Computer simulation of the effects of errors in birth registration on age-dependent anthropometric methods
Abstract
M Lozy In developing countries birth registration is likely to be delayed, leading to official ages which systematically underestimate the true ages. A computer simulation showed that such official ages underestimate the incidence of malnutrition in a community and lead to misclassification of the nutritional status of many subjects. These errors are more marked below than above 2 years of age. Estimation of the mean delay and its addition to the official ages led to improvement of the results. Above 2 years such estimated ages gave acceptable results; below that age the results were still not acceptable. It is suggested that one method of improving age estimates in large scale anthropometric studies in areas where birth records exist might be to carry out a preliminary survey to estimate the"mean delay in registration. This mean dealy would be added to all official ages and results above 2 years would be accepted. Below 2 years the estimated birth date would be used as a starting point from which the true birth date can be obtained by careful questioning of the parents. Anthropometric results referred to the ages so obtained would give better results than those of uncorrected ages or age-independent anthropometric methods.