By David Tuller, DrPH
It is a truth universally acknowledged (or at least universally acknowledged by smart researchers), that if the list of authors on an article includes Trudie Chalder, King’s College London’s mathematically and factually challenged professor of cognitive behavior therapy, then the article in question should most assuredly be expected to be short on, or utterly devoid of, intelligence and logical reasoning.
This is certainly the case with a recent publication in European Respiratory Journal titled “Post-Hospitalisation COVID-19 Rehabilitation (PHOSP-R): A randomised controlled trial of exercise-based rehabilitation.” Professor Chalder is one of more than three dozen co-authors, so it is unclear how much she can be held responsible for the article’s poor quality and unwarranted claims. Nonetheless, this new trial continues what appears to be her impressive streak of being involved with scholarship that can accurately be described as rubbish.
The new trial, conducted at the University of Leicester and Northumbria University, was funded by the Medical Research Council and the National Institute for Health Research. The design was not terrible. Specifically, unlike some other Long Covid rehabilitation studies, this one excluded participants with post-exertional malaise (PEM). That’s a good move, given that the presence of PEM is a contra-indication for an exercise-based rehab program.
So this was a study of Long Covid patients who do not meet criteria for an ME/CFS diagnosis—a key point.
The real problem here is that the reporting of the results stinks.
In the study, 181 participants who were experiencing prolonged symptoms after a Covid-related hospital stay were randomized to either an eight-week face-to-face exercise rehabilitation program, an eight-week remote exercise rehabilitation program, or care as usual. The primary outcome was the change in the Incremental Shuttle Walking Test (ISWT). Among the many secondary outcomes were questionnaires measuring health quality of life and symptom burden.
The conclusion: “Exercise-based rehabilitation improved short-term exercise capacity in Post-COVID syndrome following an acute hospitalization.”
First, let’s note that the conclusion does not explicitly state that the study is not about “post-COVID syndrome” overall but only about non-ME/CFS “post-COVID syndrome.” That’s a major limitation of the findings that should have been emphasized prominently throughout, given the significant numbers of Long Covid patients who do experience PEM and qualify for ME/CFS diagnoses. Health care providers will read the conclusions and assume they can be extrapolated to all Long Covid patients. That is clearly not the case.
Beyond that unacceptable oversight, let’s review the data in a bit more detail and see if that claim holds up.
The face-to-face intervention group had a drop-out rate of 29% and the remote intervention group had a drop-out rate of 39%. These drop-out rates are quite high. Remarkably, the article includes no substantive discussion of this. It is hard to argue persuasively that an intervention is successful or effective when so many participants apparently decided not to continue with it, for whatever reasons. Did they find it unhelpful? Too difficult? Harmful? We don’t actually know.
The article also overlooks the fact that the participants were, on average, almost as unhealthy after the eight-week intervention as they were beforehand. After the intervention, the average increase in the IWST for the face-to-face and remote groups was, respectively, 52 meters and 34 meters more than the increase in the care as usual group. But the average meters walked remained way, way below the levels of healthy people in the same age range.
The average age of participants in the trial was 59. A 2013 study called “Age-specific normal values for the incremental shuttle walk test in a healthy British population” found that the average distance walked during the ISWT by those in their 40s, 50s, 60s, and over 70 were, respectively, 824 meters, 788 meters, 699 meters, and 633 meters. By comparison, those in the face-to-face group increased from 285 to 312 meters, and those in the remote group from 353 to 388 meters.
It should have been obvious to any intelligent or even minimally competent researcher with that an exploration of both of these issues–the high drop-out rate and the continued poor health of the participants–was essential to put the purportedly “positive” findings in context. There is simply not that much positive to report about interventions that left participants severely disabled and that substantial numbers were unable or unwilling to complete.
Oh, and on top of that, there were null results for the trial’s seven quality-of-life and symptom burden questionnaires: EuroQol five-dimension five-level questionnaire (EQ5D), Patient Health Questionnaire (PHQ9), the Generalised Anxiety Disorder (GAD7) 7-item scale, Dyspnoea-12, the Functional Assessment of Chronic Illness Therapy Fatigue Scale (FACIT), the DePaul Symptom Questionnaire, and the Montreal Cognitive Assessment (MoCA). In other words, whatever incremental improvements might have occurred, participants did not report any overall benefits in the trial’s many subjective measures of well-being.
And to mention one other odd point…According to a 2019 study cited by the authors, the “minimal clinically important difference” (MCID) for the ISWT is 35 meters. In other words, while the results for the face-to-face intervention surpassed that threshold, the results for the remote intervention did not quite reach it. Yet here’s the opening of the discussion section:
“In this fully powered randomised controlled trial, we demonstrated that both face-to-face and remote exercise-based rehabilitation significantly improve exercise capacity compared to usual care alone in those previously hospitalised with COVID-19. These between group improvements exceed the established MCID (35m), highlighting improvements of clinical relevance in those with post-COVID syndrome.”
This last statement is simply not true in relation to the remote intervention, at least when it comes to the final, adjusted, intention-to-treat analysis. It is either a mistake or a deliberate effort to fudge the facts. I assume the latter, because it is very obvious that 35 is a bigger number than 34. With more than three dozen people on the manuscript, it is hard to believe that no one noticed this discrepancy. Either way, this indisputable error requires a correction. (To be clear, a correction won’t make the rest of the paper any better.)
In fact, the authors might have pointed out that there are multiple studies of the MCID of the ISWT, such as a 2008 analysis that found it to be 47.5 meters, and one from 2015 concluding that it was 70 meters. Even thought they chose to cite the MCID most favorable to their argument and ignore the others, they still presented false information to bolster their case.
So here’s the bottom line: Despite some marginal improvements among those who actually were able to or decided to complete the interventions, the trial documented that exercise-based rehabilitation failed dramatically to restore participants’ health. Moreover, participants felt no better subjectively on any measures than beforehand. Given those telling details, along with the fact that significant numbers of participants abandoned the trial’s intervention arms, the boast that these rehabilitation programs “improved short-term exercise capacity” is hard to take seriously.
**********
Disclosure: My academic position at the University of California, Berkeley, is largely supported by donations to the university via the campus crowdfunding platform from people with ME/CFS, Long Covid, and related disorders.