In this prospective 10-year study in elderly aged 60 years and over, there was a 1.3% per year reduction in the standardized incidence of hip fracture in women but not in men. This decrease was mainly due to changes in the standardized incidence of hip fracture in institution-dwelling women. INTRODUCTION: A decrease in age-adjusted hip fracture incidence has been recently demonstrated in some countries. Since a large proportion of hip fractures occur in nursing homes, we analyzed whether this decreasing trend would be more detectable in institution-dwelling elderly compared with community-dwelling elderly. METHODS: All hip fracture patients aged 60 years and over were identified in a well-defined area. Incidence of hip fracture, age- and sex-adjusted to the 2000 Geneva population, was computed in community- and institution-dwelling elderly. RESULTS: From 1991 to 2000, 1,624 (41%) hip fractures were recorded in institutionalized-dwelling elderly and 2,327 (59%) in community-dwelling elderly. The standardized fracture incidence decreased by 1.3% per year in women (p = 0.039), but remained unchanged in men (+0.5%; p = 0.686). Among institution-dwelling women, hip fracture incidence fell by 1.9% per year (p = 0.044), whereas it remained stable among community-dwelling women (+0.0%, p = 0.978). In men, no significant change in hip fracture incidence occurred among institution- or community-dwelling elderly. CONCLUSIONS: The decrease in the standardized hip fracture incidence in institution-dwelling women is responsible for the reversal in secular trend. Future research should include stratification according to the residential status to better identify the causes responsible for the trend in hip fracture incidence.