Handbook of Healthcare Analytics

18 Data Science

Rema Padman

Carnegie Mellon University

18.1 Introduction

Data science is conceptualized currently as a mashup of many disciplines to extract value from data (Saltz and Stanton, 2018; Cao, 2017; Fawcett, 2015; Kayyali et al., 2013; Kruse et al., 2016; Russell et al., 2010). Enabled by the explosive advances in information, communication, and decision technologies in recent decades, digitization is dramatically transforming every industry sector in economies worldwide, from manufacturing and retail to transportation and hospitality services, resulting in the creation of large volumes of data (L'Heureux et al., 2017). In no industry is this more disruptive or potentially transformative than the healthcare sector for many reasons, including the diversity of users and usage, technologies, data sources and types, heterogeneity of content, lack of interoperability among the software systems that collect and manage data, and insufficient standards across the data generation, collection, analysis, and usage spectrum (Raghupathi et al., 2014).

Healthcare today is so complex that it has surpassed the human mind's capacity to operate without aid (Figure 18.1). The greatest challenge is to take the appropriate information and apply it to individual patients or a population at the time when the information is needed. Drawing on multiple disciplines, such as statistics, machine learning, operations research, information technologies, computer science, and economics, ...