Process Mining Based Critical Path Recommendation in Healthcare Management

Abstract

From the literature it was studied that, most of the medical error was due to the faulty system/ process, because of which there is a delay in treatment management, leading to complications in later stages. Proper management of healthcare system is necessary to provide good medical care. Medical error due to failure in the healthcare system can be reduced by employing an appropriate clinical decision support system (CDSS). CDSS helps in identifying the severity of disease, predicting its progression, and recommending required resources for proper management of the disease. In the recent years, the information system is employed in the healthcare system to improve the management of healthcare. CDSS are being used to predict the disease progression and length of stay in the hospital. In our work, a CDSS was developed with the help of process mining techniques for providing improved treatment management. Process mining with its ability to build e cient process models was used for discovering this critical treatment path. The critical treatment path is a sequence of clinical and non-clinical activities that are critical. Process mining helps in stream-lining these activities along with the e cient resources for performing those activities. The gallstone disease treatment management is considered as a case study in this work. Modi ed Cascade Neural Network (ModCNN) was built upon the architecture of Cascade-Correlation Neural Network (CCNN) and, was trained and tested using the ADAptive LInear NEuron (ADALINE) circuit. In CDSS the performance of ModCNN was evaluated and compared with Arti cial Neural Network (ANN) and CCNN. CDSS, using ModCNN strati ed the cases that may need Endoscopic Retrograde CholangioPancreatography (ERCP) as the treatment progresses. Our result shows improvement in accuracy of prediction and reduction in waiting time. ModCNN showed better accuracy of 96:42% for predicting the disease progression when compared with CCNN (93:24%) and ANN (89:65%). CDSS developed in this work is aimed at providing better treatment planning to reduce medical error.