Verifying causes of death in Thailand: rationale and methods for empirical investigation

BACKGROUND: Cause-specific mortality statistics by age and sex are primary evidence for epidemiological research and health policy. Annual mortality statistics from vital registration systems in Thailand are of limited utility because about 40% of deaths are registered with unknown or nonspecific causes. This paper reports the rationale, methods, and broad results from a comprehensive study to verify registered causes in Thailand. METHODS: A nationally representative sample of 11,984 deaths was selected using a multistage stratified cluster sampling approach, distributed across 28 districts located in nine provinces of Thailand. Registered causes were verified through medical record review for deaths in hospitals and standard verbal autopsy procedures for deaths outside hospitals, the results of which were used to measure validity and reliability of registration data. Study findings were used to develop descriptive estimates of cause-specific mortality by age and sex in Thailand. RESULTS: Causes of death were verified for a total of 9,644 deaths in the study sample, comprised of 3,316 deaths in hospitals and 6,328 deaths outside hospitals. Field studies yielded specific diagnoses in almost all deaths in the sample originally assigned an ill-defined cause of death at registration. Study findings suggest that the leading causes of death in Thailand among males are stroke (9.4%); transport accidents (8.1%); HIV/AIDS (7.9%); ischemic heart diseases (6.4%); and chronic obstructive lung diseases (5.7%). Among females, the leading causes are stroke (11.3%); diabetes (8%); ischemic heart disease (7.5%); HIV/AIDS (5.7%); and renal diseases (4%). CONCLUSIONS: Empirical investigation of registered causes of death in the study sample yielded adequate information to enable estimation of cause-specific mortality patterns in Thailand. These findings will inform burden of disease estimation and economic evaluation of health policy choices in the country. The development and implementation of research methods in this study will contribute to improvements in the quality of annual mortality statistics in Thailand. Similar research is recommended for other countries where the quality of mortality statistics is poor.