Mining Misdiagnosis Patterns from Biomedical Literature

TL;DR

This study analyzes PubMed abstracts to identify patterns in diagnostic errors, revealing that common misdiagnoses often involve multiple low-frequency errors and asymmetric relationships between diseases.

Contribution

It introduces a novel method for mining misdiagnosis patterns from biomedical literature using phrase-based selection and graph modeling.

Findings

Common misdiagnosed diseases are linked to many different low-frequency errors.

Misdiagnosis relationships are often one-sided rather than reciprocal.

The directed graph effectively visualizes misdiagnosis patterns.

Abstract

Diagnostic errors can pose a serious threat to patient safety, leading to serious harm and even death. Efforts are being made to develop interventions that allow physicians to reassess for errors and improve diagnostic accuracy. Our study presents an exploration of misdiagnosis patterns mined from PubMed abstracts. Article titles containing certain phrases indicating misdiagnosis were selected and frequencies of these misdiagnoses calculated. We present the resulting patterns in the form of a directed graph with frequency-weighted misdiagnosis edges connecting diagnosis vertices. We find that the most commonly misdiagnosed diseases were often misdiagnosed as many different diseases, with each misdiagnosis having a relatively low frequency, rather than as a single disease with greater probability. Additionally, while a misdiagnosis relationship may generally exist, the relationship was often found to be one-sided.