04-01-2020, 04:35 PM
(04-01-2020, 02:40 PM)danbrotherston Wrote: @tom009 For methdology, are you counting tests when the results are released, or are you regressing the data back in time to when the test was sent or when symptoms presented?
Public health is presenting data in this way right now: https://www.canada.ca/en/public-health/s...cases.html
It helps correct for the bias of the testing backlog--and even more, for the backlog decreasing (which biases results towards now, even though they were earlier). Obviously, it also has limitations, that is incomplete information for the duration of the backlog, but in any case, the results are encouraging.
I'm reporting the summary data from Waterloo Region Public Health, the raw data is not available so it's not possible to do any more advanced analysis. (The same applies to Ontario and Quebec data.) I believe their numbers include both confirmed cases as well as presumptive (tests still in the backlog) cases.
And unfortunately today was the first time they reported the size of the backlog, so it's hard impossible to account for the effect of reducing that backlog.
On the national level, they do report the total number of patients tested and the number of positive/negative tests; the difference gives us the backlog. The number of tests per day is averaging around maybe 13-14,000; the backlog has stayed around about one day (but do note that the backlog in any given province may be either higher or lower than that).