Thread by @JamesTodaroMD

1/ My thoughts on the NY study on HCQ published in NEJM.

First, use of a "composite endpoint" of death + intubation is strange. Composite endpoints are reserved for studies with rare endpoints. However, death wasn't rare in this study, occurring in >10%.

nejm.org/doi/pdf/10.105…

2/ Moreover, the authors fail to provide number of deaths or intubations in the propensity-score matched control group (n=274)—leaving us unable to calculate our own mortality rate comparison. (Authors omit this from the supplementary table below.)

3/ Secondly, the study does not address the fact patients in respiratory distress treated w/ HCQ had better outcomes.

Assuming intubation indicates severe resp distress, intubated patients treated w/ HCQ had a 68% probability of surviving (compared to only 35% in no-HCQ group).

4/ Thirdly, the HCQ treatment arm had both older & more severely ill patients at baseline than the no-HCQ group.

The authors openly admit this disparity, and adjust with the propensity-score match. However, again, they fail to provide mortality & intubation data for this group.

5/ Lastly, patients were excluded if they died or were intubated within 24 hrs. This isn't enough time for HCQ to reach therapeutic levels, which usually takes 2-3 days.

Death/intubation within 72 hrs should've been excluded.

Acknowledgment @gummibear737
ncbi.nlm.nih.gov/pmc/articles/P…