By David Tuller, DrPH
The other day I posted a blog about yet another problematic Long Covid study published by a major journal. The study concluded that physical activity (PA) can “reduce” symptoms in young women with prolonged medical complaints after Covid-19, or what the authors call post-COVID condition (PCC). The problem: the study design is capable of documenting associations but not causal relationships—such as the relationship described by the use of the verb “to reduce.”
This morning I sent the following letter to Dr Rafal Marszalek, the top editor at Scientific Reports, the journal from the Nature group that published the study.
**********
Dear Dr Marszalek:
Scientific Reports recently published a study called “Higher level of physical activity reduces mental and neurological symptoms during and two years after COVID-19 infection in young women,” from Takács et al. In outlining the goal of the research, the investigators wrote the following: “This study aimed to examine the hypothesis that regular PA [physical activity] reduces the number of symptoms during and after COVID-19 infection in young women.”
And here is an assertion from the conclusion: “The salient finding of the present study is that a higher level of regular physical activity can reduce mental and neurological symptoms such as anxiety, depressed mood, fatigue, dizziness/light headedness, and forgetfulness both in acute COVID-19 and in PCC [post-COVID condition] among young women.”
Unfortunately for the investigators, the cross-sectional study design does not allow for any examination of causality—a point that should have been obvious to Scientific Reports’ editors and peer-reviewers. Because cross-sectional data is collected at one point in time, it is impossible to determine what is cause and what is consequence. The findings generally allow for causality in both directions.
In this case, the authors state categorically that, according to their findings, physical activity “reduces” symptoms. This use of explicit causal language is unwarranted and unjustified. It is just as plausible that those who had more symptoms were simply unable to perform physical activity at the same level as those who had less symptoms. For the authors to discount that possibility in favor of their preferred explanation is antithetical to proper scientific reasoning.
The authors do mention that the cross-sectional design is a limitation. But it is pointless to acknowledge limitations when you choose to ignore the limits they clearly impose. It is therefore perplexing that this paper was published in this self-evidently deficient form. The title, the conclusion and every sentence that suggests the study has documented a causal relationship must be corrected.
Thank you for your quick attention to this matter. (I have cc’d the corresponding author.)
Sincerely–
David Tuller
Senior Fellow in Public Health and Journalism
Center for Global Public Health
School of Public Health
University of California, Berkeley