Main issue is that we don't know. The best estimates I have seen hover just around 1% with high variability across age groups. In Hubei (with overrun hospitals) it's ~2.5%. Main issue - it seems to be 5x - 10x seasonal flu (but personal risk still *very* low).
-
-
Als antwoord op @K_G_Andersen @drsanjaygupta en
Right, we have NO idea of the denominator (N), especially in the US where so little diagnostic testing has been done. If could easily turn out to be closer to flu as we get a handle on N.
7 antwoorden 2 retweets 40 vind-ik-leuks -
Als antwoord op @EricTopol @K_G_Andersen en
Updated South Korea
#COVID19 fatality rate is 0.52%, via@HelenBranswell@statnews (yes it can go up, but a lot better than 2%) https://twitter.com/HelenBranswell/status/1234335439508688896 … [seems like the best denominator data available right now globally]pic.twitter.com/8NEQ4RdEYK
2 antwoorden 2 retweets 13 vind-ik-leuks -
Als antwoord op @EricTopol @K_G_Andersen en
I think it's far too early to say anything about their CFR. From symptom onset to death is 2-8 weeks, on average,
@WHO says. The vast majority of the South Korean cases have been diagnosed in the past 2 weeks. They haven't been sick long enough for us to know outcomes.4 antwoorden 9 retweets 52 vind-ik-leuks -
Als antwoord op @HelenBranswell @EricTopol en
Those numbers are unfortunately not adjusted for the epidemic curve, so 22 deaths at this stage is high. Once all outcomes are known this would suggest a CFR in the 3-5% range, which is similar to China. Good explainer here: https://www.worldometers.info/coronavirus/coronavirus-death-rate/#correct …
2 antwoorden 11 retweets 32 vind-ik-leuks -
Als antwoord op @K_G_Andersen @HelenBranswell en
The issue is that CFR is not deaths (dayX) / cases (dayX), but rather deaths (dayX) / cases (dayX-T) with T being average time to death (6-7 days for COVID-19). A week ago S. Korea had ~500 cases, so it becomes 22/500 ~ 4.5% after correcting for new cases.
4 antwoorden 9 retweets 36 vind-ik-leuks -
Als antwoord op @K_G_Andersen @HelenBranswell en
* Note: 6-7 days for T is very conservative - the real number is likely higher, which would drive the rate higher. However, all of this assumes consistent (and unbiased) testing, which is clearly not the case and would likely lower the rate.
2 antwoorden 3 retweets 13 vind-ik-leuks -
Als antwoord op @K_G_Andersen @HelenBranswell en
@K_G_Andersen - what are expected numbers by next week? 7 days seems to be a bit high, no?1 antwoord 0 retweets 0 vind-ik-leuks -
Als antwoord op @manuelrivascruz @HelenBranswell en
T in this case is "time to death", not # cases. Given the doubling times we have seen in other countries - 6-7 days - we'd expect ~8-9k cases next week assuming we still see exponential increase - and that's a big if. My best estimate is ~6.5k in a week - but that's a pure guess.
1 antwoord 0 retweets 1 vind-ik-leuk -
Als antwoord op @K_G_Andersen @HelenBranswell en
Given the epi curve correction, the numbers from Italy look quite grim. 229 confirmed cases on February 24th and 34 deaths reported today?
1 antwoord 0 retweets 0 vind-ik-leuks
Italy failed to detect cases early so there's a bias towards more severe cases in the beginning - hence the CFR is inflated. It'll likely be lower as time goes by. Again, trying to estimate any type of CFR at this stage is fraught with issues and estimates are unreliable.
-
-
Als antwoord op @K_G_Andersen @HelenBranswell en
BTW, do we have numbers of infected for children ? Say < 10 years old? Given the epicenter was Wuhan w low child rate, exposure may not have been as high as May be in US? Thoughts ?
1 antwoord 0 retweets 0 vind-ik-leuks -
Als antwoord op @manuelrivascruz @K_G_Andersen en
Thinking of differences in population pyramids. The absolute number of cases at age group may not be sufficient to confidently determine whether infection and death rate is actually lower ?
0 antwoorden 0 retweets 0 vind-ik-leuks
Einde van gesprek
Nieuw gesprek -
Het laden lijkt wat langer te duren.
Twitter is mogelijk overbelast of ondervindt een tijdelijke onderbreking. Probeer het opnieuw of bekijk de Twitter-status voor meer informatie.