The Millennium Development Goal (MDG) program focused needed attention on unacceptably high levels of child deaths across the world, dedicating its MDG4 target to reducing the under 5 mortality rate by two thirds by 2015. Considerable gains have been achieved overall and many countries are on track toward that target.

Beneath the overall trends are more specific ones. These deserve more attention. This month a remarkable and unprecedented new dataset of individual deaths across the developing world sheds light on often hidden trends.

Findings from this dataset, produced by demographic surveillance sites in the INDEPTH Network and published in the open access journal Global Health Action, provide cause of death information collected on the ground over two decades from 13 low and middle income countries, including Bangladesh, Ghana, South Africa, Kenya, India, and Vietnam. It’s worth a look to gain a better understanding of the global burden of disease and cause specific mortality among all age groups in key parts of Africa and Southeast Asia, and it complements the mathematical models of the Global Burden of Disease studies.

When it comes to children, the new findings are particularly interesting for parts of Asia, where two thirds of the world’s children live. Clearly the number of childhood deaths from infectious diseases have declined—but there are steady increases in the proportion of deaths resulting from non-communicable causes.

That child deaths are shifting from infectious causes, such as pneumonia, malaria, or diarrheal disease, to so called external causes, such as “accidents” or violence, is one part of the epidemiological transition in developing countries. (Although it’s worth noting that this transition, according to the new data from INDEPTH, is less pronounced among the poorest people in developing countries). Increased uptake of vaccines, reduction in vitamin A deficiency, fewer births, and better hygiene are some of the key factors helping drive down infectious disease mortality. Immunization against measles alone helped prevent nearly 14 million deaths between 2000 and 2012. In their place are new, often unrecognized, problems.

For infants (younger than 1 year) the leading killers are still infections. But the leading cause of child deaths in Bangladesh is not what you would think: it’s drowning.

Incredibly, according to Unicef, up to 46 children drown daily, and 16 500 drowning deaths happen each year in flood prone Bangladesh. This makes it the largest killer of children aged between 1 and 15 years (this is also true for Cambodia, Vietnam, and Thailand). Most children are 2 years of age, and most die within 20 meters of their homes. This striking problem may have been known for years by advocates on the ground, but the issue of child drowning barely figures on the global child health agenda. Even recent commentary about the threat of climate change and flooding tends to emphasize the risks of infectious disease (cholera especially), even though flooding would seem to fuel the drowning problem.

Imagine the shift in focus needed to address a preventable cause of death such as drowning, when so much effort and investment has been centered on infectious disease in children?

Barriers to recognizing the breadth of the drowning problem include poor reporting, which makes the precise numbers from this new INDEPTH dataset especially valuable. In Bangladesh’s rural Matlab area, drowning caused 46% of deaths among children aged 1-4 years, followed by acute respiratory illness, including pneumonia (16%), malnutrition (16%), tuberculosis (3%), meningitis or encephalitis (2%), and diarrhea (2%). Among older children (age 5–14 years), where a third of deaths are owing to injury and other external causes, 18% were caused by drowning.

In a diverse landscape such as Bangladesh, geography appears to play a direct part. For example, in the country’s Bandarban region, which is 200 meters above sea level near the Myanmar border, drowning rates were much lower than in the flat river delta environments of other areas in Bangladesh, which are threatened by rising sea levels and flooding.

Children in Bangladesh and other parts of South Asia don’t usually have swimming skills, despite spending nearly every day walking on narrow paths or telephone poles across farm fields flooded with water. There are no barriers, and often no one to help if they fall in. Unlike in richer countries where time in water is recreation, in South Asia when children drown they are not playing.

I am aware of two recent initiatives to address drowning in Bangladesh. One is the SOLID project (Saving of Lives from Drowning) that is testing the use of playpens (to restrict children’s mobility) and crèches (to provide care) during the peak periods of the day when children are most at risk of drowning in rural areas. It also aims to provide family education and community awareness programs, and to reach 80 000 children.

The other Bangladesh initiative, developed by a Canadian doctor and funded in part by Grand Challenges Canada, teaches basic survival swimming skills to rural children 1-5 years of age, and offers daycare services for the younger children to keep them away from open bodies of water. The program plans to hire and train 90 Bangladeshis to help with the initiative.

Findings from these two efforts will be known in the coming years, and clearly we need to be concerned with both the effectiveness and sustainability of any interventions to help prevent drowning deaths. In the meantime, as the MDG4 target approaches, the intense focus on infectious disease killers in children should give way to new recognition of non-communicable causes of death, especially hidden issues like drowning. Just this week, in fact, the World Health Organization released a global report on drowning, which undoubtedly will raise awareness and spur action.

Jocalyn Clark (@jocalynclark) is executive editor of the Journal of Health, Population and Nutrition, and other external publications at icddr,b (a global health research organization in Dhaka, Bangladesh). She was a former senior editor at PLOS Medicine (2008-13), and a former editorial registrar, assistant editor, and associate editor at The BMJ (2002-08).

Competing interests: I currently serve as a member of the INDEPTH Network’s scientific advisory committee. I formerly consulted to Grand Challenges Canada.