Abstract
Self-report of uterine fibroids (UF) has been used for epidemiologic research in different environments. Given the dearth of studies on the epidemiology of UF in Sub-Saharan Africa (SSA), it is valuable to evaluate its performance as a potential tool for much needed research on this common neoplasm in SSA women. We conducted a cross-sectional study of self-report of UF compared with transvaginal ultrasound diagnosis (TVUS) among 486 women who are members of the African Collaborative Center for Microbiome and Genomics Research (ACCME) Study Cohort in central Nigeria. We used log-binomial regression models to compute the classification, sensitivity, specificity, and predictive values of self-report compared to TVUS, adjusted for significant covariates. The prevalence of UF on TVUS was 45.1% (219/486) compared to 5.4% (26/486) based on self-report of abdominal ultrasound scan and 7.2% (35/486) based on report of healthcare practitioner's diagnosis. Self-report correctly classified 39.5% of the women compared to TVUS in multivariable adjusted models. The multivariable adjusted sensitivity of self-report of healthcare worker diagnosis was 38.8%, specificity was 74.5%, positive predictive value (PPV) was 55.6%, and negative predictive value (NPV) was 59.8%. For self-reported abdominal ultrasound diagnosis, the multivariable adjusted sensitivity was 40.6%, specificity was 75.3%, PPV was 57.4%, and NPV was 60.6%. Self-report significantly underestimates the prevalence of UF and is not accurate enough for epidemiological research on UF. Future studies of UF should use population-based designs and more accurate diagnostic tools such as TVUS.