Reliability of death certificates. The reproducibility of the recorded causes of death in patients admitted to departments of internal medicine

: The aim of the study was to assess the reproducibility when different doctors fill in diagnoses on death certificates. Records from 40 patients who had died in 1994 during admission to a medical hospital department in Denmark were selected at random. Ten doctors filled in a death certificate for each patient (without knowledge of autopsy findings and results of examinations received after the patient's death). The agreement between the diagnoses was assessed using a rating scale with seven categories. Afterwards the 400 death certificates were mixed and coded by the Medicostatistical Section of the Danish National Board of Health. The diagnoses made by the ten doctors showed insignificant discrepancies in ten cases, larger discrepancies were found in eight cases and large discrepancies in 19 cases. In three cases the patient had died suddenly and little information was available. The coding was standardized at the Danish National Board of Health, but their diagnoses reflected the discrepancies between the doctors' diagnoses. In conclusion, the reproducibility of diagnoses on death certificates is so poor that information from the Registry of Causes of Death is of little use of administrative or scientific purposes.