Do women forget their births? A study of maternity histories in a rural area of Senegal (Niakhar).

5068 maternity histories were recorded among women aged 15-89 years in a rural area of Senegal. The quality of the estimates of fertility, child mortality, and perinatal mortality was analyzed for consistency. There was no evidence of any major underreporting of births, deaths, or still births according to age, even among the oldest women. Estimates were compared with comparable values derived from a longitudinal demographic surveillance system (DSS) in the same area. The age patterns of cumulated fertility and mortality derived from the maternity histories were consistent with those of the DSS. Differences in the levels of fertility and mortality with respect to the longitudinal records could be explained by small differences within the selected villages, by selection biases, and by recent trends in demographic parameters. Values of perinatal mortality were also equivalent to those recorded by the DSS. Women did not seem to forget their births to an extent large enough to produce strong biases, even at older ages. However, an analysis of differences by field workers revealed that some 2% of the births and 4% of the deaths may have been omitted, which gives an idea of the potential accuracy of maternity histories for the estimation of fertility and mortality levels.