Do Restrictive Abortion Laws Affect the Severity of Abortion-Related Morbidities in Africa?

Blake Erhardt-Ohren, University of California, Berkeley
Hedieh Mehrtash, UNDP/UNFPA/UNICEF/WHO/World Bank Special Programme of Research
Özge Tunçalp, World Health Organization
Rosalinda Hernandez Muñoz, Sigil Consulting Group
Adama Baguiya, Institut de Recherche en Sciences de la Santé
Kwame Adu-Bonsaffoh, Korle-Bu Teaching Hospital, University of Ghana
Rodolfo Gomez Ponce de Leon, Pan American Health Organization (CLAP/WR-PAHO/WHO)
Mariana Romero, Centro de Estudios de Estado y Sociedad
Philip Govule, Department of Epidemiology and Disease Control, School of Public Health, University of Ghana
Ausbert Thoko Msusa, Centre for Reproductive Health, College of Medicine, University of Malawi
Caron Kim, World Health Organization
Luiz Francisco Baccaro, UNICAMP
Folasade Adenike Bello, Department of Obstetrics and Gynaecology, University of Ibadan
Zahida Qureshi, Department of Obstetrics and Gynaecology, University of Nairobi
Rachidatou Compaore, Institut de Recherche en Sciences de la Santé
Ndola Prata, Bixby Center for Population Health and Sustainability, UC Berkeley

We investigate whether the severity of abortion-related morbidities differs across countries with varying degrees of abortion restrictions in sub-Saharan Africa. This cross-sectional prospective study abstracted 13,657 medical records from women presenting to 210 facilities in the WHO MCS-A. We classified each country's abortion laws into four categories: abortion on request, to preserve health, to save a life, prohibited altogether. We compare patient’s characteristics with severe and non-severe abortion complications across countries with dissimilar levels of abortion law restrictiveness; and run a multivariable model exploring associations between the severity of abortion-related morbidity and the restrictiveness of the law within and across African countries. We use post-model estimations to explore predictive probabilities of abortion severity: by country, region, and for significant covariates in the model. Results of this research will be useful to policymakers, advocates, and government officials interested in reducing preventable maternal morbidity, near misses, and deaths due to abortion complications.