DiNardo Deborah, Tilstra Sarah, McNeil Melissa, Follansbee William, Zimmer Shanta, Farris Coreen, Barnato Amber E
Department of Medicine, University of Pittsburgh School of Medicine, Pittsburgh, PA, USA.
University of Colorado School of Medicine, Denver, CO, USA.
Diagnosis (Berl). 2018 Mar 28;5(1):21-28. doi: 10.1515/dx-2017-0037.
While there is some experimental evidence to support the use of cognitive forcing strategies to reduce diagnostic error in residents, the potential usability of such strategies in the clinical setting has not been explored. We sought to test the effect of a clinical reasoning tool on diagnostic accuracy and to obtain feedback on its usability and acceptability.
We conducted a randomized behavioral experiment testing the effect of this tool on diagnostic accuracy on written cases among post-graduate 3 (PGY-3) residents at a single internal medical residency program in 2014. Residents completed written clinical cases in a proctored setting with and without prompts to use the tool. The tool encouraged reflection on concordant and discordant aspects of each case. We used random effects regression to assess the effect of the tool on diagnostic accuracy of the independent case sets, controlling for case complexity. We then conducted audiotaped structured focus group debriefing sessions and reviewed the tapes for facilitators and barriers to use of the tool.
Of 51 eligible PGY-3 residents, 34 (67%) participated in the study. The average diagnostic accuracy increased from 52% to 60% with the tool, a difference that just met the test for statistical significance in adjusted analyses (p=0.05). Residents reported that the tool was generally acceptable and understandable but did not recognize its utility for use with simple cases, suggesting the presence of overconfidence bias.
A clinical reasoning tool improved residents' diagnostic accuracy on written cases. Overconfidence bias is a potential barrier to its use in the clinical setting.
虽然有一些实验证据支持使用认知强制策略来减少住院医师的诊断错误，但尚未探讨此类策略在临床环境中的潜在可用性。我们试图测试一种临床推理工具对诊断准确性的影响，并获取关于其可用性和可接受性的反馈。
2014年，我们在一个内科住院医师培训项目中进行了一项随机行为实验，测试该工具对三年级住院医师（PGY-3）书面病例诊断准确性的影响。住院医师在有监考的环境中完成书面临床病例，分为有使用该工具提示和无使用该工具提示两种情况。该工具鼓励对每个病例的一致和不一致方面进行反思。我们使用随机效应回归来评估该工具对独立病例集诊断准确性的影响，并控制病例复杂性。然后，我们进行了录音结构化焦点小组汇报会，并审查录音以找出使用该工具的促进因素和障碍。
在51名符合条件的PGY-3住院医师中，34名（67%）参与了研究。使用该工具后，平均诊断准确性从52%提高到60%，在调整分析中，这一差异刚刚达到统计学显著性检验（p=0.05）。住院医师报告说，该工具总体上是可接受和可理解的，但没有认识到其在简单病例中的效用，这表明存在过度自信偏差。
一种临床推理工具提高了住院医师在书面病例上的诊断准确性。过度自信偏差是其在临床环境中使用的潜在障碍。
