The authors assessed the accuracy of perimenopausal estrogen use reporting by 430 women in a prospective study of bone health risk factors. Data from two time points 5 years apart indicated that 383 (89%) women could consistently report having ever used perimenopausal estrogens or not. Of the 383, 138 reported some lifetime perimenopausal estrogen use; 97 (70%) of these consistently reported duration of use. The age-adjusted relative odds that women would misreport having ever used perimenopausal estrogens was 11.7 (1.3, 100.6) for women with 11-20 years since last use, and 22.2 (1.8, 277.4) for 21+ years. Among women who inconsistently reported ever use of perimenopausal estrogen, the relative odds of reporting use at baseline and never use at follow-up as compared to reporting the converse by women aged 70-75 was 8.1 (1.2, 53.2) times that for women aged 60-69 at follow-up, and increased to 9.6 (1.8, 49.9) for women aged 76-85. This suggests that women can consistently report perimenopausal estrogen use, but accurate report of use declines in women whose last use precedes the interview by over 10 years. Accurate report of duration or dates of perimenopausal estrogen use may be compromised in women of more advanced age.