Andrea Verhulst

INED research fellow Andrea Verhulst studies mortality among children under 5 across the world and excess perinatal mortality (deaths near the time of delivery or soon after) in Southern Asia and Western Africa. 

(Interview conducted May 2024)

Yes. Despite a general decrease in child mortality in the world as a whole, there are still very strong contrasts between regions. This is made very clear in the annual report published by the UN’s Interagency Group for Child Mortality Estimation (UN IGME). According to the group’s most recent estimate, 4.9 million children under 5 died across the world in 2022, 2.7 million of them (55% of the total) in sub-Saharan Africa; as opposed to 30,000 (0.6%) in Europe. In terms of mortality risk, these sharply contrasting figures amount to 71 deaths of under-fives per 1,000 live births in sub-Saharan Africa as against 4 in Europe. The second-highest mortality risk is found in the combined regions of Central and Southern Asia: 34 deaths per 1,000 births, corresponding to 1.3 million deaths in 2022.

Since the year 2000, mortality among children under 5 has fallen over 50% in sub-Saharan Africa—a major advance that deserves to be highlighted. However, current mortality levels remain similar to those observed in Europe between the two world wars. One United Nations sustainable development goal today is to reduce child mortality to 25 deaths per 1,000 births in all countries by 2030. Despite the undeniable progress, this goal will not be reached in the sub-Saharan region. 

The data for high mortality countries is often of poor quality, a problem due first of all to the fact that some countries do not have a system for registering or recording births and deaths. Such a system is crucial for estimating and tracking child mortality. Second, in low- and middle-income countries that do have a system, some demographic events still go unrecorded.

In these contexts, the challenge is to use alternative data sources and methods to obtain sufficiently accurate mortality estimates. This is one of the main objectives of my research. The point is to assess and cross-reference imperfect, incomplete data using a range of different sources, including surveys, censuses, and local tracking systems such as population observatories. Demographers can then make use of a number of regularities observed in human populations to fill in at least some of the missing information. For example, mortality distribution between 0 and 5 years is relatively regular, a regularity often used to estimate infant mortality (death in the first year of life), since the under-five mortality indicator is easier to obtain directly. 

We are also making greater use of highly advanced statistical methods, specifically Bayesian methods aimed first and foremost at improving measures of the statistical uncertainty of research findings. These methods require specialized profiles. In this connection, I’m glad to have been able to collaborate—for several years now—with Julio Romero Prieto of the London School of Hygiene and Tropical Medicine. 

That datum are essential for high mortality countries, yes, but also for countries where mortality is low. In all cases, level of child mortality is a public health indicator that synthesizes in and of itself a great number of variables, including public health conditions, healthcare quality, nutrition, living standards, and social conditions generally. This is why it’s such an important indicator of a country’s development. It helps countries situate themselves in relation to others and assess the impact of their health policies.

From 1950 to 2010, under-five mortality in France fell from 60 to 4 deaths per 1,000 births. However, in the last decade we have seen that indicator stagnate and even rise slightly. In France as elsewhere, this fact is not yet enough to orient health policy, but it does draw attention to the potential problem, and that’s crucial. We then have to use supplementary data to find an explanation for the situation.

Nikita Kupska, who is starting a thesis on this subject at INED, has already explored several hypotheses. The rise could be due to the fact that women are increasingly having children later in life, meaning the child is at greater risk. It could also be due to a deterioration in the health of pregnant women. Conversely, it could be due to medical advances that enable increasing numbers of fragile infants to be born. Those babies, who are at higher risk of dying shortly after birth, would mechanically increase mortality among liveborn children.