Thread by @gavreilly

THREAD: As we know there are now personnel and technological limits as to how many new 'confirmed cases' of Covid-19 will be reported every day - the backlog of cases to be officially announced is 9,000+ and likely to keep growing.

So how will we know how many cases there are?

(there isn't supposed to be an element of suspense here, my browser has crashed while posting the thread)

Great, the browser crashed and took the thread with it. 🙃

Basically the gist of it was:
• hospital figures are the most reliable, but they're laggy and only reflect the sickest
• daily case numbers are affected by the backlog in reporting and an artificial ceiling of how many cases can actually be notified to authorities each day
So…

• the best* indicator of how many cases are out there in the wild is from the number of swabs which yield a positive result every day. This is imperfect (people are sometimes swabbed multiple times; time lags between lab result and HPSC notification; etc) but a good barometer…

…and if you compare the cumulative positive swabs against the number of cumulative cases, since early November the excess has been stable at just over 6%.

In other words, for every 106 positive swabs read in labs, you can estimate roughly 100 new cases.

Even this approach has its limits - who knows whether the 6% rate remains stable in a third wave, and asymptomatic cases are no longer being automatically sent for testing - but for as long as there's a backlog in official notification, it's the best rule-of-thumb we have.

I had done graphs and everything.

An important point worth adding: this shortcoming in official data does NOT impact the treatment of individual cases, and doesn’t mean any other changes in how their contacts are treated, etc. Patients still treated the same as always; it’s only that we have no national picture.