And I’m referring to D.C., not Wor-Shing-Tun. Last week, the D.C. government released a ‘ COVID-19 situational update’ (pdf). This figure, which they recently re-released and made harder to read (though the key points are clear, I suppose…; p. 11) is very different from what has been said by most public health officials:
The key point is that the mortality rate will be at least double that of the model that the federal government is basically using* (the IHME model funded by Gates Foundation researchers), and could be five times as high. As importantly, the peak won’t be until early July, not mid-April. This discrepancy isn’t a D.C.-specific phenomenon (in fact, D.C. compared to many states started earlier with restrictions).
The discrepancy (and that’s an understatement) has to do with the differences between the model the federal government is using versus the model the D.C. government is using (the CHIME model). In other words, the peak in many places might be much later than currently predicted, and the death toll considerably higher. As to which model is better, it’s not clear. The CHIME model** is more in line with the Imperial College of Medicine Report, and the CHIME model uses more pessimistic assumptions about the benefits of social distancing (it assumes they will be less effective because they won’t be followed as much as people hope), while the IHME is far more optimistic about social distancing measures. (It’s also unclear what parameters D.C. used to reach these conclusions, so I can’t figure out how robust this estimate is).
All that said, there are implications for what D.C.’s use of a different model means about returning to work, especially in the D.C. area (hint: the federal workforce). If the seat of the federal government doesn’t think that the peak will hit until July (and presumably in other states too), then we better settle in for the long haul.
In the mean time, we can learn to never forget what Trump and his Republican enablers needlessly inflicted on the U.S. Anger is the appropriate emotion.
*Though who knows what they’re doing when you get right down to it.
**There are also differences in terms of the mechanics of the CHIME and IHME models, but it’s not clear to me what role those would play in the wildly different results.
mikethemadbiologist
I don’t see a peak in early July for the CHIME model. I see mid-May.
Friend and colleague has been following Massachusetts numbers, notes that IHME, optimistic as it is, appears to be pessimistic vs observations in MA. Unclear if we’re failing to properly classify pneumonia deaths w.r.t. COVID-19, or if half-assed social distancing actually gets the job done. (I’m here, it’s half-assed. Parent hauling two kids into the supermarket, none with masks, yeesh, pay no attention to the guy flattening his ass against the edge of the aisle as they wander past.)
But for example, new hospitalizations spiked on 3/26 (100+, I’m reading off a log scale), roughly flat since then, deaths also spiked 3/31 (30+, log scale), also roughly flat since then. So the model is self-consistent. But also, I have heard that hospitals are “full” — is that the source of the flattening?
And for 4/4 and 4/5, reported numbers were about 30, but the IHME chart claims we had 80 and 78 new deaths on those two days. WTF?