Skip to content

When you choose to publish with PLOS, your research makes an impact. Make your work accessible to all, without restrictions, and accelerate scientific discovery with options like preprints and published peer review that make your work more Open.

PLOS BLOGS Speaking of Medicine and Health

Why Every Body Counts

By guest contributors Satchit Balsari, Caroline Buckee, and Jennifer Leaning

In May 2022, the WHO reported that India had the highest COVID-19 death toll in the world.  The government of India protested, citing the official number that was one-tenth the WHO estimate. WHO’s estimate of 4.7 million deaths was in line with what academic researchers and journalists were finding by triangulating other sources of death data including news reports, death registries, and observations at crematoria. Today’s article in PLOS Global Public Health reaffirms the high death toll. We explain why the counting of deaths after public health emergencies is an important exercise.

Mortality ascertainment after pandemics and natural disasters generates knowledge that can help reduce deaths in future public health emergencies by examining who died, how, where and when. The public health community has saved lives by applying lessons learned from  examining the causes of deaths in past disasters to preparedness and planning for subsequent crises, After Cyclone Bhola resulted in a half million deaths in 1972 in then-East Pakistan, scientists from Cholera Research Laboratory in Dhaka, and the CDC in the US, conducted a published a seminal article in the Lancet, based on a household survey that showed that most deaths were attributable to drowning, and mortality was higher among women. This study influenced how Bangladesh prepared for subsequent cyclones, building storm shelters on stilts, installing early warning systems and undertaking widespread awareness campaigns – interventions that over the years have vastly reduced mortality from similar cyclones.

In 2005, a highly contested mortality estimation study in Iraq, concluded that the civilian toll in Iraq greatly exceeded official counts, and demonstrated the prolonged impact of infrastructural devastation on even the lives of survivors. In other words, counting the dead after the war helped enumerate the direct and indirect impact of war. This phenomenon – of prolonged impact on morbidity and mortality – can be observed after natural disasters in high income countries as well. We were part of the team that conducted a study in Puerto Rico, after Hurricane Maria caused widespread damage to the island in September 2017. When journalists questioned the government’s official death count of 64, the government prohibited access to death registers. Our team therefore conducted a household-based survey to estimate the death toll and found that there were likely over 4000 excess deaths in Puerto Rico. The number 4645 became a rallying cry for the people of Puerto Rico who felt that their loss had finally been acknowledged, and demanded an official audit. The Government estimate was eventually revised using the previously withheld registries to 2,975, ranking Hurricane Maria one of the deadliest hurricanes in US history.

The Puerto Rico study, like the Iraq study, found that excess deaths after public health emergencies persisted after the acute event, and included deaths from indirect causes like interrupted access to healthcare, disruption in utilities, loss of wages, disrupted transportation, and so on. The undercounting of such indirect deaths, especially in the pandemic, often explains the large discrepancies in mortality estimations.

Even the counting of direct deaths during the pandemic is not simple, as we describe in an earlier article in the JAMA, most excess deaths are likely to be from COVID-19 infection, but may not be correctly attributed to COVID-19 because of low awareness, evolving diagnostic criteria, limited and heterogeneous testing capacity, and unclear guidelines on recording these deaths on the death certificate A greater proportion of excess mortality, as time passes, is likely to be from the indirect effects of the pandemic that will vary around the world. During this pandemic, not only do people die from SARS-Cov-2 infection, but they also die from resultant complications and exacerbation of underlying pre-existing conditions. Others may suffer or die from the secondary impact of the pandemic on societal functioning. Prolonged lockdowns, economic downturns, loss of livelihoods, and food insecurity, all intersect to drive up morbidity and mortality during the pandemic. There is growing consensus in the scientific and public health community that it is important to count these “indirect” deaths, as they help us understand the overall impact of disasters on society.

For the past several months, researchers and journalists across India have repeatedly shown that all-cause excess mortality exceeds official death counts. A triangulated estimate by Abhishek Anand et al places this number between 3 and 5 million. Our own analysis  of data from nearly 90 municipalities across Gujarat reveals very high mortality in April 2021, even before the peak of the delta wave. Of note, the worst affected age group was 40-65 years old, in discordance with trends observed elsewhere in the world.

We used data made publicly available by the Reporters Collective, based on actual death registration data from approximately 90 municipalities in Gujarat. Our data show extremely high excess mortality rates even in the weeks preceding the peak of the Delta wave, for which data were unavailable. Our analysis reveals that the 40-65 year age group had the worst outcomes.

What we do know is that such uncertain times warrants greater access to data, not less. Data transparency and sharing has allowed the scientific community to learn a lot about a novel disease in a short amount of time. Without access to accurate numbers on testing and clinical outcomes, the medical and public health communities could not have made the strides it has in combating the pandemic. While purportedly there may be political and social reasons to shield these numbers, there is little scientific justification to do so. Counting every death is the first step. Disaggregation by age, gender, economic conditions, and location may also reveal biological patterns or may reveal structural inequities that risk the lives of more vulnerable populations. Such analysis illuminates fissures in societies the world over. In the United States, disaggregating mortality data illustrated how the pandemic had very different impact across race and class. In India, data access will help understand the differential impact of the pandemic on rural and urban India, its impact across class and caste divides which drive healthcare access, and importantly, on the strengths and limitations of the existing healthcare system.

Although death is not the only outcome of interest in this multifaceted disaster, it is, in the end, one of the most clear cut: who died, how, when and where? Professor Amartya Sen’s examination of mortality in the Bengal famine, for example, showed, that over half of famine related deaths occurred in 1944, after the food security crisis had abated, fundamentally challenging the then prevalent understanding of why people had died during the famine. In 2020, the government of India, briskly enforced lockdowns and school closures to flatten the curve. It also inadvertently suspended school meals, screening and vaccination programs. Millions lost their daily wages. Civil society in India mobilized thousands of oxygen concentrators, built hospitals and even ICUs. How did these interventions impact outcomes? What worked and what didn’t? What can citizens, corporations, and governments do better next time? These answers matter as they influence resource allocation: does India need more hospitals, more beds, faster access? Did telemedicine work? Did clinical treatments — however well-intentioned —  do more harm than good? All interventions have consequences, regardless of intent.

Where it comes to population health, India is in fact data rich, but information poor. Health data from hundreds of millions are routinely collected but seldom readily accessible to researchers or policy makers. Healthcare in India is however poised for dramatic change, with the citizenry galvanized to invest in health. In line with the government’s own vision for a digitized health data ecosystem, the time is right for a national data commons that can securely, safely and responsibly share these data with scientists, academics and policy makers so that the biological, social, economic (and political) determinants of the impact of the pandemic in India can be rigorously analyzed.

As the sub-continent celebrates its 75th year of independence from colonial rule, new goals and aspirations will be defined.  But how will the country know what to build without knowing what is broken? Can nations rebuild after crises without acknowledging what was lost?

About the authors:

Drs Satchit Balsari (@Satchit_balsari) and Jennifer Leaning are emergency physicians and global health faculty at Harvard whose research and teaching are focused on disasters and complex humanitarian emergencies. Dr Caroline Buckee is professor in epidemiology and co-led the Hurricane Maria Mortality study at Harvard.

Leave a Reply

Your email address will not be published. Required fields are marked *

Add your ORCID here. (e.g. 0000-0002-7299-680X)

Related Posts
Back to top