Estimating the lifetime incidence of induced abortion and understanding abortion practices in a Northeastern Tanzania community through a household survey

Bilikisu Elewonibi a, Caroline Amour b, Sarah Gleason a, Sia Msuya b, David Canning a, Iqbal Shaha
Publication year: 


All Tanzanian abortion estimates rely on health facility data that do not take into account completely the incidence of abortion. This papers aims to estimate the lifetime incidence of induced abortion in Arusha, Tanzania via direct and double list-experiment methods using community data and evaluate outcomes and behaviors of women who had an abortion.


From January-May 2018, a face-to-face interview survey was conducted on a representative sample of sexually active women (n=3,658) living in Arusha, Tanzania. Participants were selected in a three-stage random process and questions were asked about reproductive history, contraceptive use, and health seeking behaviors. A direct question and double list-experiment was used to estimate lifetime incidence of abortion.


Lifetime abortion incidence was 3% using the direct question compared to 7.7% using the double list-experiment method. However, post-estimation tests revealed a key study design violation thus invalidating list the experiment estimate. We find that 45% of women received their abortion outside the formal health care system, the most frequent method used was manyono pill (traditional medicine), and only 50% of women who experienced abortion complications sought treatment.


We provide another example of the performance of list experiment in measuring abortion incidence. Nearly half of reported abortions took place outside of the formal health system highlighting the substantial underestimation while using facility data to measure abortion. Seeking health care for potential complications was low despite post-abortion care services being free and legal in Tanzania.