Read on Twitter
1. Regarding death projections and the mortality delaying effects of ventilators THREAD
2. One of the reasons I'm so bearish midterm is that the officials have messed up on their death projections by basing them on Italy. We saw that the last few days as media talked about how the models messed up and we're going to see way less deaths than we thought.
3. Italy did a hard lockdown over 2 days, March 8-9. The US did hard lockdowns of key cities from March 19 (California) to March 31 (Phoenix). Italy had a peak death span of 14 days. From The diff in lockdown period we can extrapolate a US peak death span of 25 days.
4. Ofc we must account for diff populations of diff regions. When NYC slows down the level will come down. But the numbers still will not drop as quickly as those for Italy bc there are many laggard regions. E.g., Miami locked down March 30.
5. Now the key pt on why our peak projections are too early. Italy ran out of ventilators in the worst hit region (the north). They were forced to triage patients and choose not to put older/sicker patients on ventilators, resulting in very early deaths for those patients.
6. The US has never run out of ventilators, and has put even those patients with a very low chance of survival on them.
7. Ventilators extend lifespan by 10 or more days. One estimate has it at ~20 days median. Anecdotally, I have seen 10+ days so let's be conservative and use 10 days.
8. However, case survival on a ventilator is very bad. A UK study found of 98 ventilated patients, only 33 were discharged alive. That is a mortality rate of 66% once put on a ventilator. https://www.npr.org/sections/health-shots/2020/04/02/826105278/ventilators-are-no-panacea-for-critically-ill-covid-19-patients
9. I don't expect US mortality rates to be much better than Italy's bc Italy did intubate those patients who had a decent chance of success. The US intubating even patients with a very low chance of success likely delays but does not prevent their deaths.
10. Looking at Italy's lockdown of the worst region on March 8 and Italy's exact peak death daily rate of 919 on March 27, you can infer a 19 day delay b/w lockdown and peak death. https://en.wikipedia.org/wiki/2020_coronavirus_pandemic_in_Italy
11. Now, for the US, let's add 10 days (again, a conservative estimate) for ample ventilators for all patients. That puts the peak death estimate at 29, not 19, days after lockdown.
12. Earliest lockdown was California on March 19. Latest was Arizona on March 31. 29 days gap puts the very peak range for those location at April 17-29.
13. Now, Italy had about a 7 days range in either direction from the peak on March 27 that was almost as bad. So from April 17-29 peak we get a substantial ramp from April 10-17 and then a high, sustained level from April 17-May 6.
14. Again, we must account for a drop in level when NYC begins to get better, but again due to the lag time in many other large cities (pay special attention to Miami and Phoenix in coming weeks), I do not expect the kind of drop off seen in Italy.
15. In summary, I believe we are looking at a ramp up from April 10-17 and then a peak range from April 17-May 6. This much slower ramp up (relative to Italy) comes due to ventilators while the extended timeframe comes from ventilators and slower rolling lockdowns.
16. If you begin hearing experts or those on the media saying "the models predicted the deaths to spike earlier" you should think of ventilators.
17. While NYC hospitalizations are slowing, which makes perfect sense due to lockdown more than 2 weeks ago on March 22, there are likely to be many ppl on ventilators. Median intubation is 8 days after symptoms. Ventilator adds >10 days to lifespan.
https://www.uptodate.com/contents/coronavirus-disease-2019-covid-19?source=history_widget
https://www.uptodate.com/contents/coronavirus-disease-2019-covid-19?source=history_widget
18. As somber as it is, I believe ventilators are a big confounding variable differentiating the US from Italy in the models. Thus they are a big part of why the media is hugely jumping the gun in saying "NY was supposed to be spiking earlier (19 days just like Italy).
19. This disease can take a very long time to kill. Sometimes more than a month since symptoms when factoring 8 days from symptoms to intubation and then sometimes another >3 weeks on a ventilator.
20. But a longer period on a ventilator does not mean more survival. It actually means the opposite. As Gov. Cuomo put it, "The longer they’re on, the more likely they’re not going to come off..." https://www.bloomberg.com/news/articles/2020-03-26/n-y-death-toll-up-to-385-ventilators-used-as-long-as-30-days
21. Officials and the media are adding to complacency by saying, less than 3 weeks after the earliest lockdown, that it looks like things might not get too bad. The models have also done a disservice by not identifying a clear confounding variable in ventilators.
22. In fact, for 5 days from April 2-6 we saw deaths remain stable around 1200/day. Then on April 7, later than many expected, a sudden jump to 1938. And today we are on track for well over 2000. The diff in ventilators has delayed the ramp up, but it is beginning.
23. I fear the biggest pain will be felt around April 17-23, well over a week after initially projected (by using the Italy peak timeframe) and will be sustained well into May.
24. Honestly, I don't care much about what that does to the markets, relative to what it means for American lives. But I'd be a fool to not recognize what it will mean if we near April 30, the day we are supposed to be getting back to work, and daily deaths are still >3000/day.
25. At that point, the markets will either react to opening things up with >3000 deaths/day or extending lockdowns for another month. Either option will be a death knell and reinforce a bigger collapse and 2nd wave down than we saw in the 1st.
26. I don't expect this thread will get too many likes or feedback. It's pretty macabre, and I don't really enjoy writing it in any way. But it's based on thorough research, and it explains what is going on.
