Contraceptive usage for spacing of childbirth is an important indicator for understanding family planning practices as well as fertility transitions. Fertility transition are especially important in sub-Saharan Africa where fertility remains high in many countries. However, estimates and short-term projections are generally not available for countries in this region. We developed a Bayesian hierarchical time series model to estimate and project usage of contraceptives for spacing by 5-year age groups for all countries in sub-Saharan Africa for the years 1985-2030. Estimating country-age-year specific usage is challenging due to limited data availability. We use Bayesian hierarchical models to share information across countries and spline regression to share information across age groups. Temporal changes are captured with logistic growth curves and autocorrelated distortion terms. Models are validated with out of sample exercises which test the model's ability to project into the future as well as the models ability to estimate historical trends. Validation results show the model is well calibrated. Estimates reveal noteworthy variability across countries and across age groups.
| Identifer | oai:union.ndltd.org:UMASS/oai:scholarworks.umass.edu:masters_theses_2-1902 |
| Date | 29 October 2019 |
| Creators | Guranich, Gregory |
| Publisher | ScholarWorks@UMass Amherst |
| Source Sets | University of Massachusetts, Amherst |
| Detected Language | English |
| Type | text |
| Format | application/pdf |
| Source | Masters Theses |
Page generated in 0.0016 seconds