Diagnostic errors in the new millennium: a follow-up autopsy study

Diagnostic accuracy has long been a focus in clinical medicine. Prior research has shown that in the decades before 2000, autopsy studies indicated a decline in major diagnostic errors. However, the extent of this trend remained unclear. No prior work had resolved whether this decline continued into the new millennium. Autopsy rates have also been decreasing globally, raising concerns about missed clinical diagnoses. This gap motivated a new study to assess diagnostic accuracy in the 2000s. The researchers aimed to determine if the observed decline in errors persisted. They also sought to explore the role of diagnostic technologies in this trend. The study focused on a specific hospital setting in Switzerland. This paper's contribution is to provide updated data from the early 2000s.

The study aimed to evaluate diagnostic accuracy in the early 2000s by comparing clinical diagnoses with autopsy findings. It sought to determine whether the decline in major diagnostic errors observed in prior decades continued. The researchers focused on a single tertiary-care hospital in Switzerland. They selected 100 randomly chosen deceased patients for analysis. The goal was to assess the frequency of major and minor diagnostic errors. The study also aimed to track changes in autopsy rates and diagnostic procedures. Researchers wanted to understand the impact of new technologies like CT scans. This work sought to provide updated evidence on diagnostic trends in clinical medicine.

The researchers conducted a follow-up study in 2002 at a tertiary-care hospital in Switzerland. They selected 100 deceased patients from medical wards and the ICU. Autopsy reports were compared with clinical diagnoses for each patient. The study categorized errors into class I and II based on severity. Researchers calculated error rates and compared them with historical data from 1972–1992. They also assessed sensitivity and specificity for major disease categories. The number of diagnostic procedures per patient was recorded and analyzed. The study included statistical comparisons of trends over time.

Major diagnostic errors decreased from 30% to 7% between 1972–1992 and 2002 (P<0.001). Class I errors dropped from 16% to 2% in the same period (P<0.001). Sensitivity for cardiovascular diseases rose from 69% to 92% (P=0.006). Infectious disease sensitivity increased from 25% to 90% (P=0.013). Neoplastic disease sensitivity reached 100% (P=0.053). Specificity for cardiovascular diseases improved from 85% to 98% (P<0.001). The number of diagnostic procedures per patient rose from 144 to 281 (P<0.001). CT scans and tissue sampling increased significantly in the study period.

The study found a continued decline in major diagnostic errors at the start of the new millennium. Researchers observed a significant drop in class I errors compared to the 1990s. Sensitivity improved across multiple disease categories, particularly for cardiovascular and infectious diseases. Specificity for cardiovascular diagnoses also increased notably. The number of diagnostic procedures per patient rose sharply in the decade before 2002. The authors suggest that new diagnostic tools may have contributed to this trend. However, they note that autopsy rates remained low at 54% in 2002. The findings support the idea that clinical accuracy has improved over time.