Science progresses by making and testing hypotheses. Often principles, practices or policies which look sound at one level get falsified at later stage when examined with more data. Rethinking and re-examining well accepted norms is an important part of science, failing science becomes a religion. In March 2020, when the threat of the pandemic was suddenly realized, it was inevitable to suggest and implement a set of measures to arrest the transmission and things that looked logical that time were suggested and implemented. As time passed, data started pouring in and in the light of the data it was necessary to re-examine what was working and what not. It was stupidity to expect that the same things would work all over the world given the wide variety of contexts and conditions. So according to data getting accumulated, the policies should have been re-examined, refined to suit local contexts and withdrawn if found useless. This process should have started by May or June 2020 itself as data started accumulating.
Although delayed substantially, the process of introspection appears to have begun now, at least in a small way. Two papers last week appearing in NEJM and BMJ reexamined school closure as one of the measures to arrest the epidemic. The study in Swedish schools that did not shut down shows no evidence of higher rates of illness among students and teachers in schools that remained open. The other Belgian study says that the negative impact of school closure on kids appears to outweigh the weak evidence that closing the schools might have actually reduced transmission. We need more such data based introspections and re-appraisals of every policy. This analysis might help us at least the next time a new respiratory virus appears.
I am examining below the possible impact on the spread of infection of implementing lockdowns, lifting of it and undue crowding of people without taking necessary precautions. We have ample opportunities to examine the actual impact of crowding without following the necessary norms. In October and early November there were elections in Bihar which were accompanied by large public meetings, rallies and celebrations by the winning candidates. In the following month there were elections in Jammu and Kashmir which in some areas showed record voter turnout. From late November farmer agitation started on an unprecedented scale with huge processions and gatherings which have continued for three months now.
An epidemic follows a typical shape of the curve when plotted daily or cumulatively. If lockdowns really worked we should see at least a temporary downward shift in the slope, especially on a log scale. Conversely whenever restrictions were lifted or there was large scale crowding, there should be a local upward movement of the curve. Here are the time trends plotted along with the dates on which restrictions were imposed or lifted, election campaigns, rallies, big meetings or celebrations were conducted. The upper curves are cumulative incidence on log scale, lower curves daily new cases (7 day running average) on linear scale. All data taken from https://www.covid19india.org/
The nation wide trend: See that lockdown did not decrease the slope and the unlock or crowding events did not increase it. The peak in the daily incidence trend and the following downward trend appears to be independent of all these events.
The Kisan Andolan should have theoretically been the ideal occasion for spreading the infection. But there appears to be no increase in the incidence in Delhi, nor in Punjab and Haryana (below).
Overall, the shape of the cumulative trend appears almost unaffected by any of these events when examined at national or state level. The daily trend has many ups and downs which have no association with the events of crowding and norm violation, even after allowing realistic time lag. So in regional or national time trend, there is no evidence of either the lockdown having improved, nor of crowding and norm violation having worsened the situation.
An epidemic is a complex phenomenon and just owing to its complexity, it’s no surprise that some simple minded solution might have failed to work. If some policy did not work as expected, it should not be taken to mean ignorance, stupidity, failure or lack of foresight of the policy maker. We just need to accept that this measure did not work and plan alternative measures that are likely to work better. But the inability to accept that something failed at least in some specific context, is the sign of being dogmatic. If lockdowns have not worked in the Indian context so far, it makes no sense to have them again if faced with some local surges. Saying that it worked in New Zealand has no relevance. Inability to learn from data is a sign of not understanding science and not being ready to use it appropriately. The top level Indian science organizations and health authorities need to give at least some proof that they also learn from data and rethink, re-examine and refine the policies in response to observed data. Otherwise there will be no difference in the approach to build a Ram temple and the approach to curb the epidemic. One is clearly a religious issue, but the other is expected to be scientific. So let the latter be evidence based rather than faith based.