Retrospective Birth Weight Data in Developing Countries: Can They Be Trusted?

Andrew R. Channon, University of Southampton
John W. McDonald, University of Southampton
Sabu S. Padmadas, University of Southampton

This study looks at the accuracy of birth weight reports in six DHS countries, and assesses the difference to the estimates of the proportion with low birth weight (LBW) made by different assumptions regarding this accuracy. A large proportion of weights are heaped at 500g intervals, including 2500g, the boundary for the classification of LBW. Displacement of measured birth weight in either direction can produce considerable bias in the estimates of LBW, an indicator of maternal and child health. Heaping is also assessed by the method by which the mother reported the weight, either from memory or after reference to an official health card. Card recall was expected to be more accurate than memory recall. In some countries this is not the situation, and card recall has as much heaping as memory recall, indicating that officially recorded weights may not be as accurate as assumed.