Thread by @sunsopeningband

@twoShaws Some interesting notes while waiting for a flight: 🧵

The exercise cohort was not pre-screened for PEM because they did not fill out the DSQ-PEM. It's no secret that I'm not a huge fan of how investigators use this questionnaire, but they did not use the most common survey.

@twoShaws Instead, they used a visual analog scale survey and focused on a symptom cluster they found to be prevalent after CPET in a prior study. Maybe this is fine but using a different survey tool may compare apples with oranges. Seems odd to not screen for PEM using a common approach.

@twoShaws Seems even odder they wouldn't use the DSQ-PEM in people with suspected PEM because *mounts soap box and loudly inhales* IT IS IN SET OF COMMON DATA ELEMENTS CREATED BY A *checks notes and loudly inhales again* AN NIH WORKING GROUP. SO WHY WOULDN'T THEY USE THEIR OWN GUIDELINES?

@twoShaws The article isn't clear on what criteria they used to determine PEM objectively, but it is a problem they didn't use two-day CPET. It's the *decrease* in submaximal oxygen consumption and work rate that has emerged as an indicator of PEM. It's not perfect for many reasons, but...

@twoShaws ...what it avoids is erroneous conclusions based on a single CPET, in which people with PEM often resemble people who are deconditioned. We've learned you really need the second test in the post-exertional state if you're going to do this for the purpose the authors intended.

@twoShaws What the authors seem to have done is commit the same rookie mistake we frequently see, which is conflating PEM with exercise intolerance. PEM involves exercise intolerance, but exercise intolerance is not PEM. Rather, PEM is an impaired recovery response to an exertion. So, ...

@twoShaws ...you need a second CPET to make you're positioned to observe the decline in functioning. People will point to your single CPET results and say things like, "well, they're just deconditioned." And it would be hard to rebut that because you don't have data from a second test.

@twoShaws My colleagues and I are all for trying to make life easy for our future selves as a researchers by our present selves as a researchers thinking through these issues carefully and making good choices up front, so we don't end up with dodgy data. Life is hard enough as it is.

@twoShaws I'll also contribute that there is absolutely no way you can establish prevalence if the sample isn't representative of the population, and this sample isn't. The point estimates for people with and without PEM, as oddly defined as it is for this study, can be safely ignored.

@twoShaws At the end of the day, if the CPET results are not valid and the epidemiology of PEM isn't right, I have no idea what this study contributes. Maybe I'll think about it up in the air. Or maybe I'll just try to sleep it off. Stay tuned.