RESUMEN
OBJECTIVES: Describe the completeness and quality of Jamaica's 2008 vital registration mortality database. STUDY DESIGN AND SETTING: Multiple sources (hospitals, police, forensic pathologists, Coroners courts) were used to validate deaths registered as occurring in 2008. A 10% random sample was examined to evaluate the quality of certification and coding. Jamaica, a middle-income country of 2.7 million, began vital registration in 1877; however, the mortality database was considered of limited use, and the study was commissioned to understand the problem. RESULTS: Of 19,286 deaths identified, 76% were registered by 31.12.2009 for inclusion among 2008 demographic returns. Registration was highest among deaths not requiring autopsy (94%) and lowest among Coroners cases (22%) with only 41% of deaths among 15-44 year-olds registered. The leading causes of death were cerebrovascular disease, diabetes mellitus, and homicide. Fifteen percent were coded to ill-defined causes of death. Recoding the sample increased mortality from prematurity, pregnancy complications, homicide, selected cardiovascular disorders, and human immunodeficiency virus/AIDS. CONCLUSION: Delays registering Coroners cases, certification and coding errors, introduced biases into the disease profile limiting the data's value in informing clinical care. Issuance of medical certificates by pathologists who investigate Coroners cases and training physicians and coders would eliminate most underreporting and improve data quality.