Death data some succour, but testing is still inadequate and largely RAT-driven in some states; rural spread also growing
Given how we are not even sure the data is complete and reliable, an R value of less than one may not mean much.
The numbers may suggest India is seeing fewer deaths for every Covid-19 infection—1.5%—and also that the pace of new infections is coming off, to about 5.8%. But this is not reason enough to rejoice. For one, even today, testing levels are hopelessly inadequate, with much of it is being done via the less reliable rapid antigen tests (RATs). Add to this the fact that the public has just started commuting—via trains and metros—and that many parts of the country have chosen to remain under some manner of lockdown, it doesn’t look like we are anywhere close to containing the pandemic. Indeed, despite testing levels remaining low, the daily pan-India infection count is close to 85,000. The fall in the positivity rate, as an FE analysis found, is because testing has been scaled up in states like Uttar Pradesh, Bihar and Gujarat, which now account for 40% of India’s total tests and where 70-90% of the tests are done via RATs. This is what is skewing the numbers. In Bihar, for instance, the positivity ratio has plunged from 12.1% on August 1 to 0.4% on September 22 while in Gujarat, it has dropped from 10.5% on July 1 to 2.3% on September 22, thanks to a rising share of RAT tests. One should instead try to assess the situation by keeping in mind the trends in Tamil Nadu, which relies entirely on RT-PCR tests and is reporting a stable positivity rate of 8%. Again, Maharashtra’s positivity too is a high 20% while Karnataka’s, at 15%, is not low either. Also, infections appear to be steadily moving to the hinterland. Rural and semi-urban areas now account for 40% of the infections and for a fourth of the total deaths; a month ago their share was 20% while on July 27, it was only 13%. That should tell us something about how the pandemic is spreading; the number of districts with over 100 deaths has increased to 150 from 45, two months ago. While only one rural district had recorded 100 deaths on July 27, last Friday this had jumped to 15 districts. The short point is it is simply too soon to call a retreat of the contagion.
Given how we are not even sure the data is complete and reliable, an R value of less than one may not mean much. For about a month, Delhi was maintaining an R value of less than one, but that soon reversed, and last Thursday, the value was 1.26. Indeed, given how consecutive sero-surveys have shown infection levels that are 4-5% above those indicated by the daily data, it is apparent the R value isn’t telling us the whole truth. In fact, the level of infections, as indicated by the authorities, isn’t quite accurate because, unlike in the US where a confirmatory test is needed to declare a Covid-19 patient has been cured, ICMR guidelines have no such requirements and have left the states to their own devices. While the WHO guidelines stipulate that an infected person takes at least 14 days to recover, in India, a person recovers in just 11 days. The need for caution arises because the country is woefully short of healthcare facilities; last Friday, the occupancy for hospital beds in Delhi had jumped to 44% from just 27% a month back. This is no time to be complacent.