New Delhi: A study published in Science Advances on July 19 (Friday) has claimed that the death toll in India during the first wave of Covid-19, that happened in 2020, was eight times higher than the official figure of 1.4 lakh. The study also says that the impact on life expectancy varies across socio-religious categories, with Muslims, Dalits and Tribals the worst-hit, far more than all the other groups.
Although several attempts undertaken to study the excess death estimates due to the pandemic in India have unanimously said that the government has grossly underestimated the number, experts have pointed out that the factor of eight, as claimed by the recent study, has to be interpreted with caution.
All previous such studies had estimated that the factor by which the Indian government undercounted deaths for 2020 was three-six, including the ones made by World Health Organisation (WHO), though their data sources were different as compared to this study. In other words, all these studies said the Indian government undercounted deaths by three-six times, depending on the study, thus leading to a significant difference in the multiplication factor between them and the latest one, which states the factor at eight.
Gautam Menon, dean of research at Ashoka University told The Wire, “No one disagrees with the fact that deaths were undercounted although my own leaning would have been towards an undercounting of between two and five in the first wave.”
Since all the other results in the latest study – loss of life expectancy and the very extent of caste and gender disparities as far as impact of pandemic is concerned – are based on the same subsample from which excess deaths were estimated, they too need to be read cautiously, two domain experts with whom The Wire spoke, suggested.
The authors of the latest study, Aashish Gupta, a scholar at University of Oxford and others have used the data generated in the second phase of India’s National Family Health Survey (NFHS-5) to arrive at their conclusion.
This data was collected for the NFHS through interviews conducted to study the mortality in 2020 as compared to the previous years – the non-pandemic ones. And the NFHS found that the number of deaths in a few months of 2020 (the pandemic year) were 17.1% higher than the same months of 2019 (the non-pandemic year). This difference between a pandemic year and a non-pandemic year is referred to as ‘excess deaths’.
The second phase of data collection for NFHS was done in 14 states and union territories. The authors of the current paper, say, the data from these many states was extrapolated for the entire country, and they found out that 1.19 million (11.90 lakh) deaths took place in 2020. This, in turn, was eight times higher than the official estimate of the government of India. It was also 1.5 times higher than what the WHO calculated for the year 2020.
Noted demographer, Prabhat Jha, chair in Global Health and Epidemiology at University of Toronto, said the sample size captured during the second phase of NFHS – which was, in turn, extrapolated for the entire country may not be adequate to make calculations for undercounting.
Jha, who is known his work on death estimations worldwide, and who has himself conducted a study on undercounting of deaths during Covid-19 in India, added, “using just one year of baseline [of 2019 for making a comparison with the pandemic year of 2020] is not ideal.”
Gupta, an author of the paper, contended that the sample size was large enough to allow them to make the claims made in the paper. “Small sample sizes do contribute to larger confidence intervals [a metric casting doubts over results],’’ he said, adding that it was not the case with his paper.
Gupta said he and other authors had verified their results in “several ways,” including looking at average mortality of 2018-19 for the sake of comparison with 2020.
It is pertinent to mention here that unlike Gupta’s paper, which estimated that number of deaths in 2020 were 17.1% higher than 2020, all previous estimations suggested the proportion was nearly 9% – thus leading to a significant difference in the estimation of the extent of undercounting by the Indian government.
Jha’s paper found 0.6 million excess deaths took place between June 2020 and January 2021 when compared with similar months of previous year – thus accounting for 9% relative ‘excess’.
The WHO’s estimate also pointed out that 0.8 million extra deaths took place in 2020, 0r 9% more than the previous year.
The bottomline, however, Jha and others stressed remains the same — the Indian government has significantly underestimated deaths during the pandemic years.
Government’s allegations worse than defence
The Union India may well dispute the extent of undercounting, as enumerated by the latest study, but India remains a few of the very rarest countries in the world to claim that its systems had not done undercounting at all.
As with all other previous studies done on this subject, the health ministry on July 20 (Saturday) not only dismissed the latest study but remained adamant that it hadn’t missed counting deaths during the pandemic as its systems were ‘robust’ enough.
Some of the arguments the government put to defend its stance vis-a-vis the latest study were the same it had used to reject estimations of WHO, Jha, a study published in the Lancet, and several others. At one point, the government reportedly also tried to stall publication of one such report. Most of these studies pointed out India had the highest mortality in 2020 and 2021.
One of the things the government claimed this time was that its civil registration system (CRS) is sound enough to record, or register, more than 90% of all deaths that happened in 2019 and in the later years. This is in contradiction to some of its own studies like NFHS, which say, nationally this proportion is only 71%. In other words, about 30% deaths do not make into the records of the government – thus indicating that it might have missed recording a high number of deaths during Covid-19.
Moreover, even if the higher number of death registration of the government is to be considered, the fact remains that this higher proportion of registration of deaths is not at all evenly spread in the entire country. For instance, in Bihar only 37% deaths get registered, and in Uttar Pradesh, only 47% do.
The government also said in its July 20 statement that another metric – crude death rate – as documented through sample registration system (SRS) – had shown there was no excess mortality in 2020, in the pandemic year, as compared to the previous years.
Jha said it was a flawed statement. “For all ages, the declines that were occurring steadily in crude death rates [till 2019] stalled and rose particularly in [2020] in urban areas. This is consistent with the [fact that] 2020 viral wave mostly hit urban areas,” he added.
Crude Death Rate, 2020
Mortality rates in children, 2020. Source: SRS 2020 report
While the government is unprecedentedly quick in rejecting each and every excess death study related to Covid-19, it has been sitting on crucial data sets which could have put a full stop on all such conjectures. This includes at least three years of data, starting 2021, for CRS and SRS – both crucial sources to study deaths in India, as Jha and others pointed out. Such a delay in release of these crucial datasets has hardly been recorded before.
It is pertinent to note that the latest study has made claims only for 2020, which the government has sought to dismiss. But the much deadlier wave happened in 2021.
The release of the CRS and SRS reports for, at least, 2021 could have resolved all the controversies related to Covid deaths. Alternatively, the data could have also revealed that most of these independent excess death estimates were not far from reality.
Jha said another way to resolve all these controversies is to add one simple question in the upcoming census.
“Since January 2020, has there been a death in the house? If yes, then capture age, sex and date [at least month] of each death,” he said.
The census already captures information on deaths but only for the last one year. “By changing this question to last five years and recording [the] date, India would have a direct number of excess deaths. That would be the best quality information of any country in the world,” he added.
Life expectancy differentials across socio-religious categories
Gupta’s paper also said there was a higher loss of life expectancy in females (as compared to males) and the marginalised sections like SCs, STs and OBCs due to the pandemic. Among religious groups, Muslims bore the brunt, the paper said.
Overall, the paper suggested that life expectancy declined by 2.6 years. Relative to a decline in life expectancy of 1.3 years for high-caste Hindus, who are privileged in Indian society, the loss for Muslims was 5.4 years, for STs was 4.1 years, and for SCs was 2.7 years.
Also read: Report on Rare Adverse Side Effects of Covishield Causes Panic. But Should It?
However, Jha mentioned, “[Evidence related to these religious, gender and caste] subsections also requires a cautious interpretation [since all of this was based on the NFHS subsample of 14 states].”
A former Indian Statistical Services official said that, in principle, he agrees that marginalised sections may have faced worse outcomes than others during the pandemic. “There has been a historical trend of poor access to healthcare [in the case of] the marginalised sections, and Covid-19 would have exacerbated it,” he said, wishing anonymity. The pandemic is no outlier, he added.
Menon also agreed with this. “This implies that we must pay more attention than we do now to the question of social determinants of health in pandemic situations,” he said.
Many studies have documented this in the past. For instance, this 2023 article published in one of the journals of the ‘Sage’ group pointed out that India is lagging in achieving equity in healthcare as the poor and marginalised are deprived of it. Similar facts have been recorded in several news reports.
Similarly, there has been a lot of documentation (like here and here ) about females getting relatively poor access to healthcare as compared to males. There has not been an abundance of literature pointing towards large-scale deprivation of healthcare to Muslims but some reports have pointed this out for specific services like antenatal healthcare.
Note: An earlier version of this article is updated with Gautam Menon’s quote added at 10.50 pm (IST) on July 21, 2024.