Diagnostic errors; the need to have autopsies

In geriatric patients, atypical presentation and limitations in diagnostic scope may lead to underdiagnosis. The aim of this study was to establish the frequency, nature and causes of clinical diagnostic errors in a geriatric population.

A retrospective study.

We assessed the accuracy of clinical diagnosis using autopsy results as our gold standard. Factors likely to influence accuracy of clinical diagnosis were identified. We used the (modified) classification of Goldman et al. to define discrepancy.

We analysed 93 autopsies of a total of 331 deaths (28%). Discrepancies in major diagnoses were seen in 36 cases (39%). In 17 of these, clinical management might have been different if the diagnosis had been made premortem. These were: pulmonary embolism (4); unrecognised infection (4); intestinal ischaemia (3); ruptured aortic aneurysm (2); malignancy (1); tracheal obstruction (1); intestinal obstruction (1) and acute pancreatitis (1). Discrepancies in minor diagnoses were seen in 46 cases (50%). About one third of these were clinically relevant. Discrepancies occurred more frequently if there was a degree of uncertainty about clinical diagnosis p<0.001).

Major discrepancies between clinical diagnosis and autopsy findings were seen in 39% of our study population. They occur more often in the case of uncertain clinical diagnosis. Our findings stress the continuing and important role of autopsy in improving clinical practice in geriatric medicine.