Send to:

Choose Destination
See comment in PubMed Commons below
Eur J Intern Med. 2009 Dec;20(8):784-7. doi: 10.1016/j.ejim.2009.08.005. Epub 2009 Sep 18.

Diagnostic performance in a primary referral hospital assessed by autopsy: evolution over a ten-year period.

Author information

  • 1Medizinische Klinik, Kantonsspital Münsterlingen, Münsterlingen, Switzerland.



Despite remarkable progress in modern laboratory testing and imaging technology in recent years, diagnostic errors still occur. To assess whether diagnostic performance in a primary referral hospital improves with new diagnostic tools and algorithms, autopsy reports were analyzed over a ten-year period to monitor diagnostic errors.


Medical reports from 1997 to 2006 were compared retrospectively with autopsy reports. A diagnostic error was assumed when the main clinical diagnosis was missed, independently of whether this influenced the patient's survival or whether this error led to incorrect treatment without effect on survival. Two cardiovascular markers with high sensitivity, namely cardiac troponin T and D-dimer testing and two algorithms for thoracic pain and thromboembolic disease were introduced during the study period.


970 cases were included; the autopsy rate was 50.1%. Cardiovascular diseases were misdiagnosed in 18.7%, followed by infectious diseases in 12.9%, oncological 3.6% and neurological diseases in 1.8%. The most commonly missed diagnoses were myocardial infarction, pulmonary embolism and aortic dissection; however, the rate of errors for cardiovascular diseases decreased over the 10 years (p<0.002). Overall diagnostic sensitivity and specificity rose from 67% to 87% and from 94% to 99%, respectively.


Autopsy remains a valuable tool to measure diagnostic performance. Errors occur most frequently in cardiovascular events, whereas in malignant and neurological diseases they are rare. The significant improvement of diagnostic accuracy for cardiovascular diseases is associated with the introduction of new sensitive laboratory tests and algorithms for thoracic pain and thromboembolic diseases.

[PubMed - indexed for MEDLINE]
PubMed Commons home

PubMed Commons

How to join PubMed Commons

    Supplemental Content

    Full text links

    Icon for Elsevier Science
    Loading ...
    Write to the Help Desk