Great review in the @NEJM about epidemiologic models and their pitfalls. nejm.org/doi/full/10.10…
Pitfall #1: We don‘t know how many people have been truly infected. [1/4]
Pitfall #1: We don‘t know how many people have been truly infected. [1/4]
Pitfall #2: We remain uncertain about the extent of protective immunity, which greatly impacts the course of the epidemic.
Pitfall #3: What’s the extent of transmission and immunity among ppl with no or minimal symptoms (and children) [2/4]
Pitfall #3: What’s the extent of transmission and immunity among ppl with no or minimal symptoms (and children) [2/4]
Pitfall #4: It‘s very hard to measure and model contact rates between susceptible and infectious ppl nodt only under phys distancing but also in various reopening scenarios. [3/4]
Next time you read a paper describing an epidemiologic model, ask yourself “what would happen to the predictions if one of the parameters mentioned above were set at a different value? Have they done sensitivity analyses on these changing values?”
@threadreaderapp unroll
