AHRQ Study Proposes Fault Tree Analysis To Identify Diagnostic Errors

Fault tree analysis, a technique that originated in private industry to test systems’ reliability and safety, offers promise in identifying factors that lead to medical diagnostic errors, according to a new AHRQ-funded study. Fault tree analysis has been used to reduce errors in high-risk nuclear, aerospace and chemical industries. To explore the usefulness of fault tree analysis in health care, researchers prepared fault trees for 10 patient case histories involving diagnostic errors. Those examples were then synthesized into a single fault tree to show common situations that can lead to diagnostic errors. Researchers found that the visual aspects of fault tree analysis provided a clear, unified illustration of the causes of diagnostic errors. Fault tree analysis also provided a framework for developing algorithms and computer software for the early detection of diagnostic errors, researchers concluded. The study, “Using Fault Trees to Advance Understanding of Diagnostic Errors,” and abstract appeared in The Joint Commission Journal on Quality and Patient Safety.

2017 Nov;43(11):598-605. doi: 10.1016/j.jcjq.2017.06.007. Epub 2017 Oct 2.

Using Fault Trees to Advance Understanding of Diagnostic Errors.

Rogith D, Iyengar MS, Singh H.

Abstract

PROBLEM DEFINITION:

Diagnostic errors annually affect at least 5% of adults in the outpatient setting in the United States. Formal analytic techniques are only infrequently used to understand them, in part because of the complexity of diagnostic processes and clinical work flows involved. In this article, diagnostic errors were modeled using fault tree analysis (FTA), a form of root cause analysis that has been successfully used in other high-complexity, high-risk contexts. How factors contributing to diagnostic errors can be systematically modeled by FTA to inform error understanding and error prevention is demonstrated.

INITIAL APPROACH:

A team of three experts reviewed 10 published cases of diagnostic error and constructed fault trees. The fault trees were modeled according to currently available conceptual frameworks characterizing diagnostic error. The 10 trees were then synthesized into a single fault tree to identify common contributing factors and pathways leading to diagnostic error.

KEY INSIGHTS:

FTA is a visual, structured, deductive approach that depicts the temporal sequence of events and their interactions in a formal logical hierarchy. The visual FTA enables easier understanding of causative processes and cognitive and system factors, as well as rapid identification of common pathways and interactions in a unified fashion. In addition, it enables calculation of empirical estimates for causative pathways. Thus, fault trees might provide a useful framework for both quantitative and qualitative analysis of diagnostic errors.

NEXT STEPS:

Future directions include establishing validity and reliability by modeling a wider range of error cases, conducting quantitative evaluations, and undertaking deeper exploration of other FTA capabilities.