TennTradition
Defended.
- Joined
- Aug 14, 2006
- Messages
- 16,917
- Likes
- 820
It’s almost like it’s really hard to predict exactly how that would play out with 50 different governors doing 50 different things on 50 different timetables.
The harder part is translating a policy into a number in the model. How strictly will it be adhered to. How undermining is the list of essential services. And once you have a handle on what you think it means for people movement, how does that policy actually impact the Reffective of that community.
That’s what’s interesting about the IMHE model. They called the peak fairly well which is highly dependent on those factors. They also - in their first model run around 3/20 (and published 3/26 in their pre-review paper) - predicted that the US would peak at around 2,300 deaths per day in the second week of April. I don’t know where we’ll peak but that has a chance of being right. They were never predicting millions of deaths in the first wave - or even a base case of over 100,000 deaths.
However, they were predicting upwards of 90k deaths - mainly because they saw hospital resources being overloaded. They still see that in NY - but to a lesser degree than they did. They also don’t see that for most of the rest of the country now. The primary reason seems to be because the international numbers they used for the number of patients that were hospitalized for every ultimate death, how long patients who never saw the ICU were hospitalized, and how long patients who do see an ICU are hospitalized ended up being wrong here in the US. They’ve updated their model with more US statistics as we’ve gotten them and those hospitalization rates have come down. I understand there are still issues there - it it’s also disingenuous for some to attack the model for its hospitalization predictions but ignore the pretty good job it’s done on predicting deaths thus far and peaks (to my surprise still).