09-21-2020, 10:09 AM
Eric Topol (known for having identified the safety issues in Vioxx), weighs in:
[tweet]https://twitter.com/EricTopol/status/1307381521745997824?s=20[/tweet]
It's a very good thread about what it would take to get an EUA for a vaccine. Basically, per the stats, the Pfizer study can have a look once they reach 32 events (COVID-19 infections), and the DSMB can declare efficacy if 26 of those 32 occur in the placebo arm. That is pretty overwhelming evidence of efficacy. This would be a 77% efficacy, with a confidence interval of 44 to 92%, which would easily satisfy the FDA's requirements.
What you wouldn't be able to figure out is whether the vaccine reduced severe COVID-19 pneumonia (hospitalizations) or affected deaths, because you almost certainly wouldn't have enough severe COVID-19 cases to be able to tell. It would not be unheard of for a vaccine to reduce the risk of developing a disease, but if people did get it, they were more likely to get a serious version. This happened with the dengue vaccine, for example.
BC
[tweet]https://twitter.com/EricTopol/status/1307381521745997824?s=20[/tweet]
It's a very good thread about what it would take to get an EUA for a vaccine. Basically, per the stats, the Pfizer study can have a look once they reach 32 events (COVID-19 infections), and the DSMB can declare efficacy if 26 of those 32 occur in the placebo arm. That is pretty overwhelming evidence of efficacy. This would be a 77% efficacy, with a confidence interval of 44 to 92%, which would easily satisfy the FDA's requirements.
What you wouldn't be able to figure out is whether the vaccine reduced severe COVID-19 pneumonia (hospitalizations) or affected deaths, because you almost certainly wouldn't have enough severe COVID-19 cases to be able to tell. It would not be unheard of for a vaccine to reduce the risk of developing a disease, but if people did get it, they were more likely to get a serious version. This happened with the dengue vaccine, for example.
BC