Linking uncertainty in physicians' narratives to diagnostic correctness

  • Authors:
  • Wilson McCoy;Jeff B. Pelz;Cecilia Ovesdotter Alm;Pengcheng Shi;Cara Calvelli;Anne Haake

  • Affiliations:
  • Rochester Institute of Technology;Rochester Institute of Technology;Rochester Institute of Technology;Rochester Institute of Technology;Rochester Institute of Technology;Rochester Institute of Technology

  • Venue:
  • ExProM '12 Proceedings of the Workshop on Extra-Propositional Aspects of Meaning in Computational Linguistics
  • Year:
  • 2012

Quantified Score

Hi-index 0.00

Visualization

Abstract

In the medical domain, misdiagnoses and diagnostic uncertainty put lives at risk and incur substantial financial costs. Clearly, medical reasoning and decision-making need to be better understood. We explore a possible link between linguistic expression and diagnostic correctness. We report on an unusual data set of spoken diagnostic narratives used to computationally model and predict diagnostic correctness based on automatically extracted and linguistically motivated features that capture physicians' uncertainty. A multimodal data set was collected as dermatologists viewed images of skin conditions and explained their diagnostic process and observations aloud. We discuss experimentation and analysis in initial and secondary pilot studies. In both cases, we experimented with computational modeling using features from the acoustic-prosodic and lexical-structural linguistic modalities.