Risk indicators of childhood undernutrition in Kingston, Jamaica

In Jamaica, early childhood undernutrition remains a problem; however, the health of all children cannot be monitored due to limited resources. Therefore, there is a need for the early identification of children at risk of undernutrition. A simple screening instrument for use by paraprofessionals in the primary health care system was developed. We conducted a case-control study using 649 children, aged 6 to 48 months. The cases were undernourished (weight-for-age less than 80% of the reference) and identified from a survey. The children's guardians were given a questionnaire comprising 31 variables thought to be associated with undernutrition and which were present from 6 weeks of age. Nine variables were significantly associated with undernutrition. Multiple logistic regression analysis indicated that low birth weight, short birth spacing, being born at home, poor ante-natal and post-natal clinic attendance, overcrowding and a lack of house-hold possessions were independent predictors of undernutrition. Using these variables, a simple scoring system was developed to identify high risk children. It had a sensitivity and specificity of 56% and 76% respectively and a positive predictive values of 31%. This simple screening instrument should be easy to use in the primary health care system. However, its low sensitivity indicates that it is difficult to identify children who are at risk of undernutrition from 6 weeks of age. Subsequent high morbidity and poor child care which were not measured may account for some of the missed cases.