India awaits… the flattening of the COVID-19 curve

Prime Minister Modi does not seem to be his usual self of late. He has toned down his language vis-à-vis the opposition, calls for an all-party meeting and even personally talks to the key opposition leaders and the chief ministers. But the corona magic that has brought about this behavioural change appears to have landed him on the horns of an acute dilemma as well. What decision to take on 14 April when the first 21-day lockdown period would come to an end? Modi himself has hinted that the lockdown would be extended. By what duration, he didn’t reveal.

The situation is very dicey and the government itself doesn’t seem to have made up its mind yet. Whether the lockdown would be extended by another week or, by another 21 days? With the number of COVID-19 infections going up by around 400–500 every day between April 2-8, some experts are arguing that the COVID-19 infections are yet to peak, and lifting the lockdown sooner would disrupt whatever gains were achieved by the lockdown and strengthen the need for prolonging it. In an online interview, the WHO authorities too warned countries not to lift the lockdown prematurely. But clamping another 21-day lockdown would precipitate a crisis of a different order. The supply lines in the economy already stand disrupted and another 21-day lockdown would create hoarding and the famine-like situation in parts of the country. With millions of migrants stranded and unemployment predicted at above 20 – 25 crore, the human crisis would be very acute. It is indeed a Hobson’s choice between an acute health crisis and a severe socio-economic crisis. That probably explains the change in Modi’s postures.

Related news: Big B multi-starrer inspires Mamata to emulate similar COVID-19 film

Originally, the number of infections were expected to peak by the first week of April. The senseless Nizamuddin slippage and the spurt in some clusters in Maharashtra, Kerala, Delhi and Tamil Nadu altered the trend and it got extended by another week. Even though the curve is steadily going up by 400 – 500 cases by each day for about a week, the only silver lining is that the rate at which the number of new cases is increasing seems to have flattened somewhat. Experts expect the number of cases also to peak by this weekend after hitting the 15,000 mark. So in all probability, Modi might be extending the lockdown by a week. But the situation is still shrouded in some uncertainty with many unknown variables.

But merely continuing a nationwide blanket lockdown would not only give diminishing results but might turn counterproductive. It is high time to move towards a strategy of targeted containment focusing on high-risk areas and groups and localised lockdowns. They should be accompanied by more extensive testing, monitoring, tracking of contacts of cases testing positive, arranging home-delivery of medicines and essentials to the people, arranging sanitized transportation of migrants, and a bigger relief package to the poor and unemployed and so on. Yogi government in Uttar Pradesh has ordered district-level lockdown in 15 districts. Since coronavirus so far remains mainly an urban phenomenon, the UP government has virtually sealed off Kanpur. Lockdown might be lifted in zero-case districts. Such a changed approach might give better results provided Modi and his advisors do not fall for alarming reports based on mathematical modeling giving absurd results with mind boggling death figures.

What has added to the overall uncertainty and panic is the epidemic of mathematical modeling studies of the COVID-19 pandemic. Many of them — even some by reputed institutions — absurdly forecast death in millions based on a forward projection of sketchy pre-lockdown data. What opened the floodgates of such studies and one that probably even influenced Modi’s decision for a 21-day country-wide blanket lockdown was the study by the celebrated mathematical modeler Neil Ferguson. The study by his team done for Imperial College, London, dated 16 March 2020, predicted approximately 5,10,000 deaths in Great Britain and 2.2 million deaths in the USA in the worst-case scenario of no effective social distancing. Then a day before the lockdown, an India-specific modeling exercise by researchers from Delhi School of Economics, John Hopkins and Universities of Michigan and Connecticut came up with an estimate of 2.2 million cases in India by 15 May 2020 in the worst-case scenario of no intervention. Even if a 3% mortality rate is assumed, the death projections would be mind-blowing. Several other subsequent modeling attempts estimated deaths between 18,300 to 1.1 million by the end of May or June.

Related news: Depleting essentials: Villagers struggle on Indo-Bangla border amid lockdown

What are the limitations of this modeling method and what are the common mistakes and pitfalls associated with modeling? From The Federal, we approached a young researcher and mathematical modeler Nisha Chetana for some professional opinion. According to her, “A mathematical model is only as good as the assumptions it makes of its free parameters. The assumptions that we make on how a free parameter should behave, influences how the forecast of the model will be. In COVID-19 forecast models, the free parameter is not a single variable but a group of variables which are dependent on each other like social distancing, intervention methods by the government such as percentage of people under quarantine, percentage of viral shedding (i.e., departure of the virus from the host body cell after replication), percentage of viral load a patient is carrying etc. Most of the models are not taking into account these parameters as these are variables hard to quantify.”

“From a natural selection point of view, the virus needs a human host. The tendency of the virus is not to be so virulent that it will kill the host. Killing the host is detrimental to the virus itself because it won’t be able to multiply. So the virus is continuously mutating and changing according to its environment to keep up its reproduction. What this means is that the virus might mutate to be less virulent and evolve into something like a common cold/flu virus. This possible mutation of the virus is not taken into account at all in the models. The model assumes the virulence of the virus is constant, which is faulty,” she added. “The forecast that the models make assumes the rate of change is constant throughout, which is a huge mistake. The results of most of the models are wrong because the initial data on which they are based are sketchy and the underlying assumptions about the variables are wrong,” she concluded.

But not all modelling can be wrong. If intelligently deployed, they can be of great help in targeted containment. For instance, as the database on non-communicable diseases is available with the government, the elderly people in the population who are more vulnerable, especially those with co-morbidity of other diseases like diabetes, heart conditions and breathing disorders etc., can be identified and shielded by specially quarantining them. Other high-incidence target groups and the potential high-infection areas can be projected and specially monitored. It is high time the country moves towards such a targeted approach.

(The Federal seeks to present views and opinions from all sides of the spectrum. The information, ideas or opinions in the articles are of the author and do not reflect the views of The Federal.)