Measures of schoolchild height and weight as indicators of community nutrition, lessons from Brazil

Many countries measure the heights and weights of children in primary and secondary school. Individual children are usually monitored at school against growth charts, with appropriate referrals. Schoolchild data aggregated to the community, state or nation, and validated against population surveys, led to interventions such as school-based dietary supplementation and feeding, and community or national dietary education. Nutritional programs rarely target dietary needs of children 5-19. Influences on anthropometric nutritional status for children 0- 5 are widely examined in population surveys; but similar data are lacking for children 5-19, undesirably, as differences in growth and development from 5-19 are strongly associated with health in later life. Nutritional status of older children requires study in the context of household, community and national economic conditions. Many such details were surveyed by the Pesquisa de Orcamentos Familiares (POF), across the 27 provinces of Brazil, 2002-2003, conducted by the Instituto Brasileiro de Geografia e Estatistica (IBGE), a national agency publishing industrial activity, agricultural production, employment, prices, and GDP series. The POF sampled 48,470 households, and 178,375 persons, for household and social characteristics, anthropometrics, economic activity, sources of income, and detailed expenditures (for example 3,256 foods). Here, nutritional status of 50,237 persons aged 5-19 is estimated using WHO 2007 Reference Standards for age, sex, height, weight and BMI. Characteristics of students (public or private, age in grade) and children not enrolled in school, part and full-time employment, personal income and expenditure on food, entertainment and stimulants, were related to relative and seasonal differences in nutritional status. Child nutritional status also responded to household education, income, employment, spending (particularly on food), food consumption, adult and cohort anthropometrics, and community factors, such as urbanization and location. Outcomes and explanatory factors were mapped for spatial autocorrelation and tested for Granger causation over the survey period. Child nutritional status varied seasonally; annual school censuses do not model average nutritional status. Nutritional status of enrolled children incompletely modeled children not in school. Accordingly, continuing population surveys are necessary to monitor the nutritional status of all children 5-19. Variations in nutritional status of enrolled children were sensitive to economic data available at monthly or quarterly intervals in most places, such as local prices, economic activity and employment, suggesting these as factors in active policy. Despite its size, the POF sample was insufficient to very significantly model the responses of child nutritional status to local economic conditions. These are strong reasons to support the annual collection of height and weight for all schoolchildren, and to expand monitoring of the influence of community economic conditions on child nutritional status.