Does pregnancy follow up improve reliability of under five mortality estimates in Health and Demographic Surveillance Systems? Insights from Bandafassi and Niakhar (Senegal)

Yempabou Bruno Lankoande, Institut National d'Etudes Démographiques (INED)/Institut Supérieur des Sciences de la Population (ISSP)
Gilles Pison, Muséum National D'histoire Naturelle

Pregnancy registration has been part of most HDSSs routine data collection in order to improve estimates of early life mortality. However, little research has investigated the effect of this strategy on mortality estimates. We used logistic regressions and cox models to look at associated factors of pregnancy reporting, and its net effect on the age patterns of under five mortality in Bandafasi and Niakhar HDSSs (Senegal). Reporting of pregnancies depends on frequency of follow up and on maternal factors such as maternal age at birth , ethnic group. In relation to mortality estimates in children, no standard pattern appeared according to the status of pregnancy registration. Particularly, neonatal mortality can be higher (Niakhar) or lower (Bandafassi) for children born from followed up pregnancies. The results suggest that various factors interact with pregnancy follow up so that further investigation is needed to capture its net effect on mortality estimates.