Background:the demand for residential respite care for older persons is high yet little is known about the occurrence of harm, including death in this care setting. Objective:to compare the prevalence and nature of deaths among residential respite to permanent nursing home residents. Design:retrospective cohort study. Setting:australian accredited nursing homes between 1 July 2000 and 30 June 2013. Subjects:respite and permanent residents of Australian accredited nursing homes, whose deaths were investigated by Australian coroners. Methods:prevalence of deaths of nursing home residents were calculated using routinely generated coronial data stored in the National Coronial Information System. Odds ratios (OR) were calculated to examine residency (respite or permanent) by cause of death. Results:of the 21,672 residents who died during the study period, 172 (0.8%) were in respite care. The majority of deaths were due to natural causes. A lower proportion occurred in respite (n = 119, 69.2%) than permanent (n = 18,264, 84.9%) residents. Falls-related deaths in respite as a proportion (n = 41, 23.8%) was almost double that in permanent care (n = 2,638, 12.3%). Deaths from other injury-related causes (such as suicide and choking) were significantly more likely in respite residents (OR = 2.0; 95% confidence interval: 1.1-3.6; P = 0.026). Conclusions:this is the first national cohort study examining mortality among respite residents. It established that premature, injury-related deaths do occur during respite care. This is the first step towards better understanding and reducing the risk of harm in respite care.