Predicting re-hospitalisations using intelligent systems: an exploratory study

  • Authors:
  • John W. Showalter;Girish H. Subramanian

  • Affiliations:
  • University of Mississippi Medical Centre, 2500 North State, Jackson, MS 39216, USA;School of Business Administration, E355 Olmsted Building, Penn State Harrisburg, 777 W. Harrisburg Pike, Middletown, PA 17057, USA

  • Venue:
  • International Journal of Business Information Systems
  • Year:
  • 2013

Quantified Score

Hi-index 0.00

Visualization

Abstract

Intelligent systems and knowledge management are shown to have wide ranging applications for businesses. In healthcare, applications of expert systems and knowledge management have been popular since the use of MYCIN in the 1970s. In this research, we study the use of intelligent systems in the prediction of re-hospitalisations. A Bayesian network, an artificial neural network using multilayer perception, an artificial neural network using radial basis function, a support vector machine and an expert system five intelligent systems were used in the prediction. All the models except the support vector machine were able to provide good prediction of re-hospitalisation with AUC > 0.6 and complete risk stratification for re-hospitalisation within 30 days of discharge with data available at admission, showing great promise for future application of intelligent systems in healthcare.