High value predictive analytics in digital health systems

Flagship Program: Intelligent decision support to improve value and efficiency

Project Description

This project will develop and test predictive modelling approaches designed to improve clinical decision support in real-time acute settings. Statistical methods in clinical care often focus on the prediction or retrospective evaluation of patient outcomes rather than prevention. This project seeks to examine whether statistical and machine learning methods can prospectively identify trends in the data using the case study of clinical deterioration in the hospital. This can lead to more informed, patient-centred, cost-effective care.

Project Objectives

This project has five primary objectives:

  1. Systematically identify and summarise predictive models currently available for clinical deterioration prediction models
  2. Formulate a theoretical framework for the role of predictive analytics in an operational context in healthcare delivery
  3. Assess the use of predictive analytics currently used for clinical decision support across Metro South hospitals
  4. Develop, validate, and evaluate a predictive model for near real-time decision support to predict and pre-empt clinical deterioration using embedded electronic systems
  5. Model the potential costs, quality of life, and cost-effectiveness of successful implementation for models with the best predictive power in a near real-time setting.

This work forms part of a larger project.

Industry Participants

Metro South Hospital and Health Service

Research Participant

Queensland University of Technology
Robin Blythe, PhD student


DHCRC PhD Top-up Scholarship