Predictive Risk Model Algorithm

This page contains information on the Predictive Risk Model (PRM) algorithm used in the Health Care Homes (HCH) Risk Stratification Tool (RST).

Page last updated: 21 March 2019

Introduction

The HCH Trial uses a RST developed by Precedence Health Care to identify patients who may be eligible to participate in the program. The PRM algorithm and source code forms the scientific basis for the RST to identify these patients.

What is the PRM algorithm?

The PRM algorithm uses general practice data to determine the probability of a patient having an unplanned hospitalisation over one year.

How was it developed?

The CSIRO developed and validated the PRM algorithm using a retrospective primary care cohort, linked to records of hospitalisation in Victoria, to predict risk of hospitalisation within one year.

There are 95 predictors in the PRM which can be categorised under patient demographics, history of prescriptions, pathology results and disease diagnosis. Prior hospitalisation information was not used as a predictor.

This is the first published validated risk model designed specifically to be applied in Australia.

The full article 'A risk stratification tool for hospitalisation in Australia using primary care data' is available at Nature International Journal of Science website.

Access to the PRM algorithm

The PRM algorithm is available for free under an Apache 2.0 open source license.

The downloadable PRM algorithm source code will be available on this page shortly.