[jyl]

So, I have been thinking about whether Covid therapeutic drugs still look like a large commercial market and thus an interesting investment area, given the vaccine efficacy data from Pfizer and Moderna. The basic question is if in a year, there will be enough persons getting infected and dying to generate blockbuster=type revenue for an effective Covid drug.

I have a simple model for this, of course, so I plugged in some new assumptions:

- relevant population = 300MM (US population, excluding very young children).

- vaccine efficacy = 90% (in persons vaccinated, vaccine prevents 90% of future serious Covid cases). We don't know if this is what vaccine efficacy in the real world will turn out to be, but seems like a good initial guess based on data known so far. I assume vaccine is redosed as often as needed to maintain the efficacy.

- acquired immunity efficacy = 80% (in persons infected and recovered, natural immune response prevents 80% of future serious Covid cases). I think the efficacy is far higher initially (because confirmed re-infections are so incredibly rare) but there's a lot of concern that naturally immunity will decline over time, so I used 80% rather than rebuild the model to use a high number initially then ramp it down.

- anti vaxer = 30% (30% of population refuses vaccine).

- actual/reported = 5X (for every confirmed Covid case, there are actually 4 more undetected cases that were asymptomatic/very mild - most estimates of this range from 3X to 10X).

- vax start = Dec 2020. This might be optimistic by a month.

- vax/month = 35MM (once vaccination starts, we vaccinate 35MM people/month until all who are not anti vaxers have gotten it.

- infection fatality rate "IFR" = 0.39% (of all actual cases, including undetected ones, the death rate is 0.39%; this excludes young children). It is hard to be confident in a point estimate, but it won't end up mattering too much if that is 1/2X or 2X the correct number. This assumes no new therapeutics.

- number of infected as of end Sept = 36.5MM (the cumulative total number of actual infections since the start of the pandemic, including undetected ones, is about 11% of the total US population). That is roughly inline with most estimates I've seen.

- infection attack rate = 50%/yr (of all not-immune persons in the population, in one year 50% of them will get infected, including undetected cases). This is the hardest assumption to figure out. It has not been that close to that high for 2020 to date, and probably is not that high even now and even in the worst-hit states. But I am assuming that by spring 2021, pretty much all social distancing, masking, restrictions, etc will end, and the whole country will be back to life as before, packing into bars and parties and planes and touching and breathing on each other with absolutely zero precautions. The wisdom of that could be debated, but I think it is going to be reality. So, how fast did the virus spread in early 2021, when there were zero precautions and everyone was not-immune? Incredibly fast. How fast can an exponential curve rise, if no-one bothers to slow it? Incredibly fast. That's why I'm using a high number.

So you can imagine how a simple model would work. Every month 4.17% of the naive population (not vaccinated, not previously infected) gets infected, 4.17% x 10% of the vaccinated population gets infected, 4.17% x 20% of the previously infected population gets infected, and you subtract off the overlapping population. Then apply the IFR to get deaths, update the population numbers and repeat next month.

You can also see all the problems with such a simple model, not to mention the uncertainty around the assumptions. But for my purposes, being "point accurate" doesn't matter, I just want to try different assumptions and figure out what a ballpark range of reasonable scenarios looks like.

Anyway, I'll cut to the chase: I think it is quite reasonable to think that in 2022, the US has 20-40 MM new Covid cases (of which only about 3-8MM are detected), 0.3-0.8 MM hospitalizations, and 70K-180K deaths (that's deaths if there are no effective therapeutics - the fundamental point of this exercise is that there will be effective therapeutics so that actual deaths will be much lower, but you gotta pay for those therapeutics doncha know).

These results are not all that sensitive to the reasonable changes in the assumptions, by the way.

And I now better appreciate why experts are warning that Covid is going to be with us, in some form, for some time.

So, somewhat comforted - yes, it is a venal, morbid sort of comfort - I'm going to keep looking for "Covid drugs that work [for investors]" :-)