Output details
11 - Computer Science and Informatics
University of Westminster
Towards an evidence-based decision making healthcare system management: modelling patient pathways to improve clinical outcomes
<15> Originality: We introduce a novel method based on random effects continuation-ratio logit model, suitable for detecting stage wise transitions, to patient pathways modelling. We apply it to a large dataset extracted from the national English Hospital Episodes Statistics dataset. The tool is useful in detecting the most critical threshold at which multiple readmissions are more probable.
Significance: The paper proposes a statistical data mining solution to the important problem of investigating patient readmission progression, and more generally failure repetition.
Rigour: A general statistical model is formulated and tested on a large national dataset. Results published in leading peer-reviewed journal.