Relating faults in diagnostic reasoning with diagnostic errors and patient harm

Acad Med. 2012 Feb;87(2):149-56. doi: 10.1097/ACM.0b013e31823f71e6.

Abstract

Purpose: The relationship between faults in diagnostic reasoning, diagnostic errors, and patient harm has hardly been studied. This study examined suboptimal cognitive acts (SCAs; i.e., faults in diagnostic reasoning), related them to the occurrence of diagnostic errors and patient harm, and studied the causes.

Method: Four expert internists reviewed patient records of 247 dyspnea patients, using a specially developed questionnaire to detect SCAs. The patients were treated by 72 physicians between May 2007 and February 2008 in five Dutch hospitals. The findings of the record review were discussed with the treating physicians, and the causes of SCAs were classified using Reason's taxonomy of unsafe acts. Statistical analyses were performed with descriptive statistics and independent t tests to compare groups. Furthermore, a reliability study was conducted to assess the interrater reliability.

Results: SCAs occurred in 163 of 247 cases reviewed (66%). In 34 (13.8%) of all cases, a diagnostic error occurred, and in 28 (11.3%) cases, the patient was harmed. Cases with diagnostic errors or patient harm had more SCAs. However, in 10 (4.0%) of the cases, diagnostic errors or patient harm occurred, though there were no SCAs. The causes of SCAs were mostly mistakes (i.e., the planned action was incorrect).

Conclusions: In cases with more SCAs, diagnostic errors and patient harm occurred more often, suggesting that the number of SCAs per case was predictive of the occurrence of these events. The most common causes were mistakes, meaning that physicians did not realize their actions were incorrect.

Publication types

  • Multicenter Study
  • Research Support, Non-U.S. Gov't

MeSH terms

  • Clinical Competence
  • Cognition
  • Decision Making*
  • Delphi Technique
  • Diagnostic Errors / statistics & numerical data*
  • Dyspnea / diagnosis*
  • Humans
  • Internal Medicine
  • Medical Records
  • Netherlands
  • Physicians / psychology*
  • Physicians / statistics & numerical data*
  • Reproducibility of Results
  • Retrospective Studies
  • Surveys and Questionnaires