Predicting and reducing re-admission to hospital

This project has been completed

Between 2005 and 2007, The King's Fund was commissioned to produce software tools and models to identify individuals at high-risk of re-admission to hospital. The aim was to help primary care trusts intervene and reduce future hospital admissions.

Latest news

  • The Department of Health has announced that it has no plans to commission updated versions of the predictive risk models, PARR++ and the Combined Predictive Model.
  • Following this announcement, the existing models have been withdrawn from The King's Fund website. The predictive accuracy of the existing models is likely to have declined over time and, as the models were developed to work with HRG3.5 data, they are not compatible with HRG4 data.

Project background

In 2005, Essex Strategic Health Authority - on behalf of the Department of Health, the NHS Modernisation Agency and England's strategic health authorities - commissioned The King's Fund to develop a case-finding algorithm. The Department of Health took over the management of this contract in May 2007.

The King's Fund worked with Health Dialog and New York University to develop two risk stratification models:

  • the Patients at Risk of Re-hospitalisation (PARR) tool: a software tool that used inpatient data to identify patients at risk of re-hospitalisation within a year. The last version, PARR++ was released in November 2007.
  • the Combined Predictive Model: a model that used inpatient, outpatient, A&E and GP data to stratify populations according to their risk of admission. In order to run this model, a software front-end needed to be built locally. The last version of this model was released in December 2006.

The intellectual property rights for both models is owned by the Department of Health.

Available resources on PARR++ and long-term conditions

The following resources can still be downloaded from the column on the right-hand side of this page:

  • predictive risk project literature review: summary
  • predictive risk project literature review
  • PARR case finding report
  • Combined Predictive Model Final Report
  • Croydon PCT case study