7/ In this, I will give an special commendation to @GidMK, which not only engaged with us; took the time to read it, even found a mistake in the seroprevalence numbers (which we corrected early on) which whom we discussed the work in detail and agree to disagree in a the IFR.
-
Show this thread
-
8/ The jury is still out on that one. We acknowledge it is a bit higher than we estimated, but not as high as suggested (Sweden 'the canary' bounds it). I would say I will call it a draw, because with more transmissible variants it's gonna be very messy to find out who is right.
2 replies 2 retweets 26 likesShow this thread -
Replying to @federicolois
You will be very interested in our forthcoming paper on IFR in LMICs, which includes data from Brazil, Argentina, Peru, Ecuador, Colombia, and a few other places in south/central America
2 replies 0 retweets 1 like -
Replying to @GidMK
Argentina is a shithole, we need an IFR double the one in Europe to account for our behavior. It is absurd... Peru if you use the 'excess death' adjustment they did, will just corrupt your dataset, please dont.
1 reply 0 retweets 1 like -
Replying to @federicolois
We are using confirmed+suspected where possible, but are including models looking at both only confirmed and excess deaths in secondary analyses
1 reply 0 retweets 0 likes -
Replying to @GidMK
Be careful with the 'suspected' category specially in Argentina. The data is very suspect, I have the benefit of knowing some people in charge of that data in my hometown and there are patterns that show that category is a catbag.
1 reply 0 retweets 0 likes -
Replying to @federicolois @GidMK
You cannot even trust the 'confirmed' here. Unless you have laboratory confirmation --- which we filter for. If you want I can check that data as I know it first hand.
1 reply 0 retweets 0 likes -
Replying to @federicolois
The same is true in many places - as the Zambian study showed, 'confirmed' deaths may undercount by a huge fraction, so we're looking at a the data in many different ways
1 reply 0 retweets 0 likes -
Replying to @GidMK
LOL, in here the most interesting part is that probably is overcounting it :D ... If you go and look at my hometown data and the city 70km away from it, you can see the different patterns in action.
1 reply 0 retweets 0 likes -
Replying to @federicolois @GidMK
What's different? The person in charge of filling the data. She resisted the request to do some things that were commonplace in other jurisdictions that contaminate the epidemiological records.
1 reply 0 retweets 0 likes
In general the much bigger issue is undercounting - the estimates for Mexico and India are that very large proportions of COVID-19 deaths have been missed
-
-
Replying to @GidMK
I dont doubt it, BUT that illustrate exactly why I said that. In the case of Argentina, Peru and from simple modelling exercise that I have done on Colombia. The most likely scenario is those location do not share that bias. That's why I am telling you to be extra careful.
0 replies 0 retweets 0 likesThanks. Twitter will use this to make your timeline better. UndoUndo
-
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.