Clinical vs. autopsy diagnostic discrepancies in the intensive care unit: a systematic review and meta-analysis of autopsy series

The aim of this study was to assess whether there is a discrepancy between clinical and autopsy-based diagnoses in adult intensive care unit (ICU) patients.

We conducted a systematic review of cohort studies reporting on conventional autopsy-confirmed missed diagnoses. The discrepancy rate was per study calculated by dividing the number of patients with a missed diagnosis by the number of autopsies. Missed diagnoses were classified according to the Goldman classification as ‘major’ and ‘minor’ with major missed diagnoses further differentiated into Class I missed diagnoses (i.e., diagnoses that may have altered therapy or survival) and Class II missed diagnoses (i.e., diagnoses that would not have altered therapy or survival). Class I missed diagnoses constitute the primary outcome of interest. Pooled estimates for discrepancy rates (95% confidence intervals) were calculated using a mixed-effects logistic regression model with ‘study’ as random effect. Meta-regression was used to assess relationships between major discrepancy rates and autopsy rates, start year of study, and ICU type.

Forty-two studies were identified totaling 6305 analyzed autopsies and 1759 patients with missed diagnoses. The pooled discrepancy rates for Class I and major missed diagnoses were 6.5% (5–8.5) and 19.3% (15.3–24), respectively. Meta-regression analysis revealed that autopsy rate was inversely associated with discrepancy rate. Class I discrepancy rates did not change over time. Burn and trauma ICUs had lower discrepancy rates as compared to medical ICUs, possibly because of higher autopsy rates.

Missed diagnoses remain common in ICUs. A higher autopsy rate does not reveal more major diagnostic errors. These data support a clinically driven autopsy policy rather than a systematic autopsy policy.