[BostonCard]

Maybe, they somehow overcame the statistical one-sidedness of their data by adjusting for it, but then I have to wonder if the conclusions are just a reflection of how over or under that adjustment was.

Indeed they did:

To minimise the effect of confounding factors, a propensity score matching analysis was done individually for each of the four treatment groups compared with a control group that received no form of that therapy. For each treatment group, a separate matched control was identified using exact and propensity-score matched criteria with a calliper of 0·001. This method was used to provide a close approximation of demographics, comorbidities, disease severity, and baseline medications between patients. The propensity score was based on the following variables: age, BMI, gender, race or ethnicity, comorbidities, use of ACE inhibitors, use of statins, use of angiotensin receptor blockers, treatment with other antivirals, qSOFA score of less than 1, and SPO2 of less than 94% on room air. The patients were well matched, with standardised mean difference estimates of less than 10% for all matched parameters.

Basically what they did is for each patient in the treated arms, they found a patient in the control arm that "matched" them based on the criteria (more specifically, their propensity score, which is the probability that they would be treated based on their baseline characteristics).  The question you pose is a good one, and basically amounts to a term called residual confounding, the idea that the adjustment might be inadequate and that some other factor(s) might be there that they failed to adjust for.  The authors note this as a limitation in their study.

Due to the observational study design, we cannot exclude the possibility of unmeasured confounding factors, although we have reassuringly noted consistency between the primary analysis and the propensity score matched analyses.

I think at this point we are at the limits of what observational data can tell us.  They bring up some red flags as they have consistently shown no benefit and possible harm, but there is no getting around the methodological limitations. So, we will await the results of the randomized, placebo controlled studies (for example, the WHO has one) before making a firm conclusion, but at this point, I wouldn't give a patient of mine CQ/HCQ outside of a clinical study.

BC

[BostonCard]

Maybe, they somehow overcame the statistical one-sidedness of their data by adjusting for it, but then I have to wonder if the conclusions are just a reflection of how over or under that adjustment was.

Indeed they did:

To minimise the effect of confounding factors, a propensity score matching analysis was done individually for each of the four treatment groups compared with a control group that received no form of that therapy. For each treatment group, a separate matched control was identified using exact and propensity-score matched criteria with a calliper of 0·001. This method was used to provide a close approximation of demographics, comorbidities, disease severity, and baseline medications between patients. The propensity score was based on the following variables: age, BMI, gender, race or ethnicity, comorbidities, use of ACE inhibitors, use of statins, use of angiotensin receptor blockers, treatment with other antivirals, qSOFA score of less than 1, and SPO2 of less than 94% on room air. The patients were well matched, with standardised mean difference estimates of less than 10% for all matched parameters.

Basically what they did is for each patient in the treated arms, they found a patient in the control arm that "matched" them based on the criteria (more specifically, their propensity score, which is the probability that they would be treated based on their baseline characteristics).  The question you pose is a good one, and basically amounts to a term called residual confounding, the idea that the adjustment might be inadequate and that some other factor(s) might be there that they failed to adjust for.  The authors note this as a limitation in their study.

Due to the observational study design, we cannot exclude the possibility of unmeasured confounding factors, although we have reassuringly noted consistency between the primary analysis and the propensity score matched analyses.

I think at this point we are at the limits of what observational data can tell us.  They bring up some red flags as they have consistently shown no benefit and possible harm, but there is no getting around the methodological limitations. So, we will await the results of the randomized, placebo controlled studies (for example, the WHO has one) before making a firm conclusion, but at this point, I wouldn't give a patient of mine CQ/HCQ outside of a clinical study.

BC