1/ Updated data about #COVID19 variants. I fitted a multilevel multinomial model using data of 16 variants in 42 countries to compare the growth of different variants.
Spoiler: B.1.617.2 is still likely to grow faster than B.1.1.7.
Spoiler: B.1.617.2 is still likely to grow faster than B.1.1.7.
2/ B.1.617.2 is now estimated to have a 67% (95% credible interval: 56-77%) faster growth per week than B.1.1.7 (left).
Estimated growth advantage is somewhat heterogeneous between countries (right).
Estimated growth advantage is somewhat heterogeneous between countries (right).
3/ Extrapolation of these numbers for 42 different countries.
Left: all 16 variants
Right: 7 most prevalent variants
Top: linear scale
Bottom: logit scale
Left: all 16 variants
Right: 7 most prevalent variants
Top: linear scale
Bottom: logit scale
4/ Compared to previous estimates, the growth advantage of B.1.617.2 is now more in the range of what would be expected based on the growth advantage of more transmissible new variants over time.
https://twitter.com/HJWesteneng/status/1393583752832593920
5/ Source data: github.com/hodcroftlab/co…
• • •
Missing some Tweet in this thread? You can try to
force a refresh
