Yes, not only me, it's bugging everyone. First nobody knows if the anti spike antibodies are protective or not ( if protective there must be a primary rise following at least natural infection, but all the studies measuring this antibody found out that the response is non linear, even studies with non random samples couldn't find a linear relation with time from recovery.
By studies, I mean those give us access to their all raw data and clearly mention sampling techniques used.
With genuine random sampling the antibody responses are all scattered ( beyond any semi conclusive range)
I actually tried to make a list of antibody response studies submitted to us ( only 6 has proper random sampling, only two with more than 500 sample size)
It's totally blurry to me from statistical POV.
But one thing is clear ( with strong statistical correlations) All the vaccines used today prevent serious disease ( CT and few clinical and laboratory values) incuding Covaxin. Bharat Biotech did send us some useful data ( but for some reason they haven't included approximate geolocation... don't know why...it's very important specially when their genomic sequencing data is practically non existent...it's very sparse from India, there is no way we can predict efficacy against newer strains.
If they could give us some broad location ( at least cities/ districts) we could still calculate something from the strain mapping indirectly (definitely not very accurate, but better than nothing)
Bharat Biotech's data collection was somewhat different ( and under given circumstances, rather intelligent/ interesting) instead of giving much importance on antibody response, clinical symptoms ( with or without RT PCR positivity) were taken into account. Though it pissed microbiologists, it's a nice approach from public health point of view ( they assumed 90-93 percent of adult symptomatic patients, without CT scan or RT PCR to be positive, did an absolutely brilliant calculation with available seasonal flu prevalence during the months of study in an area to come to those percentages....really high grade stuff.. everyone knows basic statistics, but how one uses it to improvise logically defines good statistician...it would have a really breakthrough paper alone if they could give the locations of study population...which they obviously have and used, but didn't give the raw data)
I felt proud for Indian statisticians after a long time. It shows, this country still has top notch people.
www.rediff.com
