Integrating Machine Learning and Big Data Analytics to Transform Patient Outcomes in Chronic Disease Management

Abstract

Chronic diseases such as diabetes, cancer, and cardiovascular diseases account for nearly 70% of the total healthcare costs that can have a much broader negative impact on the quality of life of patients with decreased life expectancies, productivity loss, and increased healthcare costs. Therefore, a concerted effort is required to alleviate the burden of chronic diseases. Machine learning is becoming more ubiquitous in healthcare because of the exponential growth of electronic health records and the substantial advancement of big data analytics capability. In this study, systematic literature review approaches are employed to identify the machine learning and big data analytics technologies that are already implemented in chronic disease management. Each technology is thoroughly examined in terms of its definition, rationale, and types. In addition, deep coverage of implementation studies of the technologies is provided regarding the motivation, objective, methodology, type of chronic disease, findings, and limitations.

This study focuses on how ML and BDA-enabled chronic disease management systems facilitate the decisions made by doctors, patients, and policymakers in detecting, predicting, managing, and integrating into the patient-centric care paradigm across disease evolution stages. Based on the analytics need and the proposed BDA architecture, this research offers integrated perspectives on how to transform patient outcomes for chronic disease management by synergistically implementing ML and BDA. This research has crucial academic, technical, and managerial implications and opens up other future research avenues. Despite the enormous potential of machine learning and big data analytics to transform chronic disease management, only a handful of innovations have been subjected to larger-scale trials, hindering a swift translation into patient-centric chronic disease care. Many of the innovations hinge on inaccurate or ambiguous clinical concepts and few have considered the social dynamics involved in chronic disease. Concerns surrounding the responsibility of machines or algorithms for unintended negative consequences and the limited accessibility and equity of AI-based technologies further hinder the adoption of these innovations. Hence, innovations should pay special attention to conveyability and accountability that maintain a balance between complexity and interpretability and engage end-users early in the design phase through participatory design principles to foster trust in technology.