As for death projections, the promising thing about all of these projections over the next couple months are that despite a rise in deaths, and still a lot of deaths, they still only have the peak at about a little over 800 per day in mid August. Since the data show a younger skew than the earlier peak, and treatment appears to be better, the current waves IFR is projected to be much lower.

Hopefully that holds true; however, while I think it will, I think the infection (and case reporting) to death lag this time around could be a bit longer due to couple of reasons, not just the obvious fact that we’re seeing an increase this week that was later than many projected.

1. With more widespread testing catching pre-symptomatic cases. For example, my FIL was tested Tuesday or Wednesday of last week but wasn’t symptomatic until Saturday (now in ICU but it seems precautionary to get his oxygen levels up a bit more so hopefully it doesn’t get worse). Now unfortunately because the lab was off Friday it didn’t come back until Monday morning after he as symptomatic (and tested with results from hospital in less than a day), but it they hadn’t been, the results were ready (and probably reported) more than a day before symptoms. In the early peak, a lot of people weren’t tested until after symptoms presented, sometimes well after, if at all.

2. The turnaround on tests seems to have been faster recently. Save for the holiday issue above, my FIL’s hospital test came back the next day, and so did my MIL’s negative (thankfully test). My brother was tested yesterday and they said 2-3 days so we’ll see.

3. Related to number 2 seems to have been less of a backlog recently whereas the first peak was full of them. That said, my aunt was tested because she’s having surgery (supports catching pre-symptomatic cases) and they tried to get her in earlier because of backlog. I’ve seen other reports of this around the country the last couple of days, so this promising trend may be trending the other way.

4. While it’s skewing younger and a lower percentage are likely to have severe complications and die, does that also mean that those who do die are able to fight for their lives longer? It just seems possible to me that if their immune systems are more likely to fight it off, those who do eventually succumb will fight it off for longer. If so, that would create a longer lag.

5. Due to this younger skew, it’s likely a significant portion of those got it because they were less at-risk of severe complications and/or death, knew this, and were doing things that make them more at-risk of catching it though. But after it started spreading and continues to spread, it might be harder to keep it from spreading to those who are at-risk of complications (not just age) who were taking more precautions because of that. If that has happened (and is happening) then proportion of infections of those at-risk may be increasing and creating a lag when compared to the spread, which creates the death lag.

6. One of the main reasons that IHME model over predicted hospitalizations but under predicted deaths is early on there were far more hospitalizations to deaths. But when deaths starting increasing quickly towards and through the peak, especially in hotspots, this ratio quickly and dramatically decreased. So paradoxically hospitals didn’t get overwhelmed, BUT this may indicate the lags were longer (at least hospitalizations to death lag) early on but then deaths (as we saw) increased quickly (people also started dying at home). With some full hospitals, it’s possible we’re getting to that point.

So while this seems like this is a really negative post, but I don’t think we’ll get anywhere near the peak deaths as the first wave and the IFR will be lower. That said, the differences (and some similarities that may now be happening) in this wave have caused a longer case to death timeline and 800ish peak deaths (over a week or so) a little more optimistic. And I could see it going between 1000-1200, still half or even less of April’s peak though.