Swedish Registers

The Swedish registry approach produced a far lower 10-year AN incidence rate than a Finnish twin study that used structured interviews, reflecting differences in case ascertainment. Registry-based estimates of anorexia nervosa prevalence are likely underestimates because the register captures only clinically diagnosed cases seen in specialist care. Swedish national registers have greater than 99% coverage for vital status and emigration, resulting in minimal loss to follow-up. Chart review evidence found the National Patient Register has a positive predictive value of approximately 0.75 for anorexia nervosa diagnoses. Records were linked with pseudo-anonymised Swedish personal identification numbers. The study linked four Swedish registers to define the population, covariates, diagnoses, and medication exposures.