we've all heard a lot about super-transmissibility of UK variant. Estimates were based on Volz, Neil Ferguson et al 2021. Here is right panel of their Figure 3 showing "additive advantage" of variant of >60% (horizontal line). medrxiv.org/content/10.110…. 
2/ Volz, Ferguson et al used data from epidemic weeks 46-49. Generating code for this figure and out of sample data (weeks 50-56) is available for at github.com/mrc-ide/sarsco…. Out of sample, super transmissibility vanishes as variant became dominant.
@BogochIsaac @NathanStall
@BogochIsaac @NathanStall
3/ the UK wave associated with the variant was real enough, but out of sample data counter-indicates the super-transmissibility hypothesis.
4/ Ferguson and colleagues are well aware of this information, but apparently haven't reported it. Here is a graphic from their github showing combined results for weeks 45-54: github.com/mrc-ide/sarsco…
5/ here's a graph showing progress of variant in UK. It has the same sort of Gompertz curve as "original" S-positive COVID. This can be produced without super-transmissibiliyty by "founder effect" in which variant got foothold in new "sub-market" and did what virus does.
6/ the convergence of S-negative (variant) transmission to S-positive (ordinary) transmission is evidence to me that data is better explained by a founder effect phenomenon, and contra super-transmissibility.
7/ one small comment on coding style which gave me bad impression of author competence. Here is a code section in which they calculate elements in a matrix one-by-one in do loops. There's no need to do that as it can be done easily and more transparently in one line.
8/ another point: authors do not provide links to the original original data, which presumably comes from UK public health authorities. I was unable to locate. So I used the Ferguson et al github compilation. It would be better if they archived data as received from public health
9/ on Jan 7, Ferguson article on UK variant reviewed for Public Health Ontario. Reviewer stated that Ferguson article "at this time should not be used to inform policy or decision-making". Nonetheless, article was repeatedly used for policy and decisions publichealthontario.ca/-/media/docume…
10/ on Feb 11, Ontario's COVID science advsory panel, Science Table, covid19-sciencetable.ca/wp-content/upl… projected explosive exponential growth in variant, showing Toronto as example reaching 50% of all cases by Mar 1, with full third wave by mid-March. This Triggered enhanced lockdown.
11/ What happened? As usual, scare projections didn't materialize. Total cases continued declining at same rate. Variant has thus far been negligible. Less than 3 weeks from projection, Toronto is at 50% of Feb 11 projections (with earlier projections even further off.)
12/ in next projections (Feb 25 covid19-sciencetable.ca/wp-content/upl…), Science Table continued to project alarm over variant but delaying prediction somewhat: on Feb 11, predicted variant at 50% by Mar 1; on Feb 25, predicted 40% by second week.
13/ so how close are variants to reaching 40%? Between Feb 15 and Feb 28, there were 226 (535-309) new cases of B.1.1.7 (UK) variant. In the same period, there were 15007 (301839-286832) total new cases. About 1.5%.
• • •
Missing some Tweet in this thread? You can try to
force a refresh
