Errors in Clinical Statements of Causes of Death

Abstract

Through private funds, 25,000 records made prior to July 1, 1941, have been completed from a planned assembly of data on 100,000 autopsy deaths between the years 1900 and 1939 occurring in general hospitals where there was full-time pathologist service and a high average ratio of autopsies to deaths. The data used in the report are abstracted from official case history and autopsy protocols made accessible through administrators and record-room clerks of 15 hospitals of New York City, Albany, Schenectady, Rochester and Buffalo, New York, and in Jersey City, New Jersey, for the 10-yr. period 1930 to 1939. The accuracy of clinical statements on causes of death varies from one possible extreme to the other. Chronic diseases requiring several long stays in the hospital and verified by use of ample diagnostic facilities, such as x-rays and laboratory tests, are diagnosed accurately by clinical examination. In the first line of accuracy among important disease stands pulmonary tuberculosis with slightly less than 98% accurate clinical diagnosis. Next in accuracy of diagnosis come the acute communicable diseases[long dash]diphtheria, measles, whooping cough, scarlet fever, where 96 to 98% of all clinical diagnoses of causes of death are correct. Most mistakes of record in death certification of these diseases are due to mentioning secondary stages of such diseases instead of the origin of the infection, for example otitis media and encephalitis. Then follow deaths due to accidents, the causes of which usually are quite accurately stated.