Written by Atanu Biswas
With the advent of the Covid-19, academic journals and the media were flooded with a lot of contradictory predictions about the number of Covid-19s and the onset of different waves, with lots of readily available data. Most of it has been proven wrong in the end. Now there is a new trend – to predict additional deaths from covid, which is not initially verifiable and thus even bizarre assumptions can rarely be proved wrong. Really funny!
Although ‘excess death’ is not an alien concept. When Covid destroyed Italy in 2020, one of my Italian acquaintances wrote me an email: “A lot of people are dying, who have never been tested Covid-19, so they do not enter the official count.” Well, is it just the story of Italy? Probably not.
There is reason to believe that many covid-related deaths have not been recorded as ‘covid’ elsewhere – although this is not the fault of the authorities concerned. No country was ready for an epidemic. Due to the initial shortage of test-kits, overwhelmed healthcare systems and, of course, social stigma, such low reporting is inevitable বিশেষ especially if the associated disease is highly contagious in nature.
Although defining ‘excess death’ is a difficult task. Leading magazines and leading medical journals argued that those who had covid but could die from other illnesses in the same period should be excluded from the count. However, it is supported that those who died of preventable causes during the epidemic কারণ because the overwhelmed health infrastructure could not treat them-should be blamed. By the same token, the quarantine system can reduce deaths from accidents and occupational injuries, and social distance can reduce deaths from flu-like illnesses. These should also be kept in mind when estimating the number of cavities.
The World Health Organization’s recent estimate of more than 4.7 million Kovid deaths in India – about 10 times more than the official record – has sparked controversy. There are other assumptions though. A study by the University of Chicago even estimated that there would be an additional 6.3 million deaths in India during the epidemic by August 2021. In fact, a March 2022 Lancet article estimated that 18 · 2 million deaths worldwide. Dynamic estimates made by The Economist magazine show an additional 14.7-25.1 million additional deaths worldwide, 2.3-4 times the official number so far. The US Centers for Disease Control and Prevention estimates that there are more than 1 million additional deaths in the United States.
Such widely varied estimates are based on a variety of public and private data and a variety of models. Understandably, covid data from different sources have errors and biases of different dimensions and directions. And in general, it is a matter of an unknown amount of error to infer anything based on potentially biased data of unknown nature. Partial data is another serious problem. For example, if covid mortality in Maharashtra or Kerala is used to draw a pan-India picture, it would provide a highly biased estimate.
Data-driven decisions are not unique, we know. Interestingly, the Lancet article mentioned above used six models to estimate the expected mortality; The final estimate was based on an ensemble of this model. Well, can’t other models make different guesses? In general, no one knows how appropriate a model is. ‘Wisdom’ certainly plays an important role in choosing a ‘suitable’ model.
While this may sound surprising, the Covid-data may not be absolutely necessary for any country to estimate the additional death of the Covid. Public death information over the past few years may be sufficient if it is available and credible. A ‘suitable’ time series model could be created for the total number of deaths by 2019, keeping in mind the country’s trends and potential catastrophes over the years. Without Covid, the expected death toll in 2020 and 2021 could have been obtained from the planted model. If we subtract it from the total number of deaths in 2020-21, it will give an approximate number of covid. In fact, using covid-data for this purpose can lead to unknown levels of error. However, public deaths cannot be fully documented or disclosed in a timely manner, and may even be suspicious in any country. With so many contradictory and vastly varied estimates, finding the exact number of Kovid’s deaths can be like chasing the unknown into the unknown. And it will probably remain elusive, unless a credible large-scale survey is conducted across the country.
Yes, seven decades of history can be recalled. During the Bengal famine of the 1940s, India’s ‘Plan Man’ Prashant Chandra Mahalanobis conducted a large-scale sample survey of famine-stricken villages, to analyze and estimate the magnitude of the disaster and to make an estimate of the number. Affected people. The survey provided extremely useful results such as one-fourth (1.5 million people) of households who owned paddy land before the famine either sold or mortgaged the land in whole or in part and had almost economic status. During the famine 4 million people suffered. Well, only a credible survey like this can reveal the true Covid story. Although it is not easy, the shoes of Mahalanobis are still incomplete. Contradictory images of the Kovid-disaster will continue to create buzz.
The author is a Professor in the Department of Statistics, Indian Institute of Statistics, Calcutta.