Accuracy of admission and clinical diagnosis of tumours as revealed by 2000 autopsies.

Admission, clinical and autopsy diagnoses of tumour were computed in 2000 consecutive cases, aged 30-80 years, using data collected in two university pathology departments in Budapest, Hungary. Based on diagnosis of tumour, regardless of site, as the underlying cause of death false-negative rates were 37.4% at admission and 8.8% clinically. Corresponding false-positive rates were 8.4 and 9.1%. General practitioners who correctly diagnosed a tumour as the cause of the terminal illness identified the primary site wrongly in 20.6% (90/436) of cases. Hospital clinicians did so in 20.4% (130/636) of cases. Overall, of site-specific tumours considered as the underlying cause of death at autopsy, 27.4% were incorrectly diagnosed clinically and 50.4% at admission. Diagnostic errors were particularly common for tumours of the lung, liver, ovary and gall bladder. Graduate and postgraduate education, planning of the health care system and quality of cancer care may benefit from statistical data derived from autopsy diagnoses.