Death certification and epidemiological research. Medical Services Study Group of the Royal College of Physicians of London.

The cause of death shown on 191 death certificates was compared with the cause indicated by the hospital case notes, the consultants'' opinions, and the necropsy findings. All 191 deaths occurred among medical hospital patients aged under 50. In 39 cases there was a major discrepancy between the two sources over the cause of death and in another 54 ther was a minor but epidemiologically important difference. Death certificates are not primarily intended for epidemiological research, but researchers often rely on them. This and other studies have shown, however, that death certificates are often inaccurate records of the cause of death--even coroner''s certificates issued after a coroner''s necropsy. The accuracy of death certificates might be improved if coroners consulted clinicians more closely and if senior hospital staff completed hospital death certificates.