To describe this phenomenon, a member of NITI Aayog, Dr V.K. Paul, found an intriguing phrase to use. At a 22 May press conference, he said India had “stopped the progression of the virus.” That’s right: you and I see an inexorable increase, but Dr Paul pronounces that we’ve “stopped the progression”.
How and why would a senior government functionary use such language? Well, here’s the rest of what he said:
“Growth rate of new cases till 3 April was increasing at the rate of 22.6%, behaviour of the virus was exponential. After 4 April there is clear slowing of the growth and then it settled at 5.5%. This shows that the country stopped the progression of the virus.”
It’s true, the growth rate at the start of April was over 22%. It’s true, it has since fallen to about 5%. Thus it’s true, the rate of increase has shown a “clear slowing”. But let’s be clear: by no means has it stopped. What’s more, if the “behaviour of the virus was exponential” in early April — which it certainly was, with a 22+% growth rate—simply slowing that growth rate does not mean it is no longer exponential, as Dr Paul’s remarks seem to suggest. (As he has seemed to suggest before). Whether 22% or 5.5%, it’s exponential either way.
No doubt Dr Paul has his reasons to characterize the spread of the virus as he does. But let’s dissect some of the numbers.
The “exponential growth” bit is quickly explained. All that phrase means, as I wrote in a previous column, is “that the number of cases multiplies daily by a fixed factor that’s greater than 1.” That’s just what is happening with this pandemic. Taking Paul’s figures, the fixed factor in early April was 1.226—a growth rate of 22.6%—and in late May that had shrunk to 1.055 (5.5%). 1.055 is lower than 1.226, but it will still eventually outstrip any other kind of growth. That exponential reality is inescapable.
In the case of this virus, it means that if the exponential growth keeps up and we don’t either find a cure or build immunity to it, sooner or later we will all be infected. This is what “sooner” means: If the number of cases had continued increasing at the 22.6% of early April, covid-19 would have infected everyone in India by 6 June—tomorrow. That’s a frightening thought, so naturally you’re relieved that the growth has slowed. This is what “later” means: If the number of cases continues increasing at the current 5%, all of India will be infected by about the beginning of November. Not quite as frightening, at least when considered in May, but it says something about the real meaning of “slowing” the growth. That is, again: Absent a cure or immunity, it’s inevitable. Every one of us will even
tually get the disease, whether by June or November. So why slow the growth of the disease?
Because what the slower growth does is spread out the numbers. The caseload, if you will. If the higher growth rate had persisted, we’d have had many more infected people on any given day over the last two months than we actually did: many more who would have needed care, hospitalization and ventilators, besides, no doubt, cremations and funerals. Demand on that scale would have quickly overwhelmed our healthcare facilities. That hasn’t happened, because the lower growth rate ensured that we never had demand on that scale.
And yet, this should not be understood to mean that with the lower growth rate, we have or will have fewer cases overall. At a press conference on 24 April (and in tweets that the Press Information Bureau posted about the conference), Dr Paul said that if it wasn’t for the lockdown, “we would have had around 100,000 covid-19 cases by now.” If anyone believed that he meant we would—with the lockdown in place—simply not reach as many as 100,000 cases, that impression was belied on 18 May, when we did cross the 100,000 cases line. Which is further proof of how the numbers only spread out as the growth rate sinks.
There were other projections at that 24 April conference. One was in a presentation that Dr Paul made, titled “India tackles the covid-19 outbreak effectively”. The slide’s title was “New, cumulative cases, doubling rate: covid 19- India”, and it plotted new cases and the total number of cases for the previous month-and-a-half. No surprise, both trended upwards, the total cases exponentially so. But the slide also had a line labelled “New Cases (Trend)”, and projected this out till 16 May, then about three weeks into the future. Why only till 16 May? Because on that day, this line actually touched zero. That is, it followed the graph of new cases for the previous several weeks, followed that generally upward trend for a few days past 24 April, and then, on no evidence presented at the conference, suddenly plummeted to zero by 16 May. In other words, this slide claimed that from about 30 April, India’s new case count would start to drop and on 16 May, there would be no new cases.
What actually happened was, of course, different. The count of new cases was 1,801 on 30 April. Instead of declining, it rose steadily from there, reaching 4,794 on 16 May. Not zero. (Needless to say, it has continued to increase since).
At the 22 May press conference, Dr Paul tried to explain this drop to zero away as a “misconception”. “I never said the cases would be zero by a particular date”, he said, adding that the graph “was a mathematical trend line.” Well yes, that “trend line” wasn’t about the cases declining to zero — because it was about the new cases. And unmistakably, it projected that count to be zero on 16 May. No misconception there. But perhaps more dramatic still are other projections we have seen. On 11 April, the ministry of health and family welfare put out a graph of their own, with three dramatically different curves plotted on it. One said that “without lockdown and containment measures”, there would be 820,000 cases by 15 April; the second said that with “containment measures but no lockdown”, 120,000 cases by 15 April; and the third, that with “nationwide lockdown and containment measures (current)”, just the 7,447 cases of that day (11 April).
There’s the dubious ploy of projecting to 15 April for the first two— remember this was 11 April—but not for the third. This inflates the hypothetical gap between these numbers more than it already is — but leave that ploy aside for two others.
The 820,000 figure was, according to that same slide, based on an assumption of 41% daily growth. Only, since 9 March when we had just 50 cases countrywide, we have never had a day with anywhere close to 41% growth. What about the 120,000 figure? The slide gave no growth rate assumption for it, but a little arithmetic shows that it is based on an assumption of 27.5% daily growth. The only plausible source for this number is the four days immediately prior to the lockdown on 24 March, which saw the case count rise from 199 to 505, which works out to a daily growth rate of 26.2%. Not 27.5%, but close enough.
But why just those four days? Because they all saw relatively steep rises — 29.6%, 29.5%, 20.65% and 25.3% — that support the case the MoHFW was making for the lockdown. Had they chosen the previous seven days instead, they would have had to make do with a projected 15 April figure of 44,000; the previous 10 days, just 25,000. The 120,000 figure sounds considerably better — that is, if you want to show that without a lockdown, case numbers would have exploded — than either 44,000 or 25,000.
I realize this column has a lot of figures to absorb and I will lose my readers. That’s the lot of anyone who tries to verify numbers that are too easily thrown about. Still, here’s the takeaway. If while a coronavirus policy is being defended or justified, you hear phrases like “would have been a huge number of cases by now”, or “we have averted an enormous number of deaths”, tell yourself: trust, but verify.
Verify, because then you will have questions.
Once a computer scientist, Dilip D’Souza now lives in Mumbai and writes for his dinners. His Twitter handle is @DeathEndsFun