Background and purpose — The risk of dislocation after primary total hip arthroplasty (THA) is affected by several factors, which increases the possibility of substantial differences among hospitals. We compared cumulative incidences of dislocation between regions and hospitals after primary THA surgery in osteoarthritis (OA) patients. Patients and methods — From the Danish Hip Arthro-plasty Register, we included 31,105 THAs performed from 2010 to 2014 with 2 years’ follow-up. Dislocations treated by closed reduction were identified in the Danish National Patient Register combined with patient file review. The results are presented as 2-year cumulative incidence on national, regional, and hospital level as proportions with 95% confidence intervals (CI) and as adjusted odds ratios (OR) analyzed by multiple logistic regression. Results — 1,861 dislocations in 1,079 THAs were identified from 59 orthopedic departments. The 2-year cumulative incidence ranged from 2.2% to 4.3% between the 5 regions in Denmark. Hospital variation was 0–12%. For hospitals with a 5-year volume of more than 100 procedures, the incidence was 0.9–7.4%. Using the highest volume hospital as the ref-erence, ORs for dislocation for the remaining hospitals were between 0.3 (CI 0.1–0.6) and 2.7 (1.9–4.0) after adjusting for age, sex, head size, and fixation method. Low-volume hospitals showed a higher dislocation risk than high-volume units with an adjusted OR of 1.2 (1.1–1.4). Interpretation — We found substantial variation in the incidence of dislocation between hospitals within Denmark, which can be explained by a combination of patient-, com-ponent-, and surgery-related factors. There is a need for continuous monitoring of THA dislocation on regional and hospital levels to reduce overall and local dislocation risk for future patients.