Every death tells a story, but for demographers and policy-makers, those individual stories only become useful when they’re counted, categorized, and compared. Mortality data is the backbone of public health planning, social welfare schemes, and even insurance economics. Yet collecting accurate death data across a country of 1.4 billion people is far from simple. From decennial enumerations to continuous registration and large-scale household surveys, India relies on a layered system to track who dies, when, where, and why. Understanding these sources helps make sense of the headlines about infant mortality, life expectancy, and disease burden that shape national priorities.
Table of Contents
- Why mortality data matters in population studies
- The Census: A foundational but indirect source
- Strengths and limits of Census-based mortality data
- The Civil Registration System (CRS): Recording every vital event
- The completeness problem
- The Sample Registration System (SRS): India’s mortality workhorse
- How the SRS sample is designed
- Why the SRS is trusted
- Demographic surveys: NFHS and beyond
- What surveys add to the picture
- Challenges in mortality data collection
- Gender and age disparities
- Cause-of-death uncertainty
- Coordination and timeliness
- Looking ahead: Towards a unified mortality data system
Why mortality data matters in population studies
Mortality data is more than a grim statistic. It feeds into nearly every demographic indicator that matters: life expectancy at birth, infant mortality rate (IMR), maternal mortality ratio (MMR), under-five mortality, and crude death rate. These numbers guide where hospitals are built, how vaccines are distributed, and how state budgets are allocated. Without reliable mortality data, public health is essentially flying blind.
The challenge is that no single source captures the full picture. Each system has strengths and weaknesses, and demographers often cross-check one against the other. Broadly, India relies on three primary sources of mortality data, supplemented by large-scale demographic surveys.
The Census: A foundational but indirect source
The decennial Census of India, conducted every ten years since 1881, is the country’s oldest continuous data exercise. While the Census is not primarily designed to capture mortality, it plays a crucial supporting role. It provides the denominator, the total population, against which death rates are calculated. Without an accurate population count, mortality rates become meaningless ratios.
The Census also collects information that helps estimate mortality indirectly. Questions on the number of children ever born versus children surviving allow demographers to estimate child mortality using techniques like the Brass method. Similarly, household questions about deaths in the preceding year have historically been used to derive crude death rates, though with mixed reliability.
Strengths and limits of Census-based mortality data
The biggest strength of the Census is its sheer scale. It covers every household in the country, which means small areas and marginalised groups are theoretically captured. However, because the Census happens only once in a decade, it cannot track year-to-year changes in mortality. It also suffers from recall bias, since respondents are asked to remember events from the past, and from underreporting of deaths, particularly of female infants and elderly women. For these reasons, the Census is treated as a baseline rather than a real-time source.
The Civil Registration System (CRS): Recording every vital event
The Civil Registration System is meant to be the ideal source of mortality data. It records every birth, death, and stillbirth as a continuous, legally mandated process. The system in its current form is governed by the Registration of Births and Deaths Act, 1969, which made registration compulsory across the country. The Registrar General of India coordinates the system at the national level, while implementation rests with state governments.
When complete, the CRS provides individual-level information on date, place, sex, age, and, where medically certified, the cause of death. This makes it potentially the richest source of mortality information. It also serves a legal purpose: a death certificate is required for inheritance, insurance claims, pension settlements, and many bureaucratic procedures.
The completeness problem
Despite being legally mandatory, registration of deaths in India has historically been incomplete. The good news is that the situation has improved dramatically. Researchers have estimated that completeness of death registration in India rose from 58% in 2000 to 81% in 2018, with notable improvements even in poorer states. However, the same study found that female death registration lagged behind male registration, and states like Bihar continued to show very low completeness.
Another challenge is the Medical Certification of Cause of Death (MCCD) scheme, which is largely confined to urban hospitals. Most deaths in rural India occur at home and are never medically certified, so the cause-of-death data from the CRS is geographically skewed.
The Sample Registration System (SRS): India’s mortality workhorse
Because the CRS was incomplete for decades, the Government of India launched the Sample Registration System in 1964-65, with full-scale operations from 1969-70. The SRS was conceived as an interim arrangement until the CRS matured, but it has become the single most reliable source of mortality and fertility indicators in the country.
The SRS is a large-scale demographic survey based on a dual recording system. A part-time enumerator continuously records births and deaths in a sample unit, while an independent supervisor visits every six months to retrospectively record the same events. The two lists are then matched, and any discrepancies are resolved through field re-verification. This matching mechanism is what gives the SRS its high level of completeness.
How the SRS sample is designed
The SRS currently operates in roughly 8,800 sample units spread across rural and urban India, covering close to 7.9 million people, or about 0.6% of the national population. The sample frame is revised after every Census to remain representative. In rural areas, a sample unit is usually a village or part of a large village, while in urban areas it is a census enumeration block.
The SRS publishes annual bulletins and detailed statistical reports containing key indicators such as crude birth rate, crude death rate, infant mortality rate, neonatal mortality rate, under-five mortality, and maternal mortality ratio. These are available separately for rural and urban areas and for major states. Since 2001, the SRS has also used a structured verbal autopsy instrument to ascertain causes of death, which has substantially improved cause-specific mortality estimates.
Why the SRS is trusted
The SRS is widely considered the gold standard for Indian mortality data for several reasons. First, the dual recording system ensures very high completeness, much higher than the CRS in most states. Second, it produces annual estimates, allowing policy-makers to track trends. Third, its sample is large enough to produce reliable state-level estimates for bigger states. Many international organisations, including the World Health Organization, rely on SRS data for India.
Demographic surveys: NFHS and beyond
While the Census, CRS, and SRS form the backbone of mortality data, large-scale demographic surveys add depth, especially for child and maternal mortality. The National Family Health Survey is the most important of these. Conducted by the International Institute for Population Sciences (IIPS) under the Ministry of Health and Family Welfare, the NFHS has had five rounds since 1992-93, with NFHS-5 covering 2019-21.
The NFHS does not record deaths in real time. Instead, it collects detailed birth histories from women aged 15-49 in sampled households. From these histories, demographers estimate infant mortality, neonatal mortality, under-five mortality, and a range of related indicators for the five years preceding the survey. Because it interviews mothers directly, the NFHS captures deaths that may have escaped both the CRS and the SRS.
What surveys add to the picture
Surveys like the NFHS add three important dimensions. First, they connect mortality to socioeconomic determinants such as the mother’s education, household wealth, and access to antenatal care, allowing researchers to study inequalities. Second, they provide district-level estimates from NFHS-4 onwards, which neither the SRS nor the CRS can do reliably for most states. Third, they capture context, like nutrition, anaemia, immunisation, and water and sanitation, that helps explain why mortality differs across regions.
Other relevant surveys include the District Level Household Survey (DLHS), the Annual Health Survey (AHS), conducted in nine high-focus states, and the Longitudinal Ageing Study in India (LASI), which is critical for understanding mortality among the elderly. Specialised data sources such as the Million Death Study, which integrated verbal autopsy with SRS households to study causes of about a million deaths between 1998 and 2014, have transformed our understanding of disease-specific mortality, especially for non-communicable diseases.
Challenges in mortality data collection
Even with multiple sources, mortality data in India faces significant gaps. The biggest issue is the rural-urban divide. Urban areas typically achieve much higher levels of death registration than rural areas, and medical certification of cause of death is overwhelmingly an urban phenomenon. As a result, cause-of-death data tends to over-represent urban disease patterns.
Gender and age disparities
A second concern is gender bias. Female deaths, particularly of infants and elderly women, are systematically underreported in both the CRS and household recall in the Census. Cultural factors, lack of legal incentive (fewer property and pension claims for women), and patriarchal household structures all play a part.
Deaths at the extremes of age are also harder to capture. Neonatal deaths that occur at home, often within the first 24 hours of birth, may not be reported at all if the birth itself was not registered. Similarly, deaths of the very elderly in rural areas, where there is no formal medical attendance, often go unrecorded or are recorded with vague causes like “old age”.
Cause-of-death uncertainty
Determining why someone died is harder than recording that they died. A large share of deaths in India are still classified with ill-defined causes, which limits the usefulness of mortality data for disease-control planning. Verbal autopsy methods have helped, but they require trained physician review, and the quality of physician coding varies across states.
Coordination and timeliness
Different sources also report different numbers for the same indicator. The NFHS estimate of infant mortality is often higher than the SRS estimate for similar years, and the CRS, when adjusted for completeness, sometimes gives yet another figure. This discrepancy is not necessarily a flaw, since each source measures something slightly different, but it does create confusion in policy debates. Coordinating these sources, and gradually moving towards a CRS-based vital statistics system, is one of the long-term goals of India’s statistical infrastructure.
Looking ahead: Towards a unified mortality data system
The future direction is clear: strengthen the CRS so that it eventually replaces the SRS as the primary source of mortality statistics, while continuing to use surveys like the NFHS for socioeconomic depth. Digital initiatives like the revamped CRS portal, integration with Aadhaar-linked civil records, and electronic medical certification of cause of death are steps in that direction. Some researchers argue that the CRS is already more reliable than the SRS for measuring adult mortality in several states, suggesting that the transition is closer than it appears.
For students of population studies, the takeaway is that mortality data is never just a number. It is the product of a system, a system shaped by law, administration, technology, and social attitudes. Reading a mortality statistic critically means asking where it came from, how it was collected, and what it might be missing.
What do you think? If you had to design a single, unified mortality data system for a country as diverse as India, which existing source would you build on, and what trade-offs would you accept? And how should we weigh administrative records like the CRS against household surveys like the NFHS when they give different numbers for the same indicator?
Leave a Reply