A Predictive Analytics Framework Using Machine Learning Techniques for Measuring and Improving Organizational Performance: A Review and Study of Apollo Hospitals

Abstract

Predictive analytics based on machine learning (ML) has been proven to be a revolutionary technique used to assess and optimize organizational performance within the realm of healthcare. The current review paper provides an overview of the literature about the frameworks for predictive analytics used in healthcare with reference to the application of the latter in the case of Apollo Hospitals, which is India's most prestigious chain of multi-specialty hospitals. The use of ML-based predictive analytics and prescriptive analysis together with the Six Sigma approach was successfully employed at Apollo Hospitals in terms of different predictive tools such as Apollo Artificial Intelligence Cardiovascular Disease Risk (AICVD) tool, AI early warning system (AI EWS), AI empirical antibiotic recommendation system (AI EARS), and operative risk prediction tools. >90% accuracy in antibiotic prescription, 80% decrease in code blue incidents, 70% cut in nurse workload, and increased emphasis on prevention, leading to fewer procedures being conducted. The suggested architecture, consisting of data input, development of ML model (using supervised learning and deep learning techniques), measurement of success via clinical, operational, and financial KPIs, and ongoing improvement cycles, can provide a template for excellence through data science in hospitals. Issues such as data security, bias in models, and compatibility with existing systems have been addressed, along with future trends including generative AI and edge computing.