Trial By Error: GET/CBT Ideologues Revive 1991 Oxford Criteria as Core Definition for Long Covid Research

By David Tuller, DrPH

The Collaborative on Fatigue Following Infection, or COFFI, was formed in 2015 to promote the theories and treatment approaches embodied in the now-discredited and arguably fraudulent PACE trial and related research. In a nutshell, PACE and related research promoted the notion that the symptoms of patients with the clinical entity or entities variously called chronic fatigue syndrome, myalgic encephalomyelitis, ME/CFS, CFS/ME and other names could all be attributed to the effects of being physically deconditioned and of being afflicted with “unhelpful beliefs” about having an organic illness.

The recommended treatments were graded exercise therapy (GET) to get patients “reconditioned” and/or a specific form of cognitive behavior therapy (CBT) designed to alleviate them of their purportedly unhelpful beliefs. Since then, however, there has been a world-wide revulsion against PACE’s “unacceptable methodological lapses” (per 100+ international experts who signed Virology Blog’s 2018 letter to The Lancet), and both the US Centers for Disease Control and the UK’s National Center for Health and Care Excellence have flatly rejected its findings. These and related developments have dealt some serious blows to the credibility of the GET/CBT strategy.

COFFI has been pushing back. In a paper published last June called “A research agenda for post-COVID-19 fatigue” published in the Journal of Psychosomatic Research, members of the group referenced Professor Michael Sharpe’s 1991 case definition for chronic fatigue syndrome, which was abandoned long ago by most investigators in this field. That definition included a “sub-type” called “post-infectious fatigue syndrome.” Adopting that subtype (and using “infective” rather than “infectious”) the COFFI members have resurrected the Oxford criteria from the dead and essentially rebranded them. This appears to be a sly way of ignoring and seeking to maneuver around the fact that this approach has already been rejected in the ME/CFS domain.** [This paragraph has been corrected. See below.]

This effort isn’t all that surprising. For the GET/CBT ideological brigades, with the COFFI crew foremost among their ranks, the pandemic has served as an enormous opportunity to promote their hypotheses about psychogenic causation of the symptoms collectively being called long Covid. A key construct is pandemic-related “psychosocial strain”–a buzz phrase that has been highlighted by both Harvard physician Adam Gaffney and New Republic journalist Natalie Shure. The suggestion is that many among those counted as having long Covid most likely weren’t even infected with the coronavirus in the first place but are experiencing a form of post-traumatic stress disorder.

Of course, no one seriously disputes that psychosocial strain—whether experienced as anxiety, depression, stress, PTSD, or whatever–has physiological correlates that negatively impact health and can lead to somatic symptoms. No one seriously disputes that some people who never had covid-19 might also sometimes report troubling symptoms. No one seriously disputes that psychological counseling can be helpful for anyone going through a difficult and challenging period in life, and people living with chronic illnesses can have great difficulties and challenges.

But so what? The relevant question is whether most of the millions of formerly hard-working and productive individuals who now find themselves severely disabled but have no identified organ damage are mostly suffering from the effects of psychosocial strain or whether post-infectious pathophysiological processes are implicated. Whatever Professor Sharpe and others maintain, the evidence points to the latter. Something is really going on with people even though the results on standard tests tend to be normal; investigators continue to learn more all the time about the protean impact of coronavirus infection.

(Recent articles on the pathophysiology of long Covid can be read here, here, and here.)

**********
Reviving an ancient definition for current usage

The COFFI research agenda paper regurgitating and rebranding ancient ideas was published in the Journal of Psychosomatic Research (JPR). As the official journal of the European Association of Psychosomatic Medicine , the JPR has essentially served as an in-house promotional organ for investigators pushing the GET/CBT treatment paradigm for ME/CFS. (One of the lead PACE investigators, Professor Michael Sharpe, is currently president of the association.)

The journal has published some real whoppers, like Professor Peter White’s long-term follow-up to his trial of self-managed GET. The paper failed to mention in its highlights section that there were no—zero—differences between the intervention and comparison groups at long term follow-up. I brought this omission to the journal’s attention. The article was subsequently corrected.

To accomplish its goals of transforming an early but outdated CFS definition into the core long Covid construct, the COFFI research agenda paper offered this definition of post-infective fatigue syndrome: “Post-infective fatigue syndrome (PIFS) is defined as persistent, severe fatigue after an infection that cannot be explained by other medical or psychiatric conditions, which has been present for at least six months and significantly affects daily functioning.”

The reference for this definition is the 1991 paper in which Professor Sharpe–the Norma Desmond of this scientific domain (“I am big, it’s the pictures that got small”)–outlined the so-called “Oxford criteria” for chronic fatigue syndrome. The paper, which had several co-authors, was titled “A report–chronic fatigue syndrome: guidelines for research.”

This definition, which required no symptoms other than fatigue, has been rejected as inadequate by the international scientific community because it conflates the symptom of “chronic fatigue” with the clinical entity called “chronic fatigue syndrome.” In 2015, a review of the literature from the National Institutes of Health recommended that the Oxford criteria were “flawed” and should be “retired” from use because they could “cause harm.” The reasonable fear was that the  heterogeneity of study samples recruited with this definition would lead to meaningless results.

But that is not a concern of the COFFI crowd, which apparently believes this ancient definition focused solely on fatigue is perfect for our times—as the group’s name itself implies. As everyone knows by now, or should know, post-infectious conditions are not just about “fatigue” but involve a host of other symptoms—in particular the phenomenon of post-exertional malaise as well as cognitive dysfunctions, POTS or other conditions involving dysautonomia, sleep dysfunctions, and other non-specific medical complaints.

Fatigue is among the most common symptoms reported to doctors and is a sign of many conditions and diseases. Targeting this incredibly broad construct for intervention, while overlooking that in post-infectious illness it is generally accompanied by a cluster of other symptoms, is misguided and kind of stupid. In a 2015 report, the Institute of Medicine (now the National Academy of Medicine) found that “systemic intolerance”—an alternate name for post-exertional malaise—was the cardinal feature of the illness.

Some of the authors of this long Covid research agenda have demonstrated something of an integrity deficit in the past. That includes Australia’s Professor Andrew Lloyd, who seems eager to impose COFFI’s 1991 agenda on his compatriots. (I interviewed Professor Lloyd when I visited Australia five years ago, and wrote about it here and here.)

The senior author of the paper, Professor Hans Knoop from the Netherlands, co-wrote a Lancet commentary accompanying the PACE trial that claimed study participants had met a “strict criterion for recovery”—which was untrue, given that a significant minority of participants had already met a key outcome threshold at baseline. This bizarre and undisclosed feature of the study appears to meet common definitions of research misconduct from the US National Institutes of Health and the UK’s Medical Research Council.

The COFFI claims are grounded in flawed and bogus science. The relevant evidence in this field is based almost exclusively on findings from unblinded studies relying on subjective outcomes for claims of improvement—a recipe for generating an unknown amount of bias. Objective measures of function, such as the amount of physical movement as measured by actometers, a fitness step-test, employment status, and other indicators have almost always shown poor results in comparison with the subjective responses.

**********
A call to investigate “psychosocial predisposing and perpetuating factors”

In the US, the Centers for Disease Control and Prevention rejected the GET/CBT paradigm and switched gears in 2017, virtually “disappearing” references to the PACE trial from its pages. In 2021, the UK’s National Institute for Health and Care Excellence issued new ME/CFS recommendations that reversed the previous guidelines. Re-analyses of the trial data have documented that the treatments did not lead to recovery and that any improvements were marginal and well within what would be expected from the bias inherent in the study design.

The COFFI research agenda paper does address up front the importance of biomedical investigations. But since it would be hard for any major research proposal on long Covid to dismiss the need for biomedical investigations, the recommendation feels like throat-clearing. After this opening, the paper makes a special plea for the importance of investigating the role of “psychosocial predisposing and perpetuating factors that play a role in PIFS”:

“Previous research has shown that psychosocial factors, such as distressing life events, may predispose to developing PIFS, whereas symptoms of depression or anxiety, cognitive factors (i.e., illness beliefs) and behavioural factors (i.e., changes in activity patterns) may act as perpetuating factors. It is therefore unfortunate that there seems to be opposition to research into psychosocial predisposing and perpetuating factors that play a role in PIFS. At the moment little research is done into the role of psychosocial factors in post-COVID-19 fatigue.

The reason there is such “unfortunate” opposition to the proposed project is because the research documenting the purported “psychosocial predisposing and perpetuating factors” in post-viral conditions and ME/CFS in particular is mostly garbage. These studies often rely only on associations and correlations, which the authors then tend to interpret as causal relationships in their favored direction. The intellectual bankruptcy at the core of some of the research is distressing. As the CDC has stated bluntly, ME/CFS is not a psychological disorder.

The paper declared that “behavioural interventions were effective in reducing symptom severity and improving functional capacity for PIFS and for similar chronic fatigue states unrelated to infection.” Moreover, “exercise therapy was previously found to be effective in chronic fatigue syndrome, but its effectiveness has not yet been studied in PIFS.”

Interestingly, the authors did not cite PACE to support any of these claims, even though the authors themselves declared it was the “definitive” trial of GET and CBT for the illness. Perhaps that piece of crap has become too toxic for even these true believers to cite at face value. Instead, the authors cited two other papers on treatments for chronic fatigue syndrome, a study of cancer-related fatigue that is pretty irrelevant in this context, and a disputed Cochrane review on exercise therapy in which PACE was the largest trial included.

Unfortunately, the authors did not inform readers of the advisory that Cochrane itself slapped on the review, which essentially indicated that even its weak reported findings could not be taken at face value. In another problematic omission, the authors also do not inform readers that CDC, NICE and other major health organizations have rejected not just PACE itself but the entire treatment paradigm based on “predisposing and perpetuating factors.” Yet that rejected paradigm is what Professor Lloyd, Professor Knoop and their colleagues are proposing to impose on the long Covid patient community.

Please. When it comes to “post-infective fatigue syndrome,” the GET/CBT ideologues already had their big chance to prove their case. It was called the PACE trial, and it bombed. Enough already.

**********

**Correction: The initial version posted yesterday did not mention that the Oxford criteria included a “post-infectious” subtype. I had used the word “rechristened” to describe the authors’ decision to cite the 1991 paper, but that implied they had changed the name. (They did, from “infectious” to “infective,” but that change is obviously non-substantive.) Rather, they had just selected a subsidiary and largely ignored section of the paper to highlight. So “rebranded” is more accurate.