Validity of self-reported hysterectomy and oophorectomy in a population-based cohort: The HUNT study

Objective: To validate self-reported hysterectomy and bilateral oophorectomy. Design: Validation study. Setting: Large population-based cohort study in Norway: The Tr & oslash;ndelag Health Study (HUNT). Population: The Trondelag Health Study 2 and 3 (HUNT2 and HUNT3) included questions on gynaecological history. Women who answered questions regarding hysterectomy and/or oophorectomy were included. In total, 30 263 women were included from HUNT2 (1995-1997) and 23 138 from HUNT3 (2006-2008), of which 16 261 attended both HUNT2 and HUNT3. Methods: We compared self-reported hysterectomy and bilateral oophorectomy with electronic hospital procedure codes. Main outcome measures: Sensitivity, specificity, positive predictive value and negative predictive value of self-reported hysterectomy and bilateral oophorectomy, by comparing with hospital procedure codes. Results: Self-reported hysterectomy and bilateral oophorectomy in HUNT2 and/or HUNT3 both had specificity and negative predictive value above 99%. Self-reported hysterectomy had a sensitivity of 95.9%, and for bilateral oophorectomy sensitivity was 91.2%. Positive predictive value of self-reported hysterectomy was 85.8%, but for self-reported bilateral oophorectomy it was 65.4%. Conclusions: Self-reported hysterectomy corresponded quite well with hospital data and can be used in epidemiological studies. Self-reported bilateral oophorectomy, on the other hand, had low positive predictive value, and results based on such data should be interpreted with caution. Women who report no previous hysterectomy or bilateral oophorectomy can safely be classified as unexposed to these surgeries.