A comparison of four methods for determining prevalence of induced abortion, Taiwan, 1970-1971.

A methodological study of four procedures for obtaining information about induced abortion was done in Taiwan in 1970 and 1971. About 2000 women were in a repeat interview group and seen every six weeks for nine rounds; one-half of these were also asked to provide urine on each round for a pregnancy test. Another 2000 women were interviewed only once; about one-half were asked "standard" knowledge, attitude, and practice questions about their reproductive history, including abortion. The other one-half were asked the same questions except they were not queried about their practice of abortion, but rather were asked to participate in the randomized response technique (RRT). Three methods produced estimates that 13-14% of the women had had an induced abortion. The RRT gave a significantly higher estimate, 28%. All methods used were judged feasible, although there is a suggestion that, without some improvement, the RRT may be difficult to use in low literacy populations. The repeat interview did not demonstrate advantage over the single interview in obtaining prevalence of women with a history of induced abortion. Ths use of a shorter interval with probing when the pregnancy test is positive might improve identification of recent abortions.