OBJECTIVE: To assess the proportion of in-hospital versus in-nursing home deaths among a population of decedent nursing home residents in British Columbia, Canada, and to identify facility and individual characteristics associated with in-hospital death. METHODS: We examined nursing home (ownership/organization, size) and individual (age, level of care, sex, previous hospitalization within 30 days) characteristics of all decedent residents of British Columbia's freestanding publicly funded nursing homes. Secondary administrative data from the Ministry of Health, supplemented with facility data were analyzed. The study population included those aged 65 years and older who died between April 1, 1996 and August 1, 1999 (n = 14,413). Mixed models were used to estimate unadjusted and adjusted odds ratios (AOR; 95% confidence intervals [CI]) for factors associated with in-hospital death. RESULTS: Almost one quarter (24.6%) of deaths occurred in hospital. In-hospital death was more frequent in nonprofit (NP) single-site facilities compared to NP facilities owned and/or operated by a health authority (AOR = 1.37, 95% CI: 1.15, 1.64). Smaller nursing home size (AOR = 1.25, 95% CI: 1.05, 1.50) and male gender (AOR = 1.17, 95% CI: 1.07, 1.27) were also associated with a greater odds of in-hospital death. Progressively lower odds ratios of in-hospital death were observed for each category of increasing age and declining function, respectively. CONCLUSIONS: While individual characteristics play a significant role in explaining variation in site of death, residence in a NP single-site and smaller-sized facility was also associated with a greater frequency of in-hospital death.