An Application of the Confidante Method to Estimate Induced Abortion Incidence in Java, Indonesia.

Context: Induced abortion is legally restricted and highly stigmatized in Indonesia, and is thus extremely difficult to measure. Indirect methods leveraging women's social networks, such as the Confidante Method, have shown promise in estimating hidden behaviors, including abortion, in similar settings.

Methods: A community-based survey was conducted among 8,696 women aged 15-49 in Java, Indonesia, in November 2018-January 2019. Data were collected via in-person interviews with respondents about their own abortions and those of up to three of their closest confidantes. One-year induced abortion incidence rates per 1,000 women were estimated using a direct-report approach and the Confidante Method.

Results: The direct-report abortion rate was 3.4 per 1,000 women in 2018, compared with the Confidante Method rate of 11.3 per 1,000. Among the confidantes of women who reported an abortion in the past five years, the abortion rate was 42.0 per 1,000. Half of the women reported that they had no confidantes with whom they shared private information. Among women reporting an abortion and at least one confidante, 58% had disclosed their abortion to their confidante, indicating that substantial transmission bias was present.

Conclusions: The Confidante Method relies on several assumptions that did not hold in this study. Although the method performed better than the direct-report approach, it underestimated the incidence of abortion in Java. More research is needed to understand how abortion-related information is shared within social networks and to assess the appropriateness of applying the Confidante Method to estimate abortion in a given context.