Information visualization for chronic disease risk assessment

Christopher A. Harle, Daniel B. Neill, Rema Padman

Research output: Contribution to journalArticlepeer-review

14 Scopus citations

Abstract

Here, the authors describe and evaluate a new information-visualization method and prototype software tool that support risk assessment for negative health outcomes. Their framework uses principal component analysis and linear discriminant analysis to plot high-dimensional patient data in 2D. It also incorporates interactive visualization techniques to aid the identification of high versus low risk patients, critical risk factors, and the estimated effect of hypothetical interventions on the likelihood of negative outcomes. The authors quantitatively evaluated the visualization method using a secondary dataset describing 588 people with diabetes and their estimated future risk of heart attack. Their results show that the method visually classifies high-and low-risk people with accuracy that's similar to other common statistical methods. The framework also provides an interactive, visualization-based tool for clinicians to explore the nuances of their patients' data and disease risk.

Original languageEnglish (US)
Article number6365201
Pages (from-to)81-85
Number of pages5
JournalIEEE Intelligent Systems
Volume27
Issue number6
DOIs
StatePublished - Dec 17 2012
Externally publishedYes

Keywords

  • dimensionality reduction
  • healthcare
  • information visualization
  • risk assessment

ASJC Scopus subject areas

  • Computer Networks and Communications
  • Artificial Intelligence

Fingerprint Dive into the research topics of 'Information visualization for chronic disease risk assessment'. Together they form a unique fingerprint.

Cite this