Large electronic datasets do not always correctly identify women (defined as individuals whose sex assigned at birth was female, although we acknowledge that biological sex does not always correlate with gender identity) experiencing menopause or their symptoms, making it challenging to generate robust evidence on the burden of menopause. This study aimed to construct and characterize cohorts of women experiencing menopause to explore differences in their identification based on the cohort criteria and to provide recommendations for defining cohorts of individuals experiencing menopause in real-world data. This was a retrospective, multi-country cohort study using databases that included a total of ~240 million individuals from five countries: three observational electronic health record databases from France, Germany, and the UK; and two administrative claims databases from Japan and the USA. Demographic, comorbidity, and concomitant medication covariates were used to characterize the study target cohorts at multiple time periods relative to the index event. Counts were calculated for the number of women (as identified in the databases) in each database that matched the target cohort criteria. Cohorts that included a more specific cohort definition of women aged 40–65 with a probable menopause diagnosis with or without symptom criteria produced age distributions that best matched those expected for women experiencing menopause, along with the highest VMS incidence rates. The prevalence of comorbidities varied between databases and cohorts. However, generally, the most common comorbidities were anxiety, depression, hypertension, hypothyroidism, and osteoarthritis. The findings provide evidence to support the development of a reproducible definition of menopause and menopause symptoms, which can be used for future analysis of observational healthcare databases.
Banks et al. (Sat,) studied this question.