Developing and Evaluating a Machine Learning-Driven Predictive Analytics Framework for Organizational Performance Measurement and Improvement: An Empirical Case Study of Apollo Hospitals, India

Abstract

This research aims at developing and validating a predictive analytics model using ML tools, which will be used for evaluating and enhancing the organizational performance of health care organizations. As a single case study, this research is conducted on the example of Apollo Hospitals and suggests using several types of ML approaches, namely, supervised learning, deep learning, and ensemble learning, in order to develop predictive tools that will be useful in predicting risks, issuing early warning signals, and optimizing operations. High efficacy of the proposed analytics model was confirmed by the results achieved, which included 80% reduction in Code Blue episodes, 70% reduction in nurses' workload, more than 90% of accuracy of empirical recommendations for antibiotic use, and even 26% of ward costs reduction, all achieved through the application of ML algorithms to the database containing over 400,000 patients' anonymous records.