Cohorts in Low- and Middle-Income Countries: From Still Photographs to Full-Length Movies

p c f f [ p 2 c In recent years, the amount of information available on the ealth status of children and—to a lesser extent—adolescents iving in lowand middle-income countries (LMICs) has been assively expanded. Population-based surveys, repeated evry 5 years or so, are now available for 100 LMICs, providing information on nutritional status, health-related behaviors, morbidity, and mortality [1]. These include Demographic and Health Surveys (http://www.measuredhs.com/aboutsurveys/ dhs/start.cfm) and Multiple Indicator Cluster Surveys (http:// www.childinfo.org/). However useful such cross-sectional surveys may be, they rovide still photographs of the health and nutrition of young hildren at a given point. They do not tell us about dynamic rocesses such as growth, repeated disease episodes, intellectual evelopment, or changes in health behaviors. Nor do these sureys allow linking early exposures to later health, developmenal, or behavioral outcomes, an area of growing importance in ife-course epidemiology since it was pioneered by David Barker pproximately 30 years ago [2]. A PubMed search for birth cohort studies produces approxiately 6,500 references. Of the 20 countries with the largest umber of articles, only Brazil (208 references) and India (43 eferences) do not fall in the high-income country category. Yet, very year there are 121 million births in LMICs, and only 11 illion in high-income countries [3]. Such a massive imbalance ay have detrimental consequences. Results from high-income ountry birth cohort studies are often used to dictate global ealth policies, for example, warnings about the potential harmul consequences of rapid weight gain in childhood. However, esults from high-income country cohorts may not be applicable o LMIC populations where poor nutrition and growth faltering re still highly prevalent, andwhere catch-up growth is essential ot only for short-term survival but also for long-termhealth and uman capital [4–6]. There are a number of reasons why we need more cohorts rom LMICs. First, the frequency of exposures and outcomes, for xample, maternal and fetal undernutrition, infectious diseases orbidity, and all-cause child mortality, is substantially higher n LMICs [3]. Second, even if their frequencies are apparently imilar, exposures or health outcomes may vary substantially rom one setting to another. For example, physical activity in MICs is primarily due to commuting (on foot or bicycle),manual