Self-Reported Data on Abortion Morbidity: Using Qualitative Techniques with Community-Based Samples

Guttmacher Institute/IUSSP know (or know about) who have had abortion complications and their health-seeking behaviors. Many factors affect whether a women will develop abortion complications and how serious they will be, including the abortion method(s) used, the skill of the provider, the gestational age of the fetus and the woman’s socioeconomic characteristics. (These are proxies for determinants that are more difficult to measure, such as the safety of the abortion and the time it takes a woman to recognize she is having a complication and seek care.) Several factors can contribute to delays in seeking appropriate treatment or forgoing it altogether with sometimes fatal consequences: These encompass social and psychological factors such as fear that the abortion will become publicly known (in societies that outlaw abortion and attach high value to pregnancy), lack of knowledge about the seriousness of the complication(s), not having control over money to pay for the abortion treatment, lacking transportation to reach a facility and cultural preference for traditional care. Mistreatment or fear of mistreatment by medical personnel further dissuades women from seeking proper medical assistance in a timely manner, if at all (Jagwe-Wadda et al. 2006; Sáenz et al. 2006; Tombros et al. 2007). In their review of abortion research methodology, Barreto and coauthors identify methodological challenges that apply equally to quantitative and qualitative studies on abortion (Barreto et al. 1992). The authors draw attention to the potential pitfall of the data misrepresenting who is having abortions if certain subgroups of women are less likely to consistently participate in a study or are differentially lost to follow-up. Even when women consent to be interviewed, some subgroups may talk more freely about abortion than others. The authors point out that abortion morbidity will be overrepresented if, for example, women with serious complications are more likely than women with mild complications to talk about their abortion experience. Finally, the authors stress the importance of a good rapport between the respondent and the interviewer, as well as the role that the interview location plays in making the respondent feel at ease. Most studies measuring abortion incidence rely on facility-based samples, which are the most cost-effective samples to use. However, such samples cannot speak to Acknowledgments: Funding for this project was provided by the Netherlands Ministry for Development Cooperation, the World Bank and the UK Department for International Development. The authors would like to thank Diana Lara, Friday Okonofua and Susheela Singh for their comments on an earlier draft. They also thank other members of the IUSSP International Seminar on Measurement of Abortion Incidence, Abortion-Related Morbidity and Mortality (November 7–9, 2007) for their helpful comments.