Skip to main content

A female official was conducting a population census during the COVID-19 pandemic in Indonesia.

Photo credit: Unsplash/Muhammad Abdul Majid

A female official was conducting population census during the pandemic era in East Kalimantan, Indonesia

Sustainable Development Goal 3 focuses on ensuring healthy lives and promoting well-being for all at all ages. Age-specific mortality estimates are needed to measure many of those targets, but adult mortality estimates are especially lacking. This has become even more pressing as decreasing fertility has led to steadily increasing proportions of adults in the population. Over the last four decades the proportion of the population under age 15 in Asia has declined, in parallel to increases in the adult population (Figure 1). This is also evident in the Pacific and Latin America and the Caribbean.


Mortality in younger ages has successfully been reduced, with under-five mortality dropping globally from 93 deaths per 1,000 births in 1990 to 37 in 2023, according to UNICEF. Declining child deaths together with a greater proportion of adults in the population means a proportionately higher number of deaths at older ages. As child mortality and fertility both continue to decline, it is important to focus attention to preventing premature adult deaths to facilitate healthy aging societies.

Further, diversity and inequality in adult mortality patterns is frequent, with one study showing that these inequalities are greater in richer countries than poorer ones. To leave no one behind and to address these inequalities, universal and inclusive adult mortality estimates are needed, as well as the ability to disaggregate by geographical region, gender, age, cause of death and socio-economic status.

Figure 1: Proportion of the population by broad age groups in selected regions. (Data source: World Population Prospects 2022. Note: Trends to the right of the vertical dashed lines are projections.)

Civil registration and vital statistics (CRVS) systems can provide high quality data on adult mortality when they are comprehensive and inclusive. However, many systems are incomplete especially with regards to deaths. For example, rural populations are less likely to have their deaths registered. While two fifths of countries in Asia and the Pacific have near complete death registration, above 90 per cent (though this figure excludes China and Indonesia), it is important to recall that the remaining unregistered deaths may be among the most vulnerable and marginalized populations.

Even robust death registration systems may be overwhelmed during crises, when the need for reliable data is most acute. The COVID-19 pandemic revealed the stress on CRVS systems, needing to generate mortality estimates while facing acute pressure. Timely and accurate estimates of death are especially important during shocks, such as conflicts or pandemics, to ensure appropriate and rapid responses.

In countries with incomplete CRVS systems, capturing adult mortality, especially recent or finer granulated estimates, is more complicated. To fill this data gap, modelled data are often used. In countries with mortality data from censuses or surveys, models can impute the missing periods, often based on some assumptions regarding trends. With no country data, estimates from neighbouring countries are used to model adult mortality, averaging out regional diversity. Although modelling data helps identify stable mortality trends, models struggle to reflect country-specific factors (like large influxes of refugees), sudden shifts in premature deaths (like Covid-19), or variations within a country.

Empirical estimates of adult mortality in countries with incomplete CRVS are mostly based on census and survey data, which typically ask about the survival status of respondents' relatives, like siblings or household members.  This limits what we can learn about inequalities within countries since they rarely include additional questions about the characteristics of the deceased, such as socioeconomic status. Moreover, questions in surveys and censuses generally use a wide observation window, leading to outdated data, and can be unreliable due to poor recall of events by respondents. Surveys and censuses also generally do not capture causes of death, although some may include questions to identify maternal mortality or violent/accidental deaths. To compensate for this gap, supplemental measures like verbal autopsies are used to estimate adult mortality by cause.

Countries in Asia and the Pacic are strengthening death registration as part of their commitments under the Regional Action Framework on CRVS in Asia and the Pacific and the Asia-Pacific CRVS Decade. However, this demands time and significant investments. While building resilient CRVS systems and emphasizing death registration, adult mortality needs to be measured by other means to meet immediate needs. Thus innovative approaches like asking respondents about the survival status of members of their social network, or of their parents, or mobile phone surveys are required.

To that end, the United Nations and New York University Abu Dhabi organized an Expert Group Meeting in 2022, gathering specialists to discuss adult mortality measurement. This led to the formation of the Measuring Adult Mortality Community of Practice (MAM-CoP), fostering collaboration and experience and knowledge sharing amongst professionals in the field. These initiatives are crucial to better understand and track adult mortality, making sure that we get everyone in the picture and no one is left behind.

Print this article
Petra Nahmias
Chief, Population and Social Statistics Section
Ashira Menashe Oren
Consultant, Statistics Division
Statistics +66 2 288-1234 [email protected]
Data and Statistics