A validation of birth and pregnancy histories in Matlab, Bangladesh

Demographers have speculated about differences between birth and pregnancy history questionnaires in terms of data quality; however, differences have never been examined in an experimental study. The present analysis is an individual-level validation of fertility and mortality information from birth and pregnancy histories in the 1994 Matlab Demographic and Health Survey (MDHS). The MDHS was conducted after the 1993--94 Bangladesh Demographic and Health Survey (BDHS), and it used the same interviewers and field procedures as the BDHS. One-half of the MDHS questionnaires contained the standard birth history and the other half had a pregnancy history module. Interviewers were instructed to administer birth history and pregnancy history questionnaires by random assignment within each village.;MDHS respondents and their pregnancy events were matched with vital-events records from Matlab's Demographic Surveillance System (DSS) for a 15-year period before the survey. The analysis compared the validity of birth and pregnancy history questionnaires in three respects: (1) the completeness of live birth, early neonatal death, neonatal death, and infant death reporting; (2) the degree to which the reported total fertility rate (TFR) and cumulative fertility rate (CFR) differ from the actual TFR and CFR; and (3) the accuracy of the placement of live births, early neonatal deaths, neonatal deaths, and infant deaths in time.;A total of 1,925 women and 5,767 events were included in the analysis. Pregnancy and birth histories performed equally well in terms of live birth coverage in crude analyses, but pregnancy histories were superior after adjustment for child's current survival status, maternal age at the time of the interview, and parity in a multivariate logistic regression model (p = 0.051). Also, pregnancy histories did significantly better than birth histories in terms of the accurate placement of infant deaths in time (p < 0.05). Thus, from the perspective of documenting fertility, the standard DHS birth history does well. However, pregnancy histories hold an advantage in terms of documenting early mortality.