Data Driven Approaches for Healthcare

This insightful book explores machine learning and data-driven methods to tackle the high utilizers problem in healthcare, offering practical solutions and strategies.

This book presents data-driven methods, particularly machine learning, to understand and address the high utilizers problem, illustrated through a large public insurance program. It outlines key objectives for data-driven strategies while addressing various aspects of the high utilizer issue, highlighting the unique challenges that arise in this context.

Health care utilization generates extensive data from numerous sources, including electronic medical records, insurance claims, vital signs, and patient-reported outcomes. The emerging field of predicting health outcomes through data modeling offers significant insights into disproportionate spending patterns. Data Driven Approaches for Healthcare delves into these methods, emphasizing the importance of understanding high utilizers and the complexities involved in analyzing their data.

Key features of the book include an introduction to fundamental elements of health care data, especially administrative claims data, such as disease codes, procedure codes, and drug codes. It offers tailored supervised and unsupervised machine learning techniques to comprehend and predict high utilizers, alongside descriptive data-driven methods for this population. Additionally, the book identifies optimal linear and tree-based regression models that account for patients' acute and chronic conditions, as well as their demographic factors.