Over the past week (till September 24), the seven-day average of daily new cases of Covid-19 has fallen by around 7,000 in India. By itself, this is important for two reasons.
One, it has never happened before in India, although it has in other countries. The US (the closest comparison to India by size, geographical area, and number of Covid-19 cases) has seen two peaks (and two troughs), with cases on the rise again (the seven-day average is the highest in that country since early September). The Indian numbers are from HT’s database and the US ones from The New York Times’s. In contrast to the US, India’s numbers have never fallen – until now, that is.
Two, 7,000 is a significant number, whichever way you look at it, and already there have been a few seemingly analytical pieces celebrating the slowing of cases in India. The tone of some of these suggests that we can stop worrying — and start living.
In early September, it looked like it would only be a matter of days before India saw at least 100,000 daily new cases of Covid-19, but that hasn’t happened (which is definitely cause for some cheer). But it isn’t easy to translate these numbers into some intelligence on which way the pandemic’s curve is headed. There’s a reason why.
Over the two weeks to September 24, India’s Covid-19 numbers have not stuck to script , and show no discernible pattern. Until then, it was easy to predict what the number would be even a month down the line. Indeed, HT’s Covid-19 data specialist Jamie Mullick correctly predicted the dates when India would cross 1, 2, 3, 4, and 5 million cases. India ended September 24 with a total of 85,909 cases, and September 25 with 85,506. And most people have got their 6-million date prediction horribly wrong.
That’s because, over the past two weeks, India’s daily new case numbers have fluctuated, from a low of 74,693 on September 21 to a high of 97,932 on September 16. In the same period, the number of tests has fluctuated even more wildly, from a low of 731,534 tests on September 20 to a high of 1,492,409 (almost 1.5 million) tests on September 24. Not surprisingly, the positivity rate (the number of people testing positive expressed as a proportion of the tests conducted) has been volatile too – from a low of 5.7% on September 24 to a high of 11.9% on September 20. As the chart accompanying this column shows, even the movement within this range has been jagged, not unidirectional.
There has been no change in testing protocol or anything else that can explain this behaviour (or rather misbehaviour) of the numbers.
Sure, several states and Union Territories continue to rely on unreliable rapid antigen tests, which they continue to use in contexts where the use of these makes little sense, but they were doing this even in July and August. Regular readers of this column will know that this is one of my pet peeves — but my objection is based on science and math. Rapid antigen tests have a high proportion of false negatives — the identification of infected individuals as uninfected, and any state or UT using a lot of them is artificially (and disingenuously) pushing down the positivity rate, apart from allowing a lot of infected people to walk around (infecting others). Delhi, Uttar Pradesh, and Bihar continue to depend on these for the bulk of their testing.
My fear is that the home ministry, which is expected to issue new guidelines for what is called Unlock 5.0 over the weekend or sometime next week, may use this data as the basis. Sure, the pandemic may be slowing in India, but we have no way of arriving at this conclusion based on the numbers of the past two weeks.