Stanford researchers have determined that more than 15 million children are living in high-mortality hotspots across 28 Sub-Saharan African countries, where death rates remain stubbornly high despite progress elsewhere within those countries.

The study, published online Oct. 25 in The Lancet Global Health, is the first to record and analyze local-level mortality variations across a large swath of Sub-Saharan Africa.

These hotspots may remain hidden even as many countries are on track to achieve one of the U.N. Sustainable Development Goals: reducing the mortality rate of children under 5 to 25 per 1,000 by 2030. National averages are typically used for tracking child mortality trends, allowing left-behind regions within countries to remain out of sight — until now.

The senior author of the study is Eran Bendavid, MD, MS, an assistant professor of medicine and core faculty member at Stanford Health Policy. The lead author is Marshall Burke, PhD, an assistant professor of Earth System Science and a fellow at the Freeman Spogli Institute’s Center on Food Security and the Environment.

Decline in under-5 mortality rate

The authors note that the ongoing decline in under-5 mortality worldwide ranks among the most significant public and population health successes of the past 30 years. Deaths of children under the age of 5 years have fallen from nearly 13 million a year in 1990 to fewer than 6 million a year in 2015, even as the world’s under-5 population grew by nearly 100 million children, according to the Institute for Health Metrics and Evaluation.

“However, the amount of variability underlying this broad global progress is substantial,” the authors wrote.

“Mortality numbers are typically tracked at the national level, with the assumption that national differences between countries, such as government spending on health, are what determine progress against mortality,” Bendavid said. “The goal of our work was to understand whether national-level mortality statistics were hiding important variation at the more local level — and then to use this information to shed light on broader mortality trends.”

The authors used data from 82 U.S. Agency for International Development surveys in 28 Sub-Saharan African countries, including information on the location and timing of 3.24 million births and 393,685 deaths of children under 5, to develop high-resolution spatial maps of under-5 mortality from the 1980s through the 2000s.

Using this database, the authors found that local-level factors, such as climate and malaria exposure, were predictive of overall patterns, while national-level factors were relatively poor predictors of child mortality.

Image

Temperature, malaria exposure, civil conflict

“We didn’t see jumps in mortality at country borders, which is what you’d expect if national differences really determined mortality,” said co-author Sam Heft-Neal, PhD, a postdoctoral scholar in Earth System Science. “But we saw a strong relationship between local-level factors and mortality.”

For example, he said, one standard deviation increase in temperature above the local average was related to a 16-percent higher child mortality rate. Local malaria exposure and recent civil conflict were also predictive of mortality.

The authors found that 23 percent of the children in their study countries live in mortality hotspots — places where mortality rates are not declining fast enough to meet the targets of the U.N. Sustainable Development Goals. The majority of these live in just two countries: Nigeria and the Democratic Republic of Congo. In only three countries do fewer than 5 percent of children live in hotspots: Benin, Namibia and Tanzania.

As part of the research, the authors have established a high-resolution mortality database with local-level mortality data spanning the last three decades to provide “new opportunities for a deeper understanding of the role that environmental, economic, or political conditions play in shaping mortality outcomes.”  The database, available at http://fsedata.stanford.edu, is an open-source tool for health and environmental researchers, child-health experts and policymakers.

“Our hope is that the creation of a high-resolution mortality database will provide other researchers new opportunities for deeper understanding of the role that environmental, economic or political conditions play in shaping mortality outcomes,” said Bendavid.  “These data could also improve the targeting of aid to areas where it is most needed.”

The research was supported by a grant from the Stanford Woods Institute for the Environment. 

Bad weather in sub-Saharan Africa increases the spread of HIV, according to a study published in the June 2015 issue of the Economic Journal, co-authored by Stanford professor and FSE fellow Marshall Burke.

When the rains fail, farmers in rural areas often see their incomes fall dramatically and will try to make up for it however they can, including through sex work. Analysing data on more than 200,000 individuals across 19 African countries, the research team finds that by changing sexual behaviour, a year of very low rainfall can increase local infection rates by more than 10%.

The results have important policy implications for fighting the spread of the epidemic, as co-author Erick Gong of Middlebury College notes:

‘Existing approaches to stopping the spread of HIV – such as promoting condom use and the use of anti-retrovirals – remain critically important. But our results suggest that other policy approaches could be very useful too – in particular, approaches that provide safety nets to rural households when the weather turns bad.’

Policies and investments seemingly unrelated to HIV – such as the promotion of rural insurance or household savings schemes, or the development of drought-tolerant crops – might have surprising benefits in slowing the HIV epidemic. Co-author Kelly Jones of the International Food Policy Research Institute says:

‘The HIV/AIDS epidemic remains one of the world’s greatest health challenges, with over a million new infections per year in Africa alone. Our results expand the menu of options for addressing the epidemic, and highlight some surprising options that are not at the forefront of people’s minds.’

The research sheds valuable light on why HIV continues to spread in Africa. Previous studies have documented in limited settings that poor women often alter their sexual behaviour in response to an income shortfall. But until now, there has been little evidence that this response is big enough to affect the trajectory of the HIV epidemic.

To fill this gap, the researchers combined data on the HIV status of thousands of people across sub-Saharan Africa with data on the recent rainfall history in each individual’s location.

Because years of low rainfall can lead to much lower incomes in these locations, particularly in rural areas where people depend more heavily on agriculture for their livelihoods, variation in rainfall provides a way to study how changes in local economic conditions affect infection rates. Co-author Marshall Burke comments:

‘We were surprised by how strong the relationship is between recent rainfall fluctuations and local infection rates. As expected, the relationship is much stronger in rural areas, and particularly for women who report working in agriculture. These are the people who really suffer when the rains fail, and who are forced to turn to more desperate measures to make ends meet.’

Notes for editors: ‘Income Shocks and HIV in Africa’ by Marshall Burke, Erick Gong and Kelly Jones is published in the June 2015 issue of the Economic Journal.

Marshall Burke is an assistant professor of Earth System Science at Stanford University. Erick Gong is an assistant professor of economics at Middlebury College. Kelly Jones is a research fellow at the International Food Policy Research Institute (IFPRI).

For further information: contact Marshall Burke on +1-650-736-8571 (email: mburke@stanford.edu); Erick Gong on +1-802-443-5553 (email: egong@middlebury.edu); Kelly Jones on +1-202-862-4641 (email: k.jones@cgiar.org); or Romesh Vaitilingam on +44-7768-661095 (email: romesh@vaitilingam.com; Twitter: @econromesh).