Big Data, Health Informatics, and the Future of Cardiovascular Medicine

JOURNAL OF THE AMERICAN COLLEGE OF CARDIOLOGY

VOL. 69, NO. 7, 2017

ª 2017 BY THE AMERICAN COLLEGE OF CARDIOLOGY FOUNDATION

ISSN 0735-1097/$36.00

PUBLISHED BY ELSEVIER

http://dx.doi.org/10.1016/j.jacc.2017.01.006

FELLOWS-IN-TRAINING & EARLY CAREER PAGE

Big Data, Health Informatics, and the Future of Cardiovascular Medicine Joonseok Kim, MD

“Medicine is going to become an information

everlasting effort to utilize up-to-date information, we

science. In 10 years or so, we may have billions of

are now facing a health informatics and big data rev-

data points on each individual, and the real

olution. This new wave has the potential to positively

challenge will be to develop information technology that can reduce that to real hypotheses about that individual.”

A

—Leroy Hood (1) physician acquires and processes data from various sources to deliver the most appropriate treatment. The lack of high-quality

data can undermine any phase of delivering care. For example, scarce information regarding the patient’s condition could lead to a wide range of differ-

ential diagnoses with high uncertainty, ultimately leading to diagnostic delay. On the other hand, high-quality, accurate, and clinically relevant information is key to making precise and informed decisions, which results in optimal patient care. Historically, health care providers solely relied

transform the way we practice medicine (3). Unfortunately, the current medical education system during medical school, residency training, and fellowship still largely focuses on traditional methodologies for collecting data and organizing gathered information (4). A transition from these methodologies to a more modern health informatics approach, including the use of “big data,” is necessary. As such, a remarkable opportunity to help lead this effort exists for fellows-in-training (FITs) and early career (EC) cardiologists.

WHAT IS “BIG DATA” IN HEALTH CARE? There is no uniform definition of “big data” in health care, but it is commonly characterized by the 5 “Vs”: volume, velocity, variety, veracity, and value (5,6).

upon the patient’s history and a meticulous physical

Volume represents the size of a dataset, usually

examination. The quality of medical care was deter-

ranging from terabytes (10 12 bytes) to zetabytes (10 21

mined by each provider’s ability to obtain and process

bytes). Velocity pertains to data in motion and the fast

the available information. In the modern medical era,

speed of the generation of new data. Variety refers to

we now have exponentially increasing amounts of

data in various types and forms, and its resultant

objective information originating from complex diag-

complexity. Veracity indicates the trustworthiness

nostic imaging tools (especially pertinent in cardiol-

and inherited ambiguities due to data uncertainty and

ogy), advanced laboratory results (including genomic

inconsistency. Value refers to the additional worth

and other “omic” data), and entirely new sources, such

that data can bring to generate knowledge (6). These

as from wearable, mobile, or other device technologies

unique characteristics of big data have limited the

(2). More recently, electronic health records (EHRs)

application of conventional statistics and epidemio-

have assisted health care providers in using the

logical approach. As a result, this new concept of big

desired data more efficiently and precisely. Facilitated

data has necessitated the development of different

by rapid advances in data science and driven by our

analytical tools (7). In the last decade, there has been an exponential growth in health data from sources such as wearable and implantable devices, smartphones, and real-time sensors (8). System capacities

From the Division of Cardiovascular Health and Disease, Department of Medicine, University of Cincinnati College of Medicine, Cincinnati, Ohio.

for data acquisition, storage, and processing have

Dr. Kim has reported that he has no relationships relevant to the contents

become more affordable (8). Furthermore, dramatic

of this paper to disclose.

advances in big data visualization and analysis

900

Kim

JACC VOL. 69, NO. 7, 2017 FEBRUARY 21, 2017:899–902

Fellows-in-Training & Early Career Page

techniques have enabled us to derive meaningful in-

HOW SHOULD WE PREPARE AS FITs

sights from big data (9). These advances have fueled

AND EC CARDIOLOGISTS?

the discussion of the role of big data in clinical practice and research (10). The proper application of these

FITs have substantial advantages in big data and health

techniques is seen as a novel approach to answering

informatics research. Training cardiologists of the

clinical questions that may have been previously

current generation tend to be information technology

impossible to address. In this regard, it is the analytic

(IT) savvy and familiar with computer systems and

method and resultant value that is the most beneficial,

emerging technologies. Trainees are generally “digital

not the big data itself (10).

natives” and have witnessed the rapid advances in

Appropriate use of big data in health care possesses

computer science, the Internet, and smartphones. This

tremendous promise and could be applied to multiple

familiarity in IT allows for a more natural understand-

stages of research in cardiology, such as very large-

ing of health informatics and EHRs. In addition, many

scale population health management, cardiovascular

fellows and young cardiologists have background

disease risk prediction, precision medicine using

knowledge in mathematics, biology, computer science,

genomic information, and clinical decision support

and bioinformatics, which are core skill sets needed to

through machine-learning algorithms (5,11). Papers

integrate big data analytics into clinical practice and

utilizing big data and new analytics in cardiology

research. Furthermore, fellows actively take care of

have been sparse, but there is a growing number of

patients on the front line, which could inspire the most

publications supporting its potential utility in car-

valuable and clinically needed research ideas that can

diovascular research. One study demonstrated that

be readily applied to clinical practice.

analysis through machine learning using a large data

There are several ways to gain exposure and learn

registry improved cardiovascular outcome prediction

more about health informatics and data science as a

in patients with suspected coronary artery disease

cardiologist, from local to international levels. Most

compared with the conventional risk scoring method

health care organizations have at least 1 medical

(12). Similarly, Loghmanpour et al. (13) examined the

informatics committee to oversee the operational and/

Bayesian network algorithm model, showing superior

or research aspects of the EHRs and associated tech-

prediction of right ventricular failure following left

nologies. Fellows can become members of these

ventricular assist device therapy over the currently

groups and participate in the various projects being

available risk prediction model. More recently,

carried out. Second, there are numerous online lec-

Circulation: Cardiovascular Quality and Outcomes set

tures providing learning opportunities in data science

its theme on big data, highlighting innovative meth-

and big data analytics at very reasonable costs or even

odological approaches and big data analytics in

no cost. As an example, the American Medical Infor-

cardiovascular outcomes research (14–16).

matics Association offers “10  10” virtual courses to

The Big Data to Knowledge initiative at the National

train health care professionals in all aspects of health

Institutes of Health reinforces the importance of this

informatics (18). Fellows in a university or program

research in health care (17). Despite the excitement and

that offers courses in data science and health infor-

optimism surrounding big data and its potential utility

matics could take advantage of these classes outside of

in cardiovascular medicine, there is a significant gap

the cardiology division (19). These lecture series will

between enthusiasm in the field and day-to-day use in

help FITs and EC cardiologists gain health informatics

cardiology practice, as well as in clinical research (11).

literacy and skills. Finally, there are more advanced

This gap is partly due to the innate challenges in uti-

formal training opportunities, including fellowships in

lizing complex, unorganized, and large-volume medi-

clinical informatics, as well as master degree or PhD

cal data. It is difficult to understand the causal

programs in health informatics and data science. There

relationships

machine-

is no uniform way to approach and learn the method-

learning technique; therefore, generalizability of the

ologies utilizing big data and health informatics in

outcome is limited (9). Moreover, although data puri-

cardiovascular research. Any of these options will

established

through

the

fication technology has improved, the process of

provide a solid opportunity to develop a firm founda-

separating signal from noise is becoming more

tion in health informatics and big data analytics.

complicated (9). In addition, cardiologists and clinical

Big data research and health informatics have not

investigators are not yet familiar with the fast-growing

yet gained popularity in cardiology, but there is

concept of big data and its application in cardiovascular

tremendous potential for their utility in cardiology

research. FITs and EC cardiologists who are prepared to

research

surmount the barriers in the use of big data and capable

medicine has been a leader in the evidence-based

of unleashing its potential in cardiology are needed.

medicine era, with advances in epidemiological

and

clinical

cardiology.

Cardiovascular

Kim

JACC VOL. 69, NO. 7, 2017 FEBRUARY 21, 2017:899–902

Fellows-in-Training & Early Career Page

cohort studies and randomized controlled trials. As

informatics. Having proficiency in big data from very

we move on to the next era of data-driven medicine,

large-scale populations and a variety of sources will

big data research and health informatics will help

be an invaluable asset as a cardiologist and cardio-

open the door to new insights in cardiology and

vascular disease investigator. Cardiology fellows

transform our medical practice (20).

should be perceptive and knowledgeable in big data

Big data in cardiovascular practice and research is

and health informatics concepts, and be prepared to

at its nascent stage. Moving forward, translating big

collaborate to achieve the ultimate goal: prevent

data for use in clinical practice will take a great effort

cardiovascular disease and improve cardiovascular

among many collaborators, not just solitary in-

outcomes.

vestigators. This will be a tremendously multifaceted

ACKNOWLEDGMENT The

endeavor

valuable contribution of Dr. Eric Kirkendall for his

that

requires

high-level

expertise

in

different specialties. It is clear that data scientists,

author

appreciates

the

insights, manuscript review, and comments.

computer scientists, statisticians, health informatics experts, and cardiologists need to practice team sci-

ADDRESS FOR CORRESPONDENCE: Dr. Joonseok

ence in this endeavor. In this context, there is a high

Kim, Medical Science Building, University of Cincinnati,

demand for cardiologists with the skills and knowl-

231 Albert Sabin Way, MLC 0542, Cincinnati, Ohio 45267.

edge to collaborate with other specialists in health

E-mail: [email protected].

REFERENCES 1. Hood L. A vision for personalized medicine. MIT Technology Review. Available at: https:// www.technologyreview.com/s/417929/a-vision-forpersonalized-medicine/. Accessed May 10, 2016. 2. Andreu-Perez J, Poon CC, Merrifield RD, Wong ST, Yang GZ. Big data for health. IEEE J Biomed Health Inform 2015;19:1193–208. 3. Krumholz HM. Big data and new knowledge in medicine: the thinking, training, and tools needed for a learning health system. Health Aff (Millwood) 2014;33:1163–70. 4. Pageler NM, Friedman CP, Longhurst CA. Refocusing medical education in the EMR era. JAMA 2013;310:2249–50. 5. Antman EM, Benjamin EJ, Harrington RA, et al. Acquisition, analysis, and sharing of data in 2015 and beyond: a survey of the landscape: a conference report from the American Heart Association Data Summit 2015. J Am Heart Assoc 2015;4:e002810. 6. Wang L, Alexander CA. Big data in medical applications and health care. Am Med J 2015;6:1–8.

8. Scruggs SB, Watson K, Su AI, et al. Harnessing the heart of big data. Circ Res 2015;116:1115–9.

quantity, and data density. Circ Cardiovasc Qual Outcomes 2016;9:649–58.

9. Mayer-Schönberger V. Big data for cardiology: novel discovery. Eur Heart J 2016;37:996–1001.

15. Carson MB, Scholtens DM, Frailey CN, et al. Characterizing teamwork in cardiovascular care outcomes: a network analytics approach. Circ Cardiovasc Qual Outcomes 2016;9:670–8.

10. Krumholz HM. The promise of big data: opportunities and challenges. Circ Cardiovasc Qual Outcomes 2016;9:616–7. 11. Rumsfeld JS, Joynt KE, Maddox TM. Big data analytics to improve cardiovascular care: promise and challenges. Nat Rev Cardiol 2016;13:350–9. 12. Motwani M, Dey D, Berman DS, et al. Machine learning for prediction of all-cause mortality in patients with suspected coronary artery disease: a 5-year multicentre prospective registry analysis. Eur Heart J 2016 Jun 1 [E-pub ahead of print]. 13. Loghmanpour NA, Kormos RL, Kanwar MK, et al. A Bayesian model to predict right ventricular failure following left ventricular assist device therapy. J Am Coll Cardiol HF 2016;4:711–21.

16. Spertus JV, Normand S-LT, Wolf R, Cioffi M, Lovett A, Rose S. Assessing hospital performance after percutaneous coronary intervention using big data. Circ Cardiovasc Qual Outcomes 2016;9:659–69. 17. Bourne PE, Bonazzi V, Dunn M, et al. The NIH Big Data to Knowledge (BD2K) initiative. J Am Med Inform Assoc 2015;22:1114. 18. American Medical Informatics Association. AMIA 1010 courses. Available at: https://www. amia.org/education/10x10-courses. Accessed May 10, 2016. 19. Stanford Medicine. Introduction to the Biomedical Informatics Training Program. Available at: http:// bmi.stanford.edu/prospective-students/. Accessed

14. Ng K, Steinhubl SR, deFilippi C, Dey S, Stewart WF. Early detection of heart failure using

May 10, 2016.

7. Raghupathi W, Raghupathi V. Big data analytics in healthcare: promise and potential. Health Inf Sci Syst 2014;2:3.

electronic health records: practical implications for time before diagnosis, data diversity, data

data-driven medicine: translational bioinformatics’ next frontier. J Am Med Inform Assoc 2012;19:e2–4.

20. Shah NH, Tenenbaum JD. The coming age of

RESPONSE: Moving Beyond Big Data to Causal Inference and Clinical Implementation Peter W. Groeneveld, MD, MS Cardiovascular Outcomes, Quality, and Evaluative Research Center, Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania; and the Michael J. Crescenz Veterans Affairs Medical Center, Philadelphia, Pennsylvania E-mail: [email protected] Medicine is in the midst of a digital revolution, and the

being thoroughly transformed by both the unprecedented

practice of cardiovascular medicine and clinical research is

expansion in the volume and variety of patient data,

901

902

Kim

JACC VOL. 69, NO. 7, 2017 FEBRUARY 21, 2017:899–902

Fellows-in-Training & Early Career Page

as well as the development of advanced information

clinical actions might change the probability of those

technologies than can organize and analyze these data,

future events.

providing precise, actionable information to clinicians at

Hence, although it is arguably important that all

the point of care. In his excellent essay, Dr. Kim rightly

physicians-in-training have a basic understanding of the

emphasizes that cardiology fellows-in-training would be

diversity of biomedical data and the innovative ways that

wise to prepare themselves for practicing medicine in the

it can be analyzed to effect clinical care—lest they be sur-

“Big Data” era, and he highlights opportunities for

prised when their smartphones start recommending

informatics-savvy physicians to influence the evolution of

treatments and tests for their patients—it is absolutely

21st-century cardiovascular care.

vital that physicians be scientifically literate, and thus,

Although this enthusiasm is justified, it is also essential

suitably discerning in assessments of the clinical value of

not to lose sight of one of the great achievements of

these Big Data applications. Improving outcomes is clinical

20th-century cardiology—namely, the establishment of

medicine’s primary goal; better predictions are subordi-

scientific evidence as the cornerstone of effective clinical

nate. It is also critical for physicians pursuing advanced

practice. One unfortunate aspect of the application of

informatics training to recognize that neither Big Data nor

“data science” to medicine is that some data scientists—

advanced analytics can unerringly identify causal mecha-

although truly ingenious in their mastery of data taxon-

nisms, and transforming predictive information into

omies, information technology, computing architecture,

effective clinical actions that improve patient outcomes is

and advanced analytics—lack a firm grounding in, and are

equally important as, and often much harder than, pro-

occasionally dismissive of, scientific methodology (1,2). At

ducing precise predictions from petabytes of data.

its core, biomedical science entails hypothesis testing with

A comprehensive understanding of the field of

the goal of understanding causal relationships, while

biomedical informatics encompasses not only the har-

recognizing that chance, bias, and confounding inevitably

nessing of vast and varied data sources and applying

threaten the validity of causal inference. The scientific

complex and adaptive algorithms to generate accurate

approach may be unnecessary in many fields where data

predictions and actionable information, but also the

science has flourished (e.g., marketing, consumer analytics,

finance,

insurance,

meteorology,

humble recognition that these elements are not sufficient

professional

to improve patient outcomes. A robust definition of infor-

sports, and so on) (2). But, it remains essential in medi-

matics includes not only the analysis of data, but also the

cine, because physicians not only need accurate pre-

appropriate use of information in highly complex clinical

dictions of future adverse events that might befall our

settings where human decision makers (3)—clinicians

patients, but also need a clear understanding of how our

and patients—will continue to play vital roles.

REFERENCES 1. Schmitt C, Cox S, Fecho K, et al. Scientific discovery in the era of big data: more than the scientific method. RENCI White Paper 2015;6(3). Available at: http://renci.org/wp-content/uploads/2015/11/SCiDiscovery-BigData-FINAL-11.23.15.pdf. Accessed January 12, 2017.

2. Anderson C. The end of theory: the data deluge makes the scientific method obsolete. Wired Magazine. 2008 Available at: https://www.wired.com/ 2008/06/pb-theory/. Accessed January 12, 2017.

3. Wachter RM. The Digital Doctor: Hope, Hype, and Harm at the Dawn of Medicine’s Computer Age. New York: McGraw-Hill Education, 2015.