Estimating the proportion of modern contraceptives supplied by the public and private sectors using a Bayesian hierarchical penalized spline model

Quantifying the public/private-sector supply of contraceptive methods within countries is vital for effective and sustainable family-planning delivery. However, many low- and middle-income countries quantify contraceptive supply using out-of-date Demographic Health Surveys. As an alternative, we propose using a Bayesian, hierarchical, penalized-spline model, with survey input, to produce annual estimates and projections of contraceptive supply-share outcomes. Our approach shares information across countries, accounts for survey observational errors and produces probabilistic projections informed by past changes in supply shares, as well as correlations between supply-share changes across different contraceptive methods. Results may be used to evaluate family-planning program effectiveness and stability.