A cohort of 6,011 urban children born in 1982 in the hospitals of Pelotas, Rio Grande do Sul, was followed up so that their morbidity, mortality and growth could be assessed. These children accounted for over 99% of all births in the city. A 30% sample of the children were visited at home when approximately 12 months old, and the whole population was visited at about 20 months of age. It was possible to locate 81% of the children at 12 months. This proportion increased to 86% at 24 months, due to a change in the logistics of the field work which then included visiting all 69,000 households in the city to locate children whose families had moved within the urban area. The methodology and main difficulties encountered are discussed, and the characteristics at birth of children who were located at the first follow-up visit was compared to those of children lost to follow-up. Children who were not located tended to come from poorer families but the overall differences were not very marked, with more than 75% of children in any of the major socio-economic subgroups being located. The potential use of the large amount of information available on each child is demonstrated by the study of the relationships between birthweight, family income and nutritional status at 9-15 months of age. Birthweight is shown to be a very strong predictor of weight for age, length for age and, to a lesser extent, weight for length, there being virtually no malnutrition among children born with more than 3,000 g. Birthweight is also shown to be closely associated with weight at 9-15 months within all five family income groups, with a given absolute difference in weight at birth persisting to the end of the first year. However, due to the important effects of income on weight gain, a child born with 2,500 g in the lowest income group would be on average 1,200 g below the standard NCHS weight at 12 months, whereas a child with the same birthweight from the highest income group would have caught up with the standard. This study has shown that it is possible to identify a population based cohort of children and to follow them up for two years in a medium-sized Brazilian city.
Child; Epidemiologic methods
