Accuracy of recall of use of an intrauterine device.
Researchers analyzed interview and physician records' data on 45 women with breast cancer (cases) admitted to the Central Institute of Cancer Research in Berlin, East Germany between November 1982-June 1986 and born in 1983 or later and 194 women (controls) admitted to the Klinikum Berlin-Buch also in East Germany for conditions other than breast cancer to compare recall accuracy in women who had ever used an IUD. These women were drawn from case control study of the relationship between breast cancer and oral contraceptive use was part of the WHO Collaborative Study of Neoplasia and Steroid Contraceptives. Agreement between patient recall and physician records was exceptional for duration of IUD use (p<.001), number of IUD episodes (kappa=0.79), time since 1st IUD and time since last IUD use (p<.001). Agreement rates did not differ between cases and controls. 75% of the women could not name the IUD brand used so the researcher could not examine agreement of brand name. Thus, other than brand name, this study showed that validity of information on IUD use obtained from interviews is significant. In fact, it also pointed out that case control studies probably yield sound relative risk estimates.