Thread by @tmprowell

🧵As a scientist & physician, when our #pandemic predictions are off, I think it's important to circle back & analyze why. In late Dec 2021, I tweeted a🧵predicting widespread disruption of everything from health care & pharmacies to grocery stores & schools in Jan/Feb '22. 1/x

The reason for my prediction was the pattern of a huge spike in #COVID19 cases the prior winter (driven in part by holiday travel) & the fact that #omicron was considerably more contagious than prior #COVIDvariants & more able to evade immunity from vaccines or prior infxn. 2/x

The pace in growth of cases in Dec '21 in countries w/ early #Omicron outbreaks was stunning. The graphs were vertical. In the US, we appeared on target to have >500K cases/d in the US. We had no idea yet how long the surge would last. 3/x

The risk of reinfection with #Omicron was reported by the UK to be >5 fold higher than it had been for #Delta or prior #CovidVariants (imperial.ac.uk/news/232698/om…). And we were seeing breakthrough #Omicron even in people who were vaccinated & boosted (time.com/6130704/breakt…). 4/x

We had <2/3 of US fully #vaccinated & far fewer #boosted, which meant a lot of our pop was at ⬆️ risk to get infected. Even if #Omicron was less likely to cause severe illness in a given person than prior variants (which is true), the societal impact would still be large. 5/x

So where did the prediction go awry & why? First, up to that point, CDC had recommended a 10d period of isolation for those who were infected. As daily #Omicron cases surged to levels several times higher than ever before, the risk of societal disruption became clear. 6/x

Even if most w/ #Omicron didn't get severely ill, if all who were infected (& all of their close contacts) had to stay home for 10d, we'd likely see sig workforce shortages impacting mult lines of work. On 1/4/22, CDC cut the isolation & quarantine period in half to 5d. 7/x

This thread is about my errors, & not about whether I believe that CDC call to shorten was wrong (I'll come back to the latter), but their decision to ⬇️isolation & quarantine by 50% clearly shortened absences from work per infected person & mitigated labor disruption. 8/x

For essential public needs (schools, healthcare, groceries, meds), esp those where it's not straightforward to find a sub--we can't instantly get more pharmacists or drs or nurses or school bus drivers--that prob had the greatest impact on observed vs predicted disruption. 9/x

The second reason I believe my prediction was off is public behavioral change. There was widespread messaging that #Omicron was more contagious & immune-evasive. We saw a spike in # of people getting vaxxed/boosted, & many returning to or upgrading masks, w/ or w/o mandates. 10/x

And some (incl us) stopped doing activities we had returned to in summer/fall '21 that no longer felt safe w/ #omicron. Scientists can tell you that observing a system (especially aloud on Twitter & the news😉) alters the system. Quantum physicists can explain this. 11/x

The third reason I believe my prediction was off is that I failed to account for regional differences. The US has marked diversity in terms of % vaxxed/boosted, prior infxns, mask use, housing situations, & more. (This is a limitation of living in 1 state for the last 25y.) 12/x

Many areas saw major disruption in some sectors, but it was more heterogeneous & more asynchronous than I predicted. Some had grocery shortages, some had pharmacy closures, some had teacher shortages, some had Nat'l Guard in hospitals. But not everywhere or all at once. 13/x

The fourth reason I believe my prediction was off was that the #omicron wave peaked & fell very rapidly. This pattern had been observed in S. Africa, but we were unsure if it would be replicated here or elsewhere. It was. This was a gift. We're lucky. 14/x

The fifth reason makes me😞. The US proved to be willing to tolerate far more #Omicron cases--& the tragedies they produce--to "keep the trains running on time" than I had hoped. We've lost 120K in US to Omicron in the last 2 mos, thousands every day. No words. 15/x

I'm sure there are more reasons. The process of error analysis is always long & arduous. But the point of this thread, for #MedTwitter, #MedStudentTwitter, & most of all the public is this: the point of science is NEVER to be right. Or it shouldn't be anyway. 16/x

The point of science is to find our way to the truth, or our best estimation of it. The scientific process requires us to look at data, form a hypothesis that represents our best guess, get more data, & evaluate whether we we guessed right or not. And then ASK WHY & LEARN. 17/x

So my biggest error of all was in expressing far more certainty than I should have about what was to come next with #Omicron. We've been repeatedly humbled in this #pandemic, & it's not over. Learn from that mistake so you don't make it yourself. It was a big one. /End