Anecdotal Science

One of the most serious problems concerning our understanding of the COVID-19 virus and its life cycle is that we are all too often dealing with anecdotal results. Aside from, in my opinion, the near fraudulent papers in NEJM and Lancet, we have one anecdote after another. Take the mortality rate. A colleague asked me if we could ascertain the mortality rate from the data provided. After some thought, I really tried to see if we could, it was clear that to obtain this we needed an experiment. Yes, an experiment, like one does in science.

So here we go.

1. We select a large cohort to monitor. We must recruit them and the best way is to pay them to participate. Say $500 per week, and that would be cheap. It must be a random sample and across a large base of age, sex, race.

2. The we test them for the virus and Ab, and we eliminate them if they are infected or have been infected.

3. We keep the remaining in the cohort to be used in the experiment.

4. We may then divide them into subcohorts by age; 20-35, 36-50, 51-65, >65 for example. These sub-cohorts may help to identify certain characteristics.

6. We then test them every week.

7. We then count infected

8. We then count the dead

9. We do this for N months where N is large. I would suggest six months or it may be none.

10. We then calculate mortality, namely dead/infected, the ration based on actual data. Remember we started with uninfected people and let them go about whatever they did.

11. The size of the cohorts should be large enough to give a p value of 1%, that is really large cohorts, about 1,000 each. We have examined this in detail in an earlier post.

This is frankly the only way to ascertain mortality rates. Now some folks have tried to do this with what is available. The most interesting was some alleged economist. As best as I can figure, the current data has a multiplicity of problems.

First and most important it is often noisy, dates are wrong and numbers wrong.

Second, and this is critical, deaths are ascribed to the virus when the true cause is an underlying disease.

Third, reporting is a mess.

Overall, this is a critical metric to have. Unfortunately as we continue to gather data we get more and more confused with anecdotes and political nonsense. In New Jersey, as compared to New York, the Governor refuses to provide the basis for his decisions, and he seems to make it up on the fly. No one really knows what the basis is for opening up anything.

Moreover the above metric, which I would consider sine qua non should be done by the Feds, unfortunately they also seem clueless on this issue.

Is this difficult? No, the FDA does trials like this every day. So why not do something folks?