Thread by @vb_jens

New preprint with BC N501Y screening data is out! First time I have seen proper VOC data out in public. Data is old now, but let's take a look what we can learn about data up to week 9, so end of February. medrxiv.org/content/10.110…

Before we go there, the preprint also has this neat summary of the point-prevalence study that @PennyDaflos also wrote about. The data in the preprint is about the left branch, N501Y screening. Positive results are presumed to be VOC. In BC that will be B.1.1.7 in most cases.

There are also some (small) false positives. This shows 1% prevalence of VOC at the beginning of February. This is still news because our PHO has still not corrected this after repeatedly claiming prevalences was 0.1%, and did not correct this even when asked directly.

Our main interest in this preprint is the following table. It shows weekly N501Y screening data with proper dates and denominators. That's exactly what's needed to understand VOC growth. And how VOC and regular COVID have evolved.

Let's add this to our usual graph we have been using to guess at the BC VOC share using Ontario and Denmark as a guide. The trend derived from the data fits out guess trajectories pretty well, although it over-shoots a little. Data lags a lot, but what can you do?

Actually there is something we can do, there was a (public?) BCCDC presentation today where more recent data was presented. (And some older data got re-jigged a bit.) Let's make another curve with that newer data! That moves the curve down to the lower end of our estimates.

End result does not change much, the rise of VOC shares in BC follows the expected pattern. The virus apparently can't tell the difference between Ontario and BC after all. Let's check out the VOC cases derived from the (fitted) VOC share from above. OMG! Exponential Growth!

How exponential. Let's graph this on a log scale just to be sure. For good measure we should throw in points for the times when we had measures of the VOC shares and multiply that with that days case numbers. Yap, very exponential.

Let's look at all cases thus broken out into VOC and non-VOC. What's the story? Looks like the initial decline of regular covid slowed and turned (maybe) into a small increase in early February. And the explosive case growth we have seen? All due to VOCs!

The two grew at different rates. (Duh!) VOCs did not "replace" regular COVID, they added to it. If we had not failed to keep VOCs out, that red part would not have been replaced by the salmon colours regular COVID. It would not be there. At all.

Let's see how the official data in today's Situation Report is keeping up. It added data for week 11 and reports 861 extra VOC cases. At the same time N501Y screening found 1036 presumptive VOC in week 11. Sequencing is not keeping up, and it's just going to get worse.

I don't think the BCCDC presentation or the slides are publicly available (yet?). (And I haven't seen the whole thing either.) If others want the data I have seen, here is a link to the (limited) N501Y data I transcribed from the preprint and the talk. docs.google.com/spreadsheets/d…