Validity of certified causes of death in breast carcinoma patients.

The accuracy of the cause of death certification was surveyed in 802 breast carcinoma patients reported to the Swedish Cancer Registry from Stockholm county during 1961-1963. A total of 502 deaths occurred during the follow-up period of which 484 were eligible for the analysis. In 7 per cent the officially recorded underlying cause of death failed to accurately reflect the cancer status at death, which was mostly due to diagnostic errors or errors in the coding of the death certificate data. There seemed to be a net overestimation of intercurrent deaths in the official mortality statistics and a corresponding net underestimation of breast carcinoma deaths. Breast carcinoma was mentioned as a contributory cause of death in only 52 per cent of those certified to have died from intercurrent causes but who had recurrent disease at death. Among those with breast carcinoma as a certified contributory cause, 52 per cent did not have recurrent disease according to the clinical records. It is therefore concluded that official data on breast carcinoma as a contributory cause of death are of limited use in descriptive epidemiology.