3 ms·
Seems like the sample size is probably too small to draw any conclusions there. 131 got covid in the trial, if 70% effective than 39 should be from the group th
by bawolff 6y ago
Seems like the sample size is probably too small to draw any conclusions there. 131 got covid in the trial, if 70% effective than 39 should be from the group that got the vaccine (i think. Did i do that right given they're combining the two dosing regime)?
Maybe its chance that none of those 39 got really sick. I'm not sure what the age breakdown of trial participants was- were the people in the oxford trial younger? Did the distribution match the distribution from the mRNA trials? I have no idea.
- Mumps 6y ago(afaict) You're right about the 39 (risk reduction = 1-39/131 ~= 70%). But you're less correct about saying "smple size is probably too small" (in general, people really really really need to stop dismissing things based on sample sizes until they actually go through a Power calculation!). Your question about demographic breakdown is entirely valid. Hospitalization rate for covid is about 228.7 hospitalizations per 100,000 population [0]. Or, more easily interpretted, about 10-20% of cases turning to hospitalization. Taking the more conservative (10%) rate, we'd need 31 samples for a 95% CI and 5% margin of error. Thus, the interpetation is that is it unlikely that the _lack of seeing any hospitalizations in the vaccine+covid-positive group is by chance_. Hope that helps! [0] https://www.cdc.gov/coronavirus/2019-ncov/covid-data/covidview/index.html https://www.cdc.gov/coronavirus/2019-ncov/covid-data/covidvi...
- colinmhayes 6y agoI don't think it's fair to say " that is it unlikely that the _lack of seeing any hospitalizations in the vaccine+covid-positive group is by chance" until we see the demographic breakdowns. The difference in hospitalization likelihood between demographics is too big to ignore.
- bawolff 6y agoThanks, this is a good point. I just blindly assumed it was too small because it was a small number. I think there was a bit of an implication in the parent post that it would eliminate hospitalizations (instead of just reduce), i still think that is unrealistic.