If I understand it correctly, IFR is estimated. The denominator is the number of infections - estimated somehow from actual tested infections and some other method to estimate untested infections. Ignoring changes due to different estimates of untested infections, some of the drop could be due to a smaller proportion of the high risk people catching it due to efforts to protect them and those in lower risk groups doing little to protect themselves. As you say, better treatment definitely appears to be improving IFR as well. Some of it could also be that our behavior and/or “seasonality” lowered the initial viral load obtained by those infected. There is evidence that the initial load is directly related to the severity of the illness. If that’s the case, we may see IFR go up this winter - or, not drop as much as it would if we were in our summer.
you are right. It might go up. We don’t know. But if it does it probably won’t be by much and it might also just stop falling or slow it’s fall. One way the check is to see what happened in Switzerland. Don’t have those figures but they are pretty much on the downslope and without massive government intervention...would be a possible approximate or comp of the worst case...what happened there?