Big Data and Health Analytics PDF

PURPOSE

The practice of medicine and the business that encompasses it are rapidly changing. But while changes in health care delivery are widely recognized and discussed, much less is known about where health care as an industry will be 5, 10, or even 15 years from now. In the editors' lifetimes, this highly regulated market has gone through several evolutions (or as some would posit revolutions). The Accountable Care Act (ACA) has set the direction for the near future. But what happens after health care reform?

Contributing to the changes in health care delivery is the role of computing and information technology. While many writers focus on the revolution at the computerized bedside, in this book, we will be focusing on the current and future uses of the material being computed: the data and their ancillary analytics. Our authors write about how individual bits of health information can be organized into big data to improve the business of delivering services and to communicate to consumers. As an industry, we are just beginning to realize the potential that myriad information health deliveries hold, for health care both in the United States and globally.

Informatics has been dealing with data for years. What is new is the availability of large volumes of data, the degree to which these data are viewed as mission critical, and the scale of technologies required to make the data provide those critical missions. We call this state of affairs big data. In Chapter 8, Bruce Johnson points out, "The concept of big data is just that: a concept for the value an organization can realize from in-depth analysis of all data. The concept of big data is therefore not a database or data architecture but is more the solutions that leverage any and all data, wherever they come from. In health care, the concepts of big data are enabled only in organizations that focus on data—capture, management, and usage." However, the reader will find several overlapping definitions in this book.

The purpose of this book is to provide frameworks using cases and examples of how big data and analytics play a role in modern health care, including how public health information can inform health delivery. This book is written for health care professionals, including executives. It is not a technical book on statistics and machine-learning algorithms to extract knowledge out of data or a book exploring the intricacies of database design. It represents some of the current thinking of academic and industry researchers and leaders. It is written in a style that should interest anyone interested in health information and its use in improving patient outcomes and the business practices that lead to improved outcomes.

We stress usage, because without providing the right information to the people who need it, when they need it, data capture will not add value. The authors in this volume thus provide examples of how big data's management and use can improve access, reduce cost, and improve quality.

Big data and health analytics have been criticized for their unrealized potential. In some ways, the authors of these criticisms are correct. In a 2014 article that appeared in Health IT News (p. 1), Carl Shulman talks about how "fast, easy tech" matters. At this point, fast and easy electronic health information is rarely available. Data are collected, but the business plan of making it comprehensive and valid for a variety of purposes is missing. Some of the challenges for big data and health analytics today include the following:

• Incorporating new information, such as biomedical data, and new technologies into electronic health records (EHRs) that store big data. Text data require special algorithms, genetic data may be voluminous, and continuously monitored physiological data can be at arbitrary levels of granularity.
• The eventual movement to ICD-10-CM/PCS coding. While this coding provides a wealth of specific diagnostic information, the investment in data systems and associated business practices to handle complex codes is large. More generally, there is a potential loss of information between the raw data collected and the standard tagging required.
• Harnessing the potential of unstructured data for analysis, such as medical imaging and text.
• Building a culture of data sharing and the architecture, including interoperability, to meet health system needs, including future meaningful use requirements.
• Building data systems that meet requirements of accountable care organizations (ACOs) and other types of payment reforms.
• Producing understandable information for both providers and consumers.
• Maintaining patient privacy while aggregating data that increasingly can identify the individual, even without the Health Insurance Portability and Accountability Act (HIPAA) 18 safe-harbor data items.

The National Academy of Sciences talks about teaching students to extract value of big data. This imperative assumes we know what to teach them. For those of us in the health care industry who are involved in big data and health analytics, showing added value to the many different health professions is our challenge for health big data and analytics.