Estimating Global Abortion Incidence from 1990 until 2015 Using Bayesian Methods and a Framework of Abortion Determinants

Leontine Alkema, National University of Singapore
Anisa Asafi, World Health Organization (WHO)
Akinrinola Bankole, Guttmacher Institute
Bela Ganatra, Ipas
Caitlin Gerdts, University of California, San Francisco
Heidi Johnston, ICDDR,B
Anna Popinchalk, Guttmacher Institute
Ozge Tuncalp, Johns Hopkins University
Paul Van Look, Independent Consultant
Sanqian Zhang, National University of Singapore
Gilda Sedgh, Guttmacher Institute

Abortion incidence estimates are needed to monitor progress in achieving universal access to reproductive healthcare. To date, estimates abortion incidence worldwide have been made for 1995, 2003 and 2008. We propose a new approach to estimating abortion incidence and use this approach to estimate current abortion levels and past trends. The estimates are based on available data for all countries pertaining directly to abortion incidence, as well as information on factors indirectly related to abortion, such as contraceptive method mix and the unmet need for contraception. In the proposed model, abortion incidence is estimated on the basis of determinants of abortion using Bayesian methods. The final paper will present selected findings to illustrate the modelling approach.