Detecting induced abortions from reports of pregnancy terminations in DHS calendar data

This study considers whether pregnancy terminations reported in Demographic and Health Survey (DHS) calendar data can be classified accurately as having been spontaneous or induced based upon other information collected in the survey interview. A classification scheme is proposed that is an adaptation of the method developed by the World Health Organization for categorizing cases in which women admitted to hospitals experienced complications of pregnancy termination. The scheme is evaluated using data from the 1993 Turkey DHS. Evaluation results indicate that the method identifies true cases of induced abortion accurately, but tends to classify a relatively large number of reported spontaneous terminations as induced abortions. However, when it is corrected for likely respondent misreporting of induced abortions as spontaneous terminations, both the sensitivity and specificity of the method appear to be acceptable.