A lot of people are eager to say chloroquine is the answer and say, "forget the small details, look at that p-value!". But this article raises a bunch of important points. The important ones IMO are:
1) 6 / 26 of the individuals in the experimental group got removed from the stats because their condition got worse and the treatment ceased. One of them died.
2) The metric used to test the efficacy is viral load. You can get a negative on this and still have the disease. This is going to be a function of how long you've had the disease, which isn't controlled in the study.
1) The combination of HCQ and AZ was hyped as superior because all 6 patients did better. But looking at the data there is no significant difference in viral load outcomes between both drugs and just HCQ.
2) Many of the author's claims of statistical significance disappear if you use more conservative estimates for when people were supposed to be PCR positive except for Day 6 (now 0.01 instead of 0.001).
3) The paper turnaround was incredibly fast, with less than 24 hours for peer review. One of the authors of the paper is an editor of the journal in which it was published. Could be indicative of poor study controls.
I hope this turns out to be useful. I would not be surprised if it did not. Regardless, the best case takeaway of THIS information does not inform us about how the drug performs in the most critical area (severe ICU patients).
If you're a non-bio programmer please try to recognize that saying the current paper is good enough to justify mass distribution of chloroquine tablets to everyone is about on par with the cliche clueless product manager making bold decisions about how to move forward with your codebase despite having no clue how it works themselves.
1) 6 / 26 of the individuals in the experimental group got removed from the stats because their condition got worse and the treatment ceased. One of them died.
2) The metric used to test the efficacy is viral load. You can get a negative on this and still have the disease. This is going to be a function of how long you've had the disease, which isn't controlled in the study.
If you have a gripe with the author because they're not a medical scientist, read the peer review comments here: https://pubpeer.com/publications/B4044A446F35DF81789F6F20F8E...
They are easy to understand as a layman I feel.
Other noteworthy takeaways:
1) The combination of HCQ and AZ was hyped as superior because all 6 patients did better. But looking at the data there is no significant difference in viral load outcomes between both drugs and just HCQ.
2) Many of the author's claims of statistical significance disappear if you use more conservative estimates for when people were supposed to be PCR positive except for Day 6 (now 0.01 instead of 0.001).
3) The paper turnaround was incredibly fast, with less than 24 hours for peer review. One of the authors of the paper is an editor of the journal in which it was published. Could be indicative of poor study controls.
I hope this turns out to be useful. I would not be surprised if it did not. Regardless, the best case takeaway of THIS information does not inform us about how the drug performs in the most critical area (severe ICU patients).
If you're a non-bio programmer please try to recognize that saying the current paper is good enough to justify mass distribution of chloroquine tablets to everyone is about on par with the cliche clueless product manager making bold decisions about how to move forward with your codebase despite having no clue how it works themselves.