Objectives.Hip replacement is the preferred treatment for displaced femoral neck fractures, whereas other less expensive procedures are preferred for nondisplaced fractures. The authors determined whether there was geographic variation in the use of hip replacement to treat displaced and nondisplaced fractures.Methods.The authors studied 332 patients, age 65 years or older, hospitalized with a femoral neck fracture in three cities.Results.The population was 55% over age 80, 80% female, and lived in Houston (17%), Pittsburgh (29%), and Minneapolis (54%). Rates of hip replacement varied by city (Houston-84%, Pittsburgh-77%, Minneapolis-63%;P= 0.002), with great variability among patients with nondisplaced fractures (Houston-88%, Pittsburgh-77%, and Minneapolis-56%;P= 0.0001), and no variation among those with displaced fractures (P= 0.72). Other factors associated with hip replacement are history of hip fracture (P= 0.003) and cerebrovascular disease (P≤ 0.10), APACHE II-APS score (P= 0.09), and impacted fracture (P= 0.001). Sociodemographic and functional status (perceived health; activities of daily living and instrumental activities of daily living dependencies) were not associated with hip replacement (P> 0.10). In a logistic model controlling for prior history, APACHE II-APS, and fracture characteristics, city remained a significant predictor of hip replacement (P < 0.001).Conclusions.Despite an absence of evidence supporting its appropriateness and a much higher cost, hip replacement is used to treat nondisplaced fractures much more frequently in Houston and Pittsburgh than in Minneapolis.