From what I've read elsewhere, this doc's guesstimate is based on three elements:
* The population of those who have received at least one of two vaccine shots, currently ~42 million people;
* The population of those who have recovered from Covid, currently ~ 28 million people (total cases - total deaths);
* His estimate of the population of those who had and recovered from Covid without being tested, possibly not knowing they had it.
That later, unknown but estimated, number is significant to his April guesstimate. One can quibble that number, but unlike some "experts'" pronouncements, at least he's spelled out the bases for his guesstimate, i.e. transparency. One thing that is obvious is that as those three populations increase the cumulative susceptibility to Covid in the US will decrease. Had states like CA, OR, WA, and NY done as well as the Dakotas, WV, and NM (especially early on!) the US would clearly be seeing the effects of vaccination numbers on new cases numbers instead of thinking we might be seeing the effects of vaccinations numbers.