OBJECTIVE To investigate necropsy request practices at the University Hospital of the West Indies, Jamaica, to determine the extent to which these might influence the declining necropsy rates. This is the first such study from a developing country. METHODS The necropsy service was audited prospectively over a six month period, and data relating to non-coroner's (hospital) necropsy requests, including the clinical service and post of the clinician involved, were documented. The reasons for non-request were recorded for deaths in which a necropsy was not requested, in addition to the reasons given by pathologists for not performing necropsies in cases that were requested but not done. The overall, non-coroner's, and coroner's necropsy rates in addition to the non-coroner's necropsy request and success rates were calculated. RESULTS There were 364 deaths comprising 323 non-coroner's and 41 coroner's cases. The overall, non-coroner's, and coroner's necropsy rates were 29.2%, 20.2%, and 38.7%, respectively. The non-coroner's necropsy request rate was 35.3% with a success rate of 65%. Seventy five per cent of the requests were made by non-consultant clinicians and on the internal medicine service, which accounted for most of the non-coroner's deaths; necropsy requests were biased towards younger patients (p < 0.0001). Confident clinical diagnosis was the main reason for not requesting a necropsy, and the primary reason for refusing to perform a necropsy was that the request had been made too long after death. CONCLUSIONS These findings show a relatively high necropsy success rate in the face of a comparatively low necropsy request rate, and indicate that necropsy rates can be increased if clinicians make more necropsy requests in a timely manner in patients of all ages.