You can add location information to your Tweets, such as your city or precise location, from the web and via third-party applications. You always have the option to delete your Tweet location history. Learn more
We can compare #COVID19 1% infection fatality rate with an expected >80% attack rate (R0 of ~3) to seasonal influenza with 0.1% IFR and yearly attack rate of ~10%. Simple math would put unmitigated spread as >80X worse than a typical flu season.https://twitter.com/trvrb/status/1253398325245603840 …
To be clear, these are numbers for NYC. Demographics and health system will affect IFR (can easily be 0.5% in other settings). Density / behavior will affect R0 (can easily be R0 of 2 / 50% attack rate). In these other settings, we then are "just" 25X worse than a flu season.
Without age-stratification, these discussions are more misleading than informative.
Yes. Age-stratification hugely important. But the 10X differential in IFR between COVID-19 and seasonal flu appears to be perhaps somewhat constant across age groups: https://github.com/clauswilke/COVID19-IFR ….
Where does the 0.1% IFR number for seasonal flu come from? According to this meta-analysis of 2009 flu, it's about 0.01%, a whole order of magnitude smaller. See Figure 3. Can it be true?https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3809029/ …
H1N1pdm was a significantly more mild virus compared to more typical H3N2 seasonal flu.https://twitter.com/trvrb/status/1253410056059547648 …
I get how expect attack rate is calculated from R0, but in practice don’t we observe that the attack rate is usually below this theoretical value for real world infections? (Btw I’m in the much-worse-than-flu camp, but I do want to be careful with our estimates)
Yes. Agreed. Even without drastic measures, people will behave differently and this will reduce R0. BTW, an SIR epidemic model would have R0 of 3 result in 94% of the population infected due to "overshoot" (https://twitter.com/CT_Bergstrom/status/1252077930286444545 …). I chose ~80% to be between 66% and 94%.pic.twitter.com/KBKWcOH5pd
By now we'd probably be able to identify an actual population where 80% had been infected and 1% had died for pop. mortality of ~ 0.8%, if these numbers were correct. This does not disprove your assumption, but it gives me hope that things are better than this.
Estimates of population fatality rate of ~0.6% in Bergamo with 70% attack rate: http://medrxiv.org/content/10.1101/2020.04.15.20067074v2 …
Question: For a population the size of the United States (330 million), does this suggest a potential death toll of 330m * 0.8 * 0.01 = 2.6 million? I feel like I don't understanding the attack rate. My (low confidence) interpretation is 80% of population will get COVID-19. ???
Yes, it does. Btw., that's basically the prediction Imperial College made back in mid-March, using much more sophisticated modeling. Nothing has fundamentally changed since then in our basic understanding of the fatality rates etc. https://www.imperial.ac.uk/media/imperial-college/medicine/sph/ide/gida-fellowships/Imperial-College-COVID19-NPI-modelling-16-03-2020.pdf …pic.twitter.com/yVUbged8Y6
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.
