Abstract

AIMS: To assess the frequency with which the cause of death on death certificates included the relevant details requested of certifying doctors, especially in deaths due to malignant disease, but also including certain other deaths where specific information would be expected to be included. METHODS: Consecutive series of certificates attributing death to malignancy, pneumonia, an acute cerebrovascular event, and renal failure were inspected and compared with the categories identified in the International Classification of Disease. Review of clinical notes and of laboratory data was used to determine the number of cases in which detailed histological diagnoses were available. RESULTS: A histological diagnosis was available in 79.1% of cases of deaths due to malignancy, but was recorded on only 23.6% of certificates. Haematologists performed best (69.6%) and general surgeons worst (2.8%). The sites of primary tumours were recorded in detail in only 23 of 89 cases of tumours of the large bowel (22/36), lung (1/35) and stomach (0/18). In cases of pneumonia the causative organism was recorded in only 4 of 330. In cases of an acute cerebrovascular event one of 70 was recorded as being due to haemorrhage. A distinction between cerebral or precerebral arterial occlusion (embolism/thrombosis) and cerebral haemorrhage was not recorded in any of the other cases. In cases of renal failure a cause was not recorded in 75 of 95. CONCLUSIONS: Despite consistent encouragement to record all relevant details on death certificates this study shows that doctors fail to do so in most cases. Such a failure diminishes information available to the Office of Population Censuses and Surveys, affecting mortality statistics and gives further cause for concern about standards of certification. Means by which the standard of certification might be improved are discussed, including screening of certificates by a medically qualified person prior to registration.