With the target date of 2015 for meeting the Millennium Development Goals (MDGs) nearing, studies that examine trends in target indicators by subgroups are valuable contributions to the understanding of whether targets will be met, and can help to identify inequities in indicator trends for certain at-risk populations. For the nutritional target within MDG 1, target 2 aims to halve the proportion of undernourished people between 1990 and 2015. Within this target are two measurable indicators—the prevalence of underweight children younger than 5 years of age and the proportion of the population below the minimum level of dietary energy consumption.1 Improving trends in average indicator values for national or regional groups can mask substantial lags in vulnerable populations within these larger geographical areas. In the paper by Christopher Paciorek and colleagues in this issue of The Lancet Global Health,2 differences in trends in under-5 underweight between rural and urban populations are examined by countries and regions from 1985 to 2011. On the basis of the data presented in this paper, it seems that, on average, children in urban areas fare better than their rural counterparts. The urban—rural differences are greater in some regions than others and the trends show variability in the narrowing of the urban—rural gap by regions.
The comparison of rural versus urban populations has relevance because of the rapid urbanisation now occurring in low-income and middle-income countries, with the potential for mixed results. How urbanisation might affect the nutritional status of children is not clear. Migration to urban areas could improve access to food and health services, but this might depend on wealth and other inequalities in urban areas. Additionally, the quantity and quality of food is likely to be different in rural and urban areas. Ultimately, it would be valuable to analyse these data not just by the rural/urban divide but also by socioeconomic status within these two strata because the reduction in underweight over time, especially within urban areas shown in this analysis, might mask socioeconomic differences with the urban environment. Such an analysis would also help address one of the indicators of MDG 7—improvement in the lives of slum dwellers.
Paciorek and colleagues’ analysis focuses on two measures of nutritional status: weight-for-age and height-for-age. Whereas weight-for-age is the indicator of focus for MDG 1, height-for-age is a better measure of chronic malnutrition, and weight-for-height would be a better measure of acute malnutrition. It would have been interesting to know whether the prevalence of acute malnutrition was also declining over time and with urbanisation. Presumably, urbanisation reduces seasonal fluctuations in the availability of food, but if the urban poor cannot access this more stable food supply, this benefit would not be evenly distributed within the urban population.
Although this analysis concentrates on underweight and stunting as measures of improvement in nutritional status, another question of importance to address would be trends in overweight and obesity as populations move from rural to urban environments. This question would be important to address in terms of socioeconomic differences within the urban environment, because access to more diverse and higher-calorie foods might not necessarily imply a higher-quality diet. The recently publishedLancet Nutrition Series indicates that overweight and obesity are increasing, especially in middle-income countries.3 In those analyses, the prevalence of obesity was greatest in lower wealth quintiles and the data suggested that obesity was higher in urban than rural areas. This finding might reflect the ongoing, but as yet incomplete, nutrition transition in low-income and middle-income countries, compared with the USA where obesity is higher in lower wealth quintiles.
Paciorek and colleagues’ analysis was a huge undertaking that included reanalysis of existing data and modelling of missing country-specific and time-specific data. The paper has broader value for the nutrition research community in that it provides a set of methods (Bayesian hierarchical mixture models) by which to do these analyses, and indicates the importance of examining time trends but also subgroup analyses of these trends. It is also important to be able to see trends in the context of relative versus absolute changes over time. The MDGs have expressed most changes in relative rather than absolute numbers, which makes sense, but as we reach lower levels of burden, absolute change could become more important. These changes are pertinent because decisions are currently under discussion regarding MDGs beyond 2015.
Much of Paciorek and colleagues’ analyses relied on stringing cross-sectional data together to produce time trends from a wide range of sources of varying quality and representativeness. This type of analysis and its challenges provide a good opportunity to think about ways to improve the collection of data that would inform such analyses going forward. It would benefit the global community if there were ways to further harmonise sample selection and measurement indicators. Additionally, might we interest donors invested in improving the tracking of global health in the idea of having sentinel sites that are regionally representative where cross-sectional but also repeated measures on individuals are collected over time? These data could also be linked to demographic and socioeconomic data to better track changes in health indicators over time and ensure that average improvements do not mask inequities in these indicators among subgroups of the population such as the urban poor.
I declare that I have no conflicts of interest.
