Innovation at ACC | Harnessing the Power of Big Data in Cardiovascular Disease

The cardiovascular research community has seen an emergence of small groups of investigators publishing papers with immense sample sizes, sometimes numbering in the thousands or even millions of patients. Much of this is due to the increasing availability of large-scale databases accessible to independent researchers. It is fitting and unsurprising that these datasets have become such popular sources for investigation given that healthcare data is increasing at an exponential rate, previously noted at approximately zettabyte levels (1021 bytes of data points).1,2

Calling All Innovators

ACC’s new Health Care Innovation Member Section launched last month at ACC.18 in Orlando. The section has four initial aims:

  • To become the professional home for cardiovascular specialists and other multidisciplinary team members interested in the field of health care innovation.
  • To serve as a resource for information and action regarding the latest in health care innovation –particularly in the specific areas of Digital Health, Health Care Analytics, Emerging Technologies, Precision Medicine and Entrepreneurship.
  • To promote activities that will range from education and career development for professionals involved or interested in health care innovation to assessment and guidance to the ACC regarding the latest advancements and developments in the field.
  • To promote value to the ACC and its members by advancing the ACC’s mission and its Strategic Plan through the expertise of the Section’s leadership and membership.

These datasets include information on demographics, clinical outcomes, health care system performance, costs, resource utilization, electronic health records, genomic variation and biomarkers. Given their sheer size, they provide unique opportunities to investigate rare diseases or treatment options that individual clinicians may not see often within their local community. The datasets provide an opportunity to look broadly, beyond regional and national boundaries. Big data also provide a real-world perspective on post-market surveillance of therapeutics as well as trends in health care. Access to genomic and metabolomic data linked to detailed phenotypes provides an unprecedented opportunity for translational science.

Some of the most popular datasets in the cardiology community include ACC’s NCDR registries; the Agency for Healthcare Research and Quality’s Healthcare Cost and Utilization Project (HCUP) databases; the Society of Thoracic Surgeons’ National Database; the American Heart Association’s Get With The Guidelines® databases; and the European Society of Cardiology’s EuroObservational Research Programme registries.

Each dataset has its own rules, regulations and oversight committees. All require formal applications, though they vary in the extent of detail required. For example, among the HCUP databases, the National Inpatient Sample, which represents approximately 50 percent of the U.S. population and 50 percent of U.S. hospitalizations, requires potential investigators to complete data usage agreements, provide credentialing information and payment for the database. After approval by HCUP, investigators can download the entire database from the HCUP central distributor website.

Other datasets require upfront discussion about the investigative strategy and intent for using the data. For example, ACC’s ACTION Registry, considered to be among the most trusted data sources for outcomes-based quality improvement on high-risk STEMI/NSTEMI patients, requires potential investigators to submit a research protocol outlining their hypothesis and analytical plan. A committee then reviews the application at predetermined times during the calendar year; if approved the investigators are provided with specific statistical data output and data elements as requested in the application.

The availability of public genomic data linked to phenotypes, sometimes in prospective cohorts, is also exploding. There are several biobanks which collectively provide genomic data linked to clinical phenotypes on millions of patients. These cohorts are enriched with cardiovascular disease or risk factors. The largest are funded by governments, such as the U.K. Biobank, which started enrolling in 2006 and already has data on >500,000 patients.3 Its equivalent in the U.S. is the Million Veterans Program, which started enrolling in 2001 and now has about 500,000 of the planned one million patients enrolled.4 The most promising is the All of Us Research Program, part of the National Institutes of Health’s Precision Medicine Initiative. It launched only last year and is promising an enrollment of one million people.

Other biobanks include those that are funded commercially, such as deCODE Genetics (Amgen) and Geisinger MyCode Community Health (Regeneron Pharmaceuticals and others), or by institutions such as Kaiser Permanente Research Bank or Partners Healthcare Biobank.

While the general ethos of biobanks is to make data public to enhance scientific research, access requirements to data vary depending on the biobank and the funding source. The majority require a scientific proposal and an approval process.

Additionally, academic and private electronic cohorts targeting people directly are becoming more popular. Health eHeart and Evidation are two examples. With the expansion of the use of wearable and sensor technology, we will continue to see exponential growth in high-fidelity data, such as heart rate, actigraphy, accelerometry and even electrocardiography.

ACC’s newly established Innovation Section prioritizes the precision health agenda within the College. One area of interest is promoting the availability and use of big data to guide evidence-driven, patient-centered care models and translational research.5 Big data applications within the field of cardiology have great potential in the settings of implementation in drug and device monitoring, clinical guideline development, waste reduction and inevitably improving health care outcomes.

We should continue to accelerate the development of these large databases and their ability to interlink, as well as continue to improve access for investigators. Fortunately, multiple efforts are underway in the movement towards providing more open access to publicly and privately funded clinical trial databases,6,7 as well as sharing genetic, radiographic and outcomes data from biomedical repositories.3 Together with the use of novel artificial intelligence algorithms, big data holds promise for more innovation in cardiovascular health.


  1. Rumsfeld JS, Joynt KE, Maddox TM. Nat Rev Cardiol 2016;13:350-9.
  2. Raghupathi W, Raghupathi V. Health Inf Sci Syst 2014;2:3.
  3. Sudlow C, Gallacher J, Allen N, et al. PLoS Med 2015;12:e1001779.
  4. Gaziano JM, Concato J, Brophy M, et al. J Clin Epidemiol 2016;70:214-23.
  5. Bhavnani SP, Parakh K, Atreja A, et al. J Am Coll Cardiol 2017;70:2696-2718.
  6. Krumholz HM, Waldstreicher J. N Engl J Med 2016;375:403-5.
  7. Taichman DB, Backus J, Baethge C, et al. N Engl J Med 2016;374:384-6.

Clinical Topics: Arrhythmias and Clinical EP, Genetic Arrhythmic Conditions

Keywords: ACC Publications, Cardiology Magazine, American Heart Association, Myocardial Infarction, Translational Medical Research, Quality Improvement, Public Health, Electrocardiography, Information Storage and Retrieval, Electronic Health Records, Genomics, Patient-Centered Care, Demography, Biomarkers, Phenotype, Patient Care Team, Health Services Research, Health Care Costs, Risk Factors, Quality Improvement, Registries

< Back to Listings