Global estimation of unintended pregnancy and abortion using a Bayesian hierarchical random walk model
share
Unintended pregnancy and abortion estimates are needed to inform and motivate investment in global health programmes and policies. Variability in the availability and reliability of data poses challenges for producing estimates. We developed a Bayesian model that simultaneously estimates incidence of unintended pregnancy and abortion for 195 countries and territories. Our modelling strategy was informed by the proximate determinants of fertility with (i) incidence of unintended pregnancy defined by the number of women (grouped by marital and contraceptive use status) and their respective pregnancy rates, and (ii) abortion incidence defined by group-specific pregnancies and propensities to have an abortion. Hierarchical random walk models are used to estimate country-group-period-specific pregnancy rates and propensities to abort.
READ FULL TEXT