This article critiques a Norwegian randomized clinical trial on a "Brief Outpatient Rehabilitation Program" for Post-COVID-19 Condition (Long Covid), published in JAMA Network Open. Author David Tuller argues that the trial's investigators misrepresented their findings by claiming the intervention was "effective," despite the primary outcome (SF-36 Physical Function Subscale) failing to meet its pre-specified 10-point "clinically significant" threshold, showing only a 9.2-point inter-group difference. Tuller highlights that the authors deceptively emphasized within-group improvement instead of the critical inter-group comparison. The critique also points out significant methodological flaws, including the trial's unblinded nature, reliance on subjective measures, and a high participant dropout rate (26%). Tuller extends his criticism to JAMA Network Open for publishing the study with such misrepresentations and to media outlets for credulous reporting, linking it to a broader pattern of questionable research by proponents of cognitive behavioral therapy (CBT) and graded exercise therapy (GET).
Critique of a Norwegian Long Covid Rehabilitation Trial: Misrepresentation and Methodological Flaws
Central Argument: The article strongly criticizes a Norwegian randomized clinical trial for Long Covid rehabilitation for misrepresenting its findings and exhibiting significant methodological flaws, including bias.
Trial Overview:
Published in JAMA Network Open, titled "Brief Outpatient Rehabilitation Program for Post–COVID-19 Condition: A Randomized Clinical Trial."
Senior author: Vegard Bruun Wyller, University of Oslo.
Involved 314 patients with "mild to moderate" Long Covid (PCC).
Intervention group received 2-8 sessions based on a cognitive and behavioral approach (e.g., "cognitive reassurance," "socratic dialogue," "guided discovery"); control group received "care as usual."
Misrepresentation of Primary Outcome:
The primary outcome was the SF-36 Physical Function Subscale (SF-36-PFS), with a pre-defined "clinically significant" threshold of a 10-point change.
The actual difference between the intervention and non-intervention groups was 9.2 points, below the clinically significant threshold.
Despite this, the abstract misleadingly claimed scores "improved statistically and clinically significantly in the intervention group" by focusing on within-group change, not the critical inter-group comparison.
This lack of clinical significance was not mentioned in the discussion or limitations sections of the paper.
Methodological Concerns:
The trial was unblinded and used subjective measures, creating a "recipe for bias" and suggesting that observed modest positive results might be an artifact of the study design (placebo effect).
A high dropout rate was noted, with 26% (83 of 314 participants) not providing final data.
Broader Critique:
JAMA Network Open is criticized for publishing the study and allowing the misrepresentation.
The author argues that this reflects a pattern where proponents of CBT/GET (Cognitive Behavioral Therapy/Graded Exercise Therapy) disregard basic research principles.
Credulous news coverage from outlets like NRK and MedPage Today failed to report the lack of clinical significance, propagating inflated claims.
Concerns are raised about this trial's findings potentially influencing future "living" systematic reviews of Long Covid interventions.
Reader Comments Highlight:
Questions about the study's claim of being "safe," suggesting it may be a predetermined outcome without sufficient follow-up for harm.
Requests for explanations regarding the high dropout rate and information on participants who worsened.
Concerns about the economic incoherence of an intervention with such "marginal gains" for a large population.