OBJECTIVES: To examine the validity of self-reported primary hip replacement, primary knee replacement, and cholecystectomy. STUDY DESIGN AND SETTING: The data from 28,524 participants in the Million Women Study who were recruited from Scotland were linked with routinely collected Scottish hospital admission data. We compared the reporting of three operations by the study participants on a postal questionnaire with that recorded in the hospital admission database. RESULTS: Annual incidence rates for the three operations were similar using hospital admission and self-reported data at: 2.1 (n=220) and 2.3 (n=247) per 1,000 women, respectively, for hip replacement; 1.1 per 1,000 women from both sources for knee replacement (n=118 and n=116, respectively); and 3.4 (n=357) and 3.2 (n=344) per 1,000 women, respectively, for cholecystectomy. Agreement between self-report and the hospital admission records was good for all three operations with agreements of 99.8%, 99.9%, and 99.6% for hip replacement, knee replacement, and cholecystectomy, respectively. CONCLUSION: In middle-aged women who answered postal questionnaires, self-reporting of primary hip replacement, primary knee replacement, and cholecystectomy was found to agree well with information held in a medical record database.

Original publication




Journal article


J Clin Epidemiol

Publication Date





1190 - 1194


Arthroplasty, Replacement, Arthroplasty, Replacement, Hip, Arthroplasty, Replacement, Knee, Cholecystectomy, Cohort Studies, Data Collection, Databases, Factual, Female, Hospitalization, Humans, Medical Records, Middle Aged, Scotland