Using Open Clinical Data to Create an Embeddable Prediction System for Hospital Stay
With the ever increasing availability of open, clinical health data, there exists a deficiency of platforms to take advantage of it . The global prevalence of diabetes has risen from 4.7% in 1980 to 8.5% in 2014 and continues to rise, placing an increased demand on hospital resources . The management of diabetic patients within hospital can be assisted by the accurate prediction of length of stay (LOS) of patients. This paper introduces the use of Bayesian networks (BN) to accurately predict patient LOS in hospital. The results show the tree augmented naive BN classifier to be the most effective in predicting LOS. We believe that our model can be implemented by hospitals to more efficiently utilize hospital resources.
KeywordsLength of stay Bayesian network Health care system Diabetes
The work is partially supported by Dr. Mago’s NSERC Discovery Grant. All authors also like to thank Darryl Willick for his support to run the experiments on High Performance Computing Lab at Lakehead University.
- 1.Lichman, M.: UCI machine learning repository (2013)Google Scholar
- 2.World Health Organization, et al.: Global report on diabetes. World Health Organization (2016)Google Scholar
- 3.OECD Indicators. Health at a glance (2005)Google Scholar
- 16.Tsamardinos, I., Aliferis, C.F., Statnikov, A.: Time and sample efficient discovery of Markov blankets and direct causal relations. In: Proceedings of the Ninth ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, pp. 673–678. ACM (2003)Google Scholar
- 17.Tsamardinos, I., Aliferis, C.F., Statnikov, A.R., Statnikov, E.: Algorithms for large scale Markov blanket discovery. In: FLAIRS Conference, vol. 2, pp. 376–380 (2003)Google Scholar