Gender, Power, and the Risk of Spontaneous or Induced Abortion in Zanzibar, Tanzania: A Causal Inference Approach

Divya Vohra, University of California, Berkeley
Caitlin Gerdts, University of California, San Francisco
Alison Norris, Ohio State University

Intimate partner violence (IPV) has long been recognized as a public health and human rights issue. IPV can increase the risk of both spontaneous and induced abortion, relative to live birth. Less is known about how social environments that normalize IPV might contribute to poor reproductive health outcomes. This study seeks to understand how women’s autonomy and attitudes about IPV affect their risk of induced abortion, relative to spontaneous abortion, in a population of women receiving post-abortion care (PAC) services in Zanzibar. We use G-computation to estimate the causal effect of women’s autonomy and IPV attitudes on their risk of reporting an induced abortion. G-computation provides a flexible way of using data obtained from an observational study to simulate what would have happened in a randomized trial. This is one of the first efforts to use such causal inference methods in the field of reproductive health.