to this narrative, patients believe stupid, deluded things about the nature of their illness, and so they reject the sound advice of psychologists like the authors. This sort of patient-blaming is endemic in this domain of science. The meta-analysis is essentially useless as a piece of science, but it serves as yet more evidence of the scientific cluelessness of some of the investigators in the field.