Atithee wrote: ↑Mon Jun 22, 2020 2:21 amAre you surprised by the numbers, Jay? I’m not a bit. I’ve been tracking the doubling rate from your notes, and it has steadily fallen (got worse) over the past week. Not a good sign although I don’t think we will have terrible numbers when all is said and done. But, I’ll stick to my poll estimate of > one million. Now hoping that the second wave will not be catastrophic for anyone, not just for India only.
I am not surprised. I was just disappointed that it didn't go in the good direction I wanted
... After all, it fell fully in the middle of my prediction two weeks ago (420-440K).
In fact what just happened is very interesting. The growth in the previous week was pretty low, and so this week's rates look like it is higher, but all it is doing is to catch up with the rate we had before (which was slowing down, and will slow down because the daily additions are still linearly increasing, in my opinion). It was a slower linear increase in the previous week at about extra 150 cases per day but went up to around 450 cases per day this week. The reality is that the linear increase is probably still 300-400 per day like it always was, post-lockdown.
So what happened? The unassigned cases' adjustments in the last two weeks. Negative adjustments last week and positive this week.
The thing about those case are that the states (I think UP, MAH etc, but others too) were taking ownership only after the recovery period passed (14 to 21 days). Not by design, but because they can;t find the people who took the tests that got reported to ICMR
... Those are cases that ICMR knew to have been tested positive, but the person has left for the villages or simply left quarantine somehow. The state has no record, or has some address in some other state... But they can't reach the person, who may have given some wrong address and phone number to fool the state too. So you don't have any record, and the state cannot add them to their count. The state has no idea except that the test was done in the state, unless the person shows up at a hospital. Finally, after a while, the state simply takes ownership and writes them off as recovered because they were never in hospital or in the death records or anything.
So, once we reached about 17 to 21 days seeing the unassigned cases, the states started adding those as new cases and news recoveries! They tell ICMR that test-ID-number-XXX is their case. ICMR takes the number off from their list. The process of the state telling ICMR and then ICMR taking them off takes like a week (because this is not urgent activity). So, we got some double counting over a few days at the end of the first 20 days of unassigned cases, and then ICMR took them off. So for about 5-6 days last week (about Days #21-28 of theses unassigned cases), we had a drop in the numbers by about 150 per day. After probably having a rise of double-counted cases by 50 a day in the week before last too. Complicated numbers, I know... Bottom-line: Last week had a drop in cases. Now, we were on a catchup phase this week, and the unassigned cases will be growing at the rate they were doing before, I assume ( constantly adding new unassigned cases and taaking them off, but the sum would be positive, on average, on any day). So, back to normal rate of growth after this confusion, and we should see 300-400 cases of daily growth in the number added each day.
But are there signs that this linear rate of increase in the 300-400 range per day is getting worse or improving? Right now, not much of a sign. I am pretty sure that it is not a super-linear growth. That means our overall growth rate and doubling period will keep on improving. If the doubling period keeps increasing, we can plan and handle the pandemic, as far as healthcare is concerned.
But there will be no peak unless we see the linear increase in daily cases change to a sub-linear increase. 300 per day increase needs to come down to 150 per day and 100 per day. Let us wait and hope to see that soon.