REVIEW

published: 01 April 2022

doi: 10.3389/fcvm.2022.844296

Big Data in Cardiology: State-of-Art

and Future Prospects
Haijiang Dai 1,2 , Arwa Younis 3 , Jude Dzevela Kong 2 , Luca Puce 4 , Georges Jabbour 5 ,
Hong Yuan 1* and Nicola Luigi Bragazzi 2,4,6,7*
1
Department of Cardiology, The Third Xiangya Hospital, Central South University, Changsha, China, 2 Laboratory for Industrial
and Applied Mathematics (LIAM), Department of Mathematics and Statistics, York University, Toronto, ON, Canada, 3 Clinical
Cardiovascular Research Center, University of Rochester Medical Center, Rochester, New York, NY, United States,
4
Department of Neuroscience, Rehabilitation, Ophthalmology, Genetics, Maternal and Child Health (DINOGMI), University of
Genoa, Genoa, Italy, 5 Physical Education Department, College of Education, Qatar University, Doha, Qatar, 6 Postgraduate
School of Public Health, Department of Health Sciences, University of Genoa, Genoa, Italy, 7 Section of Musculoskeletal
Disease, Leeds Institute of Molecular Medicine, NIHR Leeds Musculoskeletal Biomedical Research Unit, University of Leeds,
Chapel Allerton Hospital, Leeds, United Kingdom

Cardiological disorders contribute to a significant portion of the global burden of

disease. Cardiology can benefit from Big Data, which are generated and released
by different sources and channels, like epidemiological surveys, national registries,
electronic clinical records, claims-based databases (epidemiological Big Data), wet-
Edited by: lab, and next-generation sequencing (molecular Big Data), smartphones, smartwatches,
Victor Chien-Chia Wu,
Chang Gung Memorial and other mobile devices, sensors and wearable technologies, imaging techniques
Hospital, Taiwan (computational Big Data), non-conventional data streams such as social networks, and
Reviewed by: web queries (digital Big Data), among others. Big Data is increasingly having a more
Gustavo B.F. Oliveira,
and more relevant role, being highly ubiquitous and pervasive in contemporary society
Instituto Dante Pazzanese de
Cardiologia, Brazil and paving the way for new, unprecedented perspectives in biomedicine, including
Giuseppe Sangiorgi, cardiology. Big Data can be a real paradigm shift that revolutionizes cardiological
University of Rome Tor Vergata, Italy
Andre Rodrigues Duraes, practice and clinical research. However, some methodological issues should be properly
Federal University of Bahia, Brazil addressed (like recording and association biases) and some ethical issues should be
*Correspondence: considered (such as privacy). Therefore, further research in the field is warranted.
Nicola Luigi Bragazzi
bragazzi@yorku.ca Keywords: Big Data, epidemiological registries, high-throughput technologies, wearable technologies,
Hong Yuan non-conventional data streams, cardiology
yuanhong01@csu.edu.cn

Specialty section: CARDIOVASCULAR DISEASE: EPIDEMIOLOGY AND GLOBAL

This article was submitted to BURDEN OF DISEASE
General Cardiovascular Medicine,
a section of the journal
The global burden of disease (GBD) is the quantitative estimation of the health loss
Frontiers in Cardiovascular Medicine
because of a disorder, risk factor, or injury. It is modeled and computed as the
Received: 27 December 2021 epidemiological, clinical, and societal burden generated by a given disease, in terms of
Accepted: 24 February 2022
its economic-financial and humanistic impact, if ineffectively managed and inadequately
Published: 01 April 2022
treated. Such a quantitative approach enables practitioners and scholars as well as all relevant
Citation: stakeholders, including public and global health decisions- and policymakers, to compare
Dai H, Younis A, Kong JD, Puce L,
the burden of different diseases, risk factors, or injuries, robustly and consistently over
Jabbour G, Yuan H and Bragazzi NL
(2022) Big Data in Cardiology:
a temporal period and across various spatial settings and territories/nations. Moreover,
State-of-Art and Future Prospects. these data can inform policies in a pure data-driven and evidence-based fashion, allowing
Front. Cardiovasc. Med. 9:844296. prioritization and allocation of resources, especially in developing countries and in other
doi: 10.3389/fcvm.2022.844296 resource-limited contexts (1). This approach enables to monitor the effects of a given

Frontiers in Cardiovascular Medicine | www.frontiersin.org 1 April 2022 | Volume 9 | Article 844296

Dai et al. Big Data in Cardiology

policy or intervention and verifies if sufficient progress has been pioneering and inspiring figure in the arena of systems biology,
made toward the achievement of the Sustainable Development pointing out the shift from a “one-size-fits-all” theoretical
Goals (SDGs) set up by the United Nations (UN) General framework to one in which the individual signature of the disease
Assembly (2). In particular, SDG 3.4.1 has the ambitious goal of matters (5–9).
achieving a 30% reduction in premature mortality due to non- Moreover, thanks to its latest scientific and technological
communicable diseases, including cardiovascular disease (CVD), improvements, medicine, including cardiology, is entering a new,
by 2030 (2). unprecedented era, characterized by the production and release
To track such a target, the GBD initiative as well as other of an incredible amount of data, termed as Big Data. They are
similar taskforces and groups, like the Global Health Estimates characterized by several key dimensions, including velocity (Big
(GHE) initiative led by the World Health Organization (WHO), Data can be generated, processed, and analyzed in real-time),
have devised and implemented a set of validated and reliable volume (referring to the wealth of data, the magnitude of which
indicators. These measures include the years of life lost (YLLs), challenges classical storage, processing, and analytical capacities
the years lived with disability (YLDs), and the disability-adjusted and infrastructures), variety (referring to the diversity of data
life years (DALYs), which allow researchers to quantitatively sources, administrative, patient-reported, healthcare-generated,
evaluate life lost due to death (casualty or premature death) or etc.), veracity (credibility, reliability, and accuracy of data), and
disability, respectively, which hinder to live life at 100% health. value (raw data that, once processed, become smart, applicable,
As previously mentioned, GBD- and GHE-related metrics and actionable).
are of paramount importance in providing stakeholders with Different channels and sources can produce Big Data: from
data, especially in those settings where there is a dearth of data, large-scale surveys, databases, repositories, and registries
or data are not properly updated and/or reliable, because they (epidemiological/clinical Big Data) to next-generation
would be too much time- and resource-consuming to collect. sequencing and high-throughput technologies (molecular
CVD contributes to a significant portion of the GBD (3). CVD, Big Data) and computational approaches (infodemiological
especially stroke and ischemic heart disease (IHD), is the leading or digital Big Data). Big Data is deeply transforming clinical
cause of mortality and disability. Prevalent CVD cases have practices into disruptive ones and informing data-driven
nearly doubled from 271 million in 1990 to 523 million in 2019, approaches (Figures 1, 2).
globally. Similarly, the number of CVD deaths has increased from The “American College of Cardiology (ACC) Task Force on
12.1 million to 18.6 million, with DALYs and YLLs increasing Health Policy Statements and Systems of Care” designed the 2017
as well. YLDs have doubled from 17.7 million to 34.4 million. Roadmap for Innovation in the cardiological arena, identifying
Despite scholarly achievements and technological advancements, three major pillars: namely, i) digital health, ii) Big Data, and iii)
especially concerning the management of acute coronary artery precision health (10, 11).
disease, chronic ischaemic heart disease, and heart failure, CVD
still imposes a dramatically high burden, which is increasing even
in those settings in which it was previously decreasing (3, 4),
ROLES AND APPLICATIONS OF
pointing out the urgent need of implementing effective public EPIDEMIOLOGICAL/CLINICAL BIG DATA
health policies at a global and local level. This burden is still IN CURRENT CARDIOLOGICAL
dramatically high for diseases, like atrial fibrillation, acute heart RESEARCH AND PRACTICE
failure, or sudden cardiac death (3, 4).
In the present review paper, we will show how cardiology Epidemiological/clinical Big Data can come from large-scale,
can benefit from the use of the so-called “Big Data”, especially often nationwide surveys. These data can inform public and
in the efforts of counteracting and mitigating against the global health policies as well as evidence-based medicine and,
burden of CVD. In the next paragraphs, we will overview the more specifically, cardiology.
changes cardiological research and practice have undergone in Whilst randomized controlled clinical trials represent the
the last decades and we will make some examples of potential gold standard for building a body of rigorous and clinically
applications of Big Data in the cardiological arena, broken down relevant evidence, they may not always reflect real-life patient
according to their sources/channels (Tables 1–3), as well as their populations, as such limiting the generalizability and external
current major shortcomings and limitations (Table 4). validity of their findings. Real-life or real-world evidence,
collected during daily clinical practice, provides a complementary
perspective to rigorously and strictly randomized controlled
TOWARD A NEW WAY OF PRACTICING clinical trials (12, 13). In this respect, Big Data-based studies
CARDIOLOGY AND DOING can add to well-designed “small data”-based investigations and
CARDIOLOGICAL RESEARCH randomized controlled clinical trials (13).
A major example of real-life or real-world data is TriNetX,
Healthcare provision delivery has changed dramatically in the which is the largest global research network providing real-world
last decades. New models and pathways of managing and evidence. It contains tens of billions of clinical facts diagnosis,
treating diseases have emerged. A new biomedical approach laboratory findings, treatment received, procedures performed,
termed “P4 medicine” (preventative, predictive, personalized, on more than 250 million patients worldwide, including subjects
and participatory) has been introduced by Doctor Leroy Hood, a suffering from hypertensive disease, type 2 diabetes, or chronic

Frontiers in Cardiovascular Medicine | www.frontiersin.org 2 April 2022 | Volume 9 | Article 844296

Dai et al. Big Data in Cardiology

TABLE 1 | Types of big data and their sources/channels in the field of cardiology.

Type of big data Sources/channels

Epidemiological/clinical big data Epidemiological survey

Claims-based database (administrative database)
Electronic health records (EHRs)/electronic medical records (EMRs)
Large clinical registries (the “Society of Thoracic Surgeons (STS) National Database,” the “American Heart
Association (AHA) Get With The Guidelines (GWTG) Database,” the “American College of Cardiology (ACC)
National Cardiovascular Data Registry” (NCDR), the “Hospital Compare Database,” the “National Heart Lung and
Blood Institute (NHLBI) Percutaneous Transluminal Coronary Angioplasty (PTCA) Registry,” the “STS/ACC
Transcatheter Valve Therapy (TVT) database,” the “Hypertrophic Cardiomyopathy Registry,” and the “Cooperative
Cardiovascular Project”)
Molecular big data Microarray chips, next-generation DNA and RNA sequencing and whole-exome sequencing,
chromatin-immunoprecipitation-coupled sequencing, and mass-spectrometry-based proteomics analysis
Big data generated by information and Smartphones, apps, and gamified mobile apps Smartwatches Sensors and wearable devices/technologies
communication technologies (ICTs) Imaging techniques (i.e., radiography, radiomics, and radiogenomics)
Computational/digital big data “Non-conventional data streams”
Web searches (Google Trends)
Website page consultation (i.e., Wikipedia)

TABLE 2 | Some select examples of big data-based registries/databases for cardiovascular disease.

Country/territory Database name/acronym Major details

Japan Japanese Registry Of All cardiac and vascular Governed by the Japanese Circulation Society (JCS), more
Diseases-Diagnostic Procedure Combination than 700,000 health records’ data as of 2012 from 610
(JROAD-DPC) certificated hospitals
Japan Acute Myocardial Infarction Registry (JAMIR) >20,000 patients
Korea Prospective Cohort Registry for Heart Failure in Korea >5,000 patients
(KorAHF)
Denmark Danish Cardiac Rehabilitation Database (DHRD) Collecting data from all hospitals in Denmark
Danish Heart Registry Collecting data from five cardiology centers, eight cardiology
satellite centers, four surgical centers, and a private hospital
Sweden Swedish Primary Care Cardiovascular Database >70,000 patients
(SPCCD)
SWEDEHEART >2 million subjects
USA National Cardiovascular Data Registry (NCDR) Governed by the American College of Cardiology (ACC), it
consists of 10 registries, eight inpatient/procedure-based and
two outpatient-based from more than 2,400 hospitals and
8,500 providers with more than 60 million patient records

kidney disease. Specifically, concerning the cardiological arena, registration data, cohort data, inpatient and outpatient data,
this network has been exploited to shed light on the safety profile among others (20).
and cardiovascular outcomes of drugs (14, 15), the effectiveness These data can be retrospectively or prospectively collected:
of rehabilitation protocols (16, 17), and the cardiovascular prospective clinical registries can be defined as large/very
implications of the still ongoing “Coronavirus Disease 2019” large datasets of observational data which have been collected
(COVID-19) pandemic (18, 19), among others. prospectively and systematically and in a structured fashion,
To paraphrase what Doctor Lukas Kappenberger, pioneering to reflect real-world clinical practices and outcomes of a given
father of the so-called “computational cardiology,” has stated procedure (treatment, or surgical intervention) across large
in 2005, the science (i.e., randomized controlled clinical trials) patient populations, including specific clinical/demographic
tells scholars and practitioners what they can do, the guidelines (sub-)populations (20).
and checklists implement what they should do, and clinical Furthermore, besides being complementary, randomized
registries/databases tell them what they are doing and observing controlled clinical trials can be embedded within clinical
(20, 21). registries (20): this enables to save time and resources and
Currently, there are lots of sources generating epidemiological strengthens the generalizability of the findings (20). For instance,
Big Data, such as surveys, medical insurance data, vital the “Coronary Artery Surgery Study (CASS) registry,” which is

Frontiers in Cardiovascular Medicine | www.frontiersin.org 3 April 2022 | Volume 9 | Article 844296

Dai et al. Big Data in Cardiology

TABLE 3 | Types of big data and examples of potential uses/applications in the scientific societies, like the “Society of Thoracic Surgeons (STS)
field of cardiology. National Database,” which collects clinical outcomes for patients
Type of big data Examples of potential
undergoing cardiothoracic surgery (23), and the “American
uses/applications Heart Association (AHA) Get With The Guidelines (GWTG)
Database,” which is based on a hospital-based initiative, led
Epidemiological/clinical big data Epidemiological assessment (incidence, by the AHA and the American Stroke Association (ASA),
prevalence, co-morbidities, and mortality collecting data from >2,000 hospitals, aimed at improving
rates)
the quality of care of patients suffering from CVD, including
Epidemiological nowcasting/forecasting
for funding and resources allocation
heart failure, atrial fibrillation, and stroke (24). Another major
optimization societal database is the “ACC National Cardiovascular Data
Economic assessment (costs evaluation) Registry” (NCDR), which is composed of 10 registries (eight
Evaluation and comparison of different of which are inpatient/procedure-based and the remaining two
cardiological treatment and management are outpatient-based), collecting data from >2,400 hospitals and
options 8,500 healthcare providers with >60 million patient records (25).
Identification of diagnostic and prognostic Other databases include the “Hospital Compare Database”,
markers which collects data concerning the quality of care (overall star
Evaluation and assessment of mid-term rating and other quality measures) from >4,000 Medicare-
and long-term clinical outcomes
certified hospitals (26), and the “Cooperative Cardiovascular
Molecular big data Patient profiling and stratification
Project” (27, 28), which is one of the early examples of a
Personalized/individualized cardiology
clinical registry.
Characterization of the effects and actions Some datasets and registries are devoted to specific
of drugs at the cellular and molecular levels
cardiovascular medications or surgical procedures, like
Identification of potential druggable targets
the “National Heart Lung and Blood Institute (NHLBI)
Big data generated by information Collection of patient-reported outcomes
and communication technologies
Percutaneous Transluminal Coronary Angioplasty (PTCA)
Registry”, collecting outcomes data for patients undergoing
Customization and personalization of
healthcare provision delivery PTCA (29, 30), the “STS/ACC Transcatheter Valve Therapy
Computational/digital big data Patient health-related literacy assessment (TVT) database”, which collects outcomes data for patients
Patient education and empowerment undergoing transcatheter valve replacement and repair
procedures from >650 reporting sites (31), and the CathPCI
registry from the NCDR, collecting outcomes data for patients
TABLE 4 | Major shortcomings and limitations of big data in current cardiological undergoing diagnostic catheterization and/or percutaneous
practice and clinical research. coronary intervention (PCI) procedures (32).
Some registries and databases are specifically devoted
Type of big data Limitation/issue
to particular CVD, like the “Hypertrophic Cardiomyopathy
Epidemiological/clinical big data Discrepancies between registry-based Registry” led by the University of Virginia (USA) and the
studies and individual (single- or University of Oxford (UK), aimed at identifying biomarkers of
multi-center) investigations hypertrophic cardiomyopathy (33).
Discrepancies among database-based Epidemiological/clinical big data can be utilized for a
studies variety of purposes and aims, including i) performing an
Privacy and bioethical issues epidemiological assessment of CVD (in terms of incidence,
Molecular big data Conflicting results among studies prevalence, and mortality rates), ii) quantifying and forecasting
(depending on the type of tissue studied,
epidemiological trends, iii) investigating the determinants of
the type of molecular technique used, etc.)
CVD and related underlying co-morbidities, iv) identifying
“False discovery” of markers
diagnostic and prognostic markers and signatures, v) devising
Big data generated by information Privacy and bioethical issues due to the
and communication technologies pervasive and ubiquitous nature of the
risk score tools to better stratify CVD patients, vi) exploring
devices the mid-term and long-term clinical outcomes of a given
Computational/digital big data Lack of transparency concerning the (pharmacological or surgical) procedure and its superiority over
algorithm another one (the competitor), vii) verifying the implementation
of recommendations and decision-making processes, setting
benchmarks, and viii) computing the economic-financial costs of
a given CVD (34, 35).
also one of the early examples of clinical registries, is a database Big data can help uncover relationships between diseases
embedded within a clinical trial (the CASS investigation) (22). and/or co-morbidities, in that they tend to co-cluster. The
In the cardiological arena, there exist very large clinical diseasome is the “human disease network”: a Big Data-based
databases and registries, whose origins can be dated back to the study of the diseasome can contribute to a better understanding
eighties (20). The most popular ones include societal registries, of the so-called “system or network medicine” (36). Several
that is to say, databases endorsed, funded, and sponsored by CVDs, including heart failure, frequently coexist with various

Frontiers in Cardiovascular Medicine | www.frontiersin.org 4 April 2022 | Volume 9 | Article 844296

Dai et al. Big Data in Cardiology

FIGURE 1 | Types and sources/channels of big data.

FIGURE 2 | Potential applications of big data in the field of cardiology.

comorbidities. Meireles et al. (37) assessed the prognostic role developing acute heart failure. A set of 229 patients suffering from
and impact of several underlying comorbidities on the risk of acute heart failure was compared vs. a set of 201 patients with

Frontiers in Cardiovascular Medicine | www.frontiersin.org 5 April 2022 | Volume 9 | Article 844296

Dai et al. Big Data in Cardiology

chronic heart failure. The number of comorbidities was slightly risk factors like high blood pressure, chest pain, diabetes, and
higher in the acute heart failure patient group: these included metabolic impairment.
metabolic impairments such as hyperuricemia, and obesity, other On the other hand, despite the use of sophisticated statistical
CVDs like atrial fibrillation, or peripheral artery disease as well tools, as previously mentioned, there are still open issues that
as chronic kidney disease. Investigating the comorbidome could need to be addressed and solved. Big Data-based studies can
allow the implementation of “precision cardiology” by devising offer a different point of view, but some conflicting findings of
ad hoc multi-dimensional interventions targeting the specific randomized controlled clinical studies and small, well-conducted
patient sub-population. investigations can be found.
There exist several risk tools, ranging from the Framingham Such discrepancies could be due to the unique nature of
score to the SCORE, the “Global Registry of Acute Coronary the database used in the study: each cardiological database
Events” (GRACE), the “Thrombolysis In Myocardial Infarction” significantly varies in the methods deployed to collect and
(TIMI), the “Congestive Heart Failure, Hypertension, Age (≥75 capture data and the population(s) it specifically represents (13).
years), Diabetes, Stroke, Vascular disease, Age (65 to 74 years), Also, the format of the database (structured vs. unstructured)
and Sex category” (CHA2 DS2 VASc), and the “Meta-Analysis could impact data quality. For instance, Hernandez-Boussard
Global Group in Chronic Heart Failure” (MAGGIC) risk-score, et al. (13) mined a dataset inclusive of 10,840 clinical
among others (38). notes and found lower recall and precision rates (51.7 and
These risk calculators are fundamental components of 98.3%, respectively) in the case of structured electronic health
the so-called “personalized cardiology,” in that they enable records (HER), concerning unstructured EHR (95.5 and 95.3%,
to stratify patient cohorts and provide the patient with respectively), warranting the routine measurement of recall for
the treatment they need the most. Examples of precision each database/registry, before proceeding with data processing
and personalized management include the customized and analysis.
assessment of the risk factor for a variety of cardiovascular Summarizing, Big Data repositories, registries, and databases
diseases, such as atrial fibrillation, chronic myocardial are increasingly common in the field of cardiological practice
ischemia, heart failure, and hypertension, given the and clinical research: there are, however, significant considerable
individual biological makeup (genetic) and family history variations in socio-demographic characteristics, co-morbidities,
for cardiovascular disease. Also, pharmacological provisions, and major complication rates between individual (single-
for instance, the usage of anticoagulants, can be tailored, or multi-center) and database-based studies, and even
in such a way to minimize the insurgence of potential among registry-studies themselves (for example, clinical vs.
side-effects (6–8). administrative database). This should be accounted for when
There are, however, few published comparisons among the critically appraising cardiological research and in risk adjustment
different risk scores, which remains a field open to further modeling (20).
research and investigation (38). In particular, administrative databases (20) can provide
Some emerging applications of Big Data-based databases researchers and scholars, as well as practitioners and policy- and
are: i) addressing cardiovascular-related iniquities and decision-makers with a lot of information concerning disease
disparities, also from a gender perspective, ii) performing epidemiology, co-morbidities, disparities, and inequalities in
post-marketing analysis of different cardiovascular treatments access to healthcare and clinical outcomes. Furthermore,
and medications (39). they can inform in a data-driven fashion the decision-
Finally, Big Data is particularly helpful when the studied making processes underlying cardiological pharmacological
cardiological disease is rare, such as congenital CVD (40): treatments or surgical procedures, in terms of pre-operative
pediatric cardiology is anticipated to benefit a lot from large risk stratification parameters to significantly curb/minimize
datasets and the deployment of artificial intelligence (41). perioperative morbidity and mortality rates. On the other
Artificial intelligence is anticipated to fully leverage and hand, administrative databases (20) may suffer from clerical
harness Big Data-based databases, potentially overcoming the inaccuracies, recording bias (due to the very nature of
issue of “classical” and “conventional” statistical techniques, the database and secondary to economic-financial incentives
including propensity score analysis and multivariate regression underlying the collection, and maintenance of the dataset),
modeling (42). Ahn et al. (43) developed CardioNet, a manually temporal changes in nosology and nomenclature systems as well
curated, standardized, and validated, comprehensive CVD- as in billing codes, and, finally, a dearth of several clinically
related database based on clinical information (either structured relevant parameters, including cardiology-specific variables
or unstructured) collected from 748,474 patients, that can and outcomes.
be utilized for Artificial intelligence analyses and provide A major issue seriously limiting the deployment of databases
insights on the care of patients with CVD. Barbieri et al. (44) and registries is related to their inter-operability and sometimes
combined the classical survival analysis (Cox proportional inconsistent use of definitions. Moreover, not all databases meet
hazard modeling) with a deep learning approach on a cohort regulatory standards (13) and are enough curated/validated.
of 2,164,872 New Zealanders aged 30–74 years. Predictors As such, data standardization and meta-data are urgently
of CVD events were found to be tobacco use in women and warranted (20).
chronic obstructive pulmonary disease (COPD) with acute Conversely, clinical studies, especially those relying on
lower respiratory infection in men, besides well-established “Small Data,” even though well-designed and well-conducted,

Frontiers in Cardiovascular Medicine | www.frontiersin.org 6 April 2022 | Volume 9 | Article 844296

Dai et al. Big Data in Cardiology

are generally statistically underpowered and are plagued by disease (CAD), inflammatory heart disease, rheumatic heart
several biases, including participants sampling and selection disease, and hypertensive heart disease, among others.
bias, which hinders the generalizability of the findings, with Vakili et al. (51) made efforts to combine all the OMICS-based
samples being not representative of the entire population. It specialties within a highly integrated, coherent, multi-OMICS
is also difficult to stratify according to a given cardiological approach termed as “panomics,” to shed light on the multi-
pharmacological treatment or surgical procedure if the sample factorial pathogenesis of CVD. The authors systematically mined
is particularly heterogeneous and the sample size does not the literature and were able to find 104 CVD-related OMICS-
allow to make sufficiently statistically robust and reliable based databases, 72 of which provided genomics/post-genomics
calculations. Confidence and certainty can increase with “Big and clinical measurements. Of these datasets, 59 and 65 databases
Data,” paralleling, however, the growth of complexity and were transcriptomic, epigenomic/methylomic, 41 proteomic, 42
associated computational costs (45, 46). Also, Big Data-based metabolomic, and 22 microbiomic.
databases can be affected by biases, as previously mentioned, Combing the scholarly literature, clinical and OMICS-based
such as recording or association biases and other statistical information, and exploiting the “diseasome” approach, Sarajlić
artifacts, like “reverse epidemiology” or “reverse causality” et al. (52) assessed the structure of the human protein-protein
(47). For instance, some database-based studies have found interaction (PPI) network to discover new CVD-related genes,
that body mass index (BMI), lipid profile, and blood pressure, that could be potential druggable targets. The authors found that
which usually predict a poor clinical outcome in the general these new genes were involved in intracellular signaling cascades,
population, become inverse prognostic predictors in chronic signaling transducing activity, enzyme binding, and intracellular
heart failure patients. Greater survival has been, indeed, receptor-mediated signaling pathways.
linked to overweight and obesity, hypercholesterolemia, Moreover, the unique and unprecedented convergence
and high values of blood pressure, which is rather counter- between different disciplines, such as nano-(bio-) engineering,
intuitive. Several hypotheses have been formulated, including three-dimensional (3D) printing and computational simulation,
the presence of the “malnutrition-inflammation complex molecular and mathematical modeling, and advanced and
syndrome” or “malnutrition–inflammation–cachexia syndrome”. sophisticated biostatistical techniques and Artificial Intelligence
However, some scholars think that it is more likely (and (Data Mining, Machine, and Deep Learning), are shaping
biologically/clinically plausible) that these findings are new paths and opportunities in the field of cardiological
statistical artifacts. practice and clinical research, enriching it, making it more
multi- and inter-disciplinary and complex, and more able
to address the biomedical challenges. Similarly, Dr. Elias
ROLES AND APPLICATIONS OF Zerhouni (53–55), who has served as Director of the
MOLECULAR BIG DATA IN CURRENT National Institutes of Health (NIH) from 2002 to 2008,
CARDIOLOGICAL PRACTICE AND has indicated such a unique convergence as the future
RESEARCH roadmap in the field of scholarly research, including the
cardiological arena.
Wet-lab and high-throughput technologies, including 3D printing is being increasingly utilized in biomedicine,
microarray chips, next-generation DNA and RNA sequencing and, in particular, in cardiology. Generally, mainly rigid
and whole-exome sequencing, chromatin-immunoprecipitation- anatomic models are produced, but the incorporation of dynamic
coupled sequencing, chromatin interaction analysis by paired- functionality is expected to dramatically advance preoperative
end tag sequencing (ChIA-PET), chromatin conformation cardiovascular surgical planning as well as hemodynamics
capture with sequencing, assay for transposase-accessible (56). 3D models can shed light on different CVD-related
chromatin with high-throughput sequencing (ATAC-Seq), pathophysiological conditions, thus complementing information
and mass-spectrometry-based proteomics analysis can obtained using classical imaging.
generate a wealth of molecular big data, paving the way for Moreover, molecular Big Data, alone or combined/integrated
a personalized/individualized rather than “one-size-fits-it-all” with epidemiological Big Data, can capture the landscape of
cardiology (48, 49). several cardiological diseases and events, either idiopathic or
Molecular big data can elucidate the mechanisms underlying congenital, including dilated cardiomyopathy and heart failure
the etiopathogenesis of a given heart disease and identify new (57, 58), among others.
potential druggable targets for the development of ad hoc
pharmacological therapies. Personalized cardiology can benefit ROLES AND APPLICATIONS OF BIG DATA
from genome-wide association and post-genomics studies (50,
51), aimed at the identification of new cardiogenic transcription
GENERATED BY IMAGING TECHNIQUES
factors, genotypic and phenotypic validations of potential AND WEARABLE TECHNOLOGIES/SMART
transcriptional regulators, and molecular/cellular mechanisms. SENSORS IN CURRENT CARDIOLOGICAL
CardioGenBase (50) is a literature-based, comprehensive PRACTICE AND RESEARCH
online resource tool, which extensively collects gene-
disease associations (over 1,500) for major CVD, including Latest technological achievements in the field of mobile
cerebrovascular disease, ischemic heart disease, coronary artery health (mHealth) and ubiquitous health (uHealth), with

Frontiers in Cardiovascular Medicine | www.frontiersin.org 7 April 2022 | Volume 9 | Article 844296

Dai et al. Big Data in Cardiology

smartphones, smart devices, smartwatches, and other wearable “personalized cardiology,” also becoming culturally sensitive
sensors (59) are revolutionizing the field of cardiology, and targeting specific populations, which are disproportionately
directly involving, and engaging the patient, improving their affected by non-communicable diseases, including CVD.
therapeutical adherence and compliance, and also enabling Concerning smart devices, such as smartwatches and
remote patient monitoring. smartphones, Prasitlumkum et al. (66) have conducted a
Wearable sensors of different types (bioelectric, mechano- systemic review and meta-analysis to quantitatively evaluate the
electric, optoelectronic, and ultrasonic wearable devices) enable accuracy of utilizing wearable devices for screening, detecting,
collecting cardiovascular vital signs (such as blood pressure, and properly diagnosing atrial fibrillation. The authors were
heart rate and heart rhythm, blood oxygen saturation, and able to compute excellent areas under the summary receiver
blood glucose, as well as brain waves, air quality, exposure operating characteristic (SROC) curves at 0.96 and 0.94,
to radiations, and other metrics) continuously, allowing early for smartphones and smartwatches, respectively. Sensitivity
intervention (60). and specificity were in the range of 94–96 and 93–94%
Gandhi et al. (61) conducted a systematic review of for the two kinds of smart/wearable devices, respectively:
the literature, investigating the effectiveness of mHealth they proved to be as diagnostically accurate and reliable
Interventions for the secondary prevention of CVD. The as gold standards, like photoplethysmography and single-
authors pooled 27 studies together, totaling 5,165 patients. lead electrocardiography.
mHealth was found to increase therapeutic adherence (with Signals and data generated by imaging techniques, like
an odds ratio, OR, of 4.51) as well as overall compliance, electrocardiography, computed tomography, or magnetic
either pharmacologic or non-pharmacologic (with an OR of resonance imaging, can be further processed, analyzed, and
3.86). Different targets were more likely to be met: namely, refined using artificial intelligence (67, 68). For instance,
blood pressure (OR 2.80), exercise and physical activity with MOCOnet (69) is a next-generation convolutional neural
reduced sedentary time and sitting (OR 2.55), but not smoking network that can significantly enhance and improve
cessation (OR 1.42), and lipid profile (OR 1.16). However, the quantitative cardiovascular magnetic resonance T1 mapping,
mHealth group did not differ from the standard-of-care group making it more robust, reliable, clinically meaningful,
in terms of hospitalizations and hospital readmissions (OR less prone to motion artifacts, and in a time-efficient
0.93). Few studies showed a statistically significant reduction manner. MOCOnet, being purely data-driven, outperforms
in angina (OR 0.23) and transient ischemic attack/stroke currently available methods for motion correction, which
recurrence in cerebrovascular disease patients (OR 0.18). The are model-driven.
cardiovascular mortality rate was computed to be lower, even Finally, radiomics and radiogenomics are highly innovative
though not achieving the significance threshold (OR 0.19). translational fields of research aimed at mining, retrieving,
Similar results could be replicated in a more updated systematic merging, processing, analyzing, and extracting clinically
review and meta-analysis conducted by Akinosun et al. (62) meaningful patterns and interpretations from large-
and in the systematic review of the literature by Spaulding scale, high-dimensional datasets generated by clinical
et al. (63). imaging techniques and tools (70), including cardiac
Wali et al. (64) showed that mHealth interventions can computed tomography angiography and cardiac magnetic
be particularly useful in reaching vulnerable and underserved resonance. Latest advancements concerning more and more
communities, including aboriginal and indigenous individuals or sophisticated protocols enable the integration of imaging
subjects residing in low- and middle-income countries. Usually, features and molecular profiling to identify relevant and
these individuals are excluded or are under-represented in clinically meaningful biomarkers and signatures (such as
clinical trials. atherosclerotic lesions, coronary plaques, and myocardial
Gamification and gamified mobile applications (apps) structural abnormalities) related to diagnosis, prognosis, and
represent another interesting and promising ramification response to treatment. Supervised and unsupervised artificial
of the digital health arena. Davis et al. (65) have intelligence, including deep and machine learning, can further
performed a systematic literature review, synthesizing combine and aggregate data and assist the development of
seven studies, totaling 657 patients. The authors found risk models and tools that can facilitate clinical diagnostic and
that gamification resulted in improved adoption of prognostic procedures.
healthier lifestyles and behaviors (for instance, in terms In the field of cardiological research, radiomics, and
of the practice of exercise and physical activity), better radiogenomics can be utilized for the characterization,
biochemical profile, enhanced mood, and motivation. profiling/phenotyping, and risk stratification of coronary
Interestingly, also CVD-related health literacy and heart disease (CHD), hypertrophic cardiomyopathy, ischemic
knowledge improved in a significant way, even though some heart disease, and cerebrovascular disease (70–73), among
parameters, such as blood pressure, body mass index, self- others. However, also given its recency, still too much
management, and therapeutical compliance, were comparable has to be explored in this field. On the other hand, it
with standard-of-care. can be anticipated that radiomics, radiogenomics, and
To summarize, mHealth and digital health-based other Big Data generated by wearable/smart devices and
interventions, including telemonitoring (telecardiology) or sensors will profoundly impact both cardiological practice
text messaging, can be customized, meeting the needs of and research.

Frontiers in Cardiovascular Medicine | www.frontiersin.org 8 April 2022 | Volume 9 | Article 844296

Dai et al. Big Data in Cardiology

ROLES AND APPLICATIONS OF Wikipedia pages were deemed of moderate quality, with 8.5%
INFODEMIOLOGICAL BIG DATA IN being of good and poor quality, respectively. Despite clinical
presentation and etiopathogenesis of CVD being treated and
CURRENT CARDIOLOGY RESEARCH
discussed, several sections, including the pathophysiology, signs
Infodemiology (a portmanteau of “information” and and symptoms, diagnosis, and management, were not always
“epidemiology”) and infoveillance (a combination of the accurate and adequately scholarly referenced. Several entries
words “information” and “surveillance”) represent a highly exhibited errors and omissions. The readability was at the level
innovative discipline, at the intersection of computer, data, of collegiate subjects.
and behavioral science, aimed at shedding light on the CVD patients use the internet as a low-cost and easily available
determinants of computational and digital activities (such source of personal healthcare information, to learn more about
as web queries, use of social media, posting on social networks, their condition/disorder, as well as about potential treatment
and production/consumption of online material) (74, 75). options and CVD physicians and surgeons (81). According to
Researchers in the field of infodemiology and infoveillance a recent survey by Jones et al. (81), 74.3% of the interviewees
make use of resources that enable to assess information demand surfed the internet, with 63% utilizing it daily. In the case the
and consumption, such as Google Trends, which is an open- patient could not directly access the web, a family member was
source tool that enables to track and monitor web searches willing to do so on their behalf. The authors concluded that
conducted using the Google search engine. most patients (∼85%) utilized the internet, being particularly
Infodemiology and infoveillance enable to track the interested in local information.
effectiveness of awareness campaigns, such as the “Go Red Practitioners and residents in the field of cardiology should
for Women” (76), which is a social initiative aimed at improving be aware of these findings in that the web is often consulted
and enhancing CVD- and stroke-related literacy among women. by patients with CVD. Locally delivered Web-based information
Suero-Abreu et al. (77) investigated the impact of “Go Red service is particularly requested and appreciated by CVD
for Women” on health information-seeking behavior, utilizing patients. The web can be used to deliver high-quality, educational
Google Trends. Authors found increased search volumes related material and empower the patient, by enhancing their literacy,
to the awareness campaign and various CVD-related terms over collecting patient-generated/reported outcomes (PROMs), and
15 years. However, stroke-related digital searches were not found health-related behaviors and attitudes, devising ad hoc social
to be increased over the study period. campaigns and monitoring their impact on health-related digital
Dzaye et al. (78) have exploited infodemiology and seeking behaviors.
infoveillance techniques to assess public interest toward
CVD and related comorbidities during the “Coronavirus Disease
2019” (COVID-19) pandemic. According to some studies,
“PARTICIPATORY CARDIOLOGY”:
attention to CVD would have decreased, despite the negative INTEGRATING BASIC AND
relationship between CVD and infection. Patients suffering TRANSLATIONAL CARDIOLOGY AND
from CVD or with risk factors for CVD have been consistently CITIZEN SCIENCE
reported to exhibit worse outcomes than their CVD-free
counterparts. Authors found that digital interest in terms like Big Data can also contribute to an emerging super-specialization
exercise or physical activity and cigarettes had increased (by within the field of cardiology: the so-called “participatory
18%) and decreased (by 52.5%), respectively. Noteworthy, cardiology”, in such a way promoting public participation in
interest in terms like statin, lipid profile, low-density lipoprotein the field of cardiological practice and clinical research, creating
(LDL), and hemoglobin A1C, had significantly increased at well, “global collaborative social networks”, and integrating basic
after a previous decline over time. and translational cardiology and citizen science (82, 83). This
On the other hand, according to a research study by the is of paramount importance especially in low- and middle-
same group (79), the first months of the COVID-19 pandemic income countries and would help curb/reduce health disparities
were paralleled by a decrease in search interest for myocardial and iniquities.
infarction and acute coronary syndrome (ACS), potentially A systematic review conducted by Wali et al. (64) has shown
explaining the excess cardiovascular mortality despite a marked that establishing collaborative partnerships and relationships
reduction in hospitalization for ACS. with community members – especially those from underserved
To summarize, search engines and other non-conventional and vulnerable populations – would significantly improve and
data streams appear to be valuable and promising tools enhance the effectiveness of the cardiological intervention
that can provide insights on health information-seeking by ensuring it was devised and implemented within the
behaviors and evaluate the effectiveness of social campaigns and appropriate context.
other interventions. Participatory cardiology, as a branch of participatory
The quality of cardiology-related websites and, more in medicine, gives a new value and importance to the patient,
general, online material is highly heterogenous and variable who is the “real teacher,” quoting a famous statement of the
both in terms of content and information provided. For Canadian physician and cardiologist Sir William Osler (1849–
instance, Azer et al. (80) assessed the quality, accuracy, and 1919) enunciated in 1903. Latest scientific and technological
readability of Wikipedia pages concerning CVD. About 83% of advancements and current trends in clinical practice and

Frontiers in Cardiovascular Medicine | www.frontiersin.org 9 April 2022 | Volume 9 | Article 844296

Dai et al. Big Data in Cardiology

research, especially in the cardiological arena, have gradually uploaded and shared, enabling other scholars and researchers
shifted the practitioners’ attention and interest toward patient’s to replicate findings. However, there exist some privacy and
“subjective” outcomes (satisfaction, pain, quality of life, etc.), bioethical issues. Data de-identification or anonymization or
besides “objective” clinical outcomes (healthcare resources pseudonymization or masking can ensure re-use of potentially
uptake and consumption, healthcare processes and provisions sensitive, personal, and legally restricted data, preserving
delivery, morbidity and mortality rates). scalability and performance, also if this technique could be
However, for most cardiologists and cardiological surgeons, challenging and not trivial to implement (88).
the world of PROMs represents a still “unchartered health Molecular big data require extensive processing of data,
care environment” (84), the navigation of which, by which can be quite expensive, time- and resource-consuming.
incorporating “mission, values and culture” (85, 86), can Moreover, the results of the various studies have to be
advance cardiological practice and research. There are several reconciled, depending on the type of tissue/cell studies, the
gaps in the implementation and full incorporation of PROMs molecular technique applied, etc. This can lead to a “false
within the daily routine cardiological practice. According to the discovery” of biomarkers. Recently, meta-analyses of molecular
“International Consortium for Health Outcomes Measurement” big data pooling together various samples have enabled to
(ICHOM), while there exist several national, international, increase the statistical power and, thus, the reliability and
and trial registries for heart failure, very few of them can be trustworthiness of the discovery. Ensuring reproducibility
considered as patient-centered and standardized guidelines and clinical meaningfulness of results should be a research
and checklists guiding the process of properly, effectively, and priority (89).
meaningfully using PROMs are lacking. To fill in this gap, Big data generated by information and communication
the ICHOM has developed a 17-item dataset, which consists technologies can be affected by privacy and bioethical issues due
of several domains (functional-, psycho-social-, burden of to the pervasive and ubiquitous nature of the devices.
care-, and survival-related outcomes). This set, which also Finally, concerning computational/digital big data, there are
includes PROMs besides clinical/objective measurements, and some issues affecting their usage, like the lack of transparency
administrative data, enables to compare consistently heart failure related to the algorithm deployed to retrieve, collect, process, and
management and treatment across several healthcare providers store data.
and various regions, globally (87).
CONCLUSIONS AND FUTURE
LIMITATIONS AND SHORTCOMINGS OF PROSPECTS
BIG DATA IN THE FIELD OF CARDIOLOGY
Big Data is increasingly having a more and more relevant
Table 4 overviews the major limitations and shortcomings of Big role, being highly ubiquitous and pervasive in contemporary
Data in the field of cardiology based on the type of source/channel society, permeating it and paving the way for new, unprecedented
that generates them. Basically, these pitfalls are of a two-fold perspectives in biomedicine, including cardiology. Big Data
nature: legal/bioethical (in terms of legal requirements and can be a real paradigm shift that revolutionizes cardiological
restrictions, legislation, privacy, and data sharing policies) and practice and clinical research. However, some methodological
methodological. issues should be properly addressed, and some ethical issues
Epidemiological/clinical Big Data can be affected by should be considered. Therefore, further research in the field is
inconsistencies according to the type of study and its design urgently warranted.
(registry-based vs. individual – single or multi-center –
investigations). Also, database-based studies may give rise AUTHOR CONTRIBUTIONS
to contrasting findings based on the reason and scope data
were collected (clinical, administrative, or financial purposes). HD and NB conceived and drafted the paper. All other authors
Optimizing databases and ensuring inter-operability could critically revised it. All authors contributed to the article and
overcome these issues. Moreover, datasets can also be publicly approved the submitted version.

REFERENCES 3. Roth GA, Mensah GA, Johnson CO, Addolorato G, Ammirati E, Baddour
LM, et al. GBD-NHLBI-JACC global burden of cardiovascular diseases writing
1. Mathers CD. History of the global burden of disease assessment group. global burden of cardiovascular diseases and risk factors, 1990–2019:
at the world health organization. Arch Public Health. (2020) 78:77. update from the GBD 2019 study. J Am Coll Cardiol. (2020) 76:2982–3021.
doi: 10.1186/s13690-020-00458-3 doi: 10.1016/j.jacc.2020.11.010
2. GBD 2017 SDG Collaborators. Measuring progress from 1990 4. Mensah GA, Roth GA, Fuster V. The global burden of cardiovascular diseases
to 2017 and projecting attainment to 2030 of the health-related and risk factors: 2020 and beyond. J Am Coll Cardiol. (2019) 74:2529–32.
sustainable development Goals for 195 countries and territories: doi: 10.1016/j.jacc.2019.10.009
a systematic analysis for the global burden of disease study 5. Kirchhof P, Sipido KR, Cowie MR, Eschenhagen T, Fox KA, Katus H,
2017. Lancet. (2018) 392:2091–138. doi: 10.1016/S0140-6736(18)3 et al. ESC CRT R&D and European affairs work shop on personalized
2281-5 medicine. the continuum of personalized cardiovascular medicine: a position

Frontiers in Cardiovascular Medicine | www.frontiersin.org 10 April 2022 | Volume 9 | Article 844296

Dai et al. Big Data in Cardiology

paper of the european society of cardiology. Eur Heart J. (2014) 35:3250–7. 25. Brindis RG, Fitzgerald S, Anderson HV, Shaw RE, Weintraub WS, Williams
doi: 10.1093/eurheartj/ehu312 JF. The American college of cardiology-national cardiovascular data registry
6. Lee MS, Flammer AJ, Lerman LO, Lerman A. Personalized (ACC-NCDR): building a national clinical data repository. J Am Coll Cardiol.
medicine in cardiovascular diseases. Korean Circ J. (2012) 42:583–91. (2001) 37:2240–5. doi: 10.1016/s0735-1097(01)01372-9
doi: 10.4070/kcj.2012.42.9.583 26. Shwartz M, Ren J, Peköz EA, Wang X, Cohen AB, Restuccia JD.
7. Auffray C, Charron D, Hood L. Predictive, preventive, personalized and Estimating a composite measure of hospital quality from the hospital
participatory medicine: back to the future. Genome Med. (2010) 2:57. compare database: differences when using a Bayesian hierarchical latent
doi: 10.1186/gm178 variable model vs. denominator-based weights. Med Care. (2008) 46:778–85.
8. Sagner M, McNeil A, Puska P, Auffray C, Price ND, Hood L, et al. The P4 doi: 10.1097/MLR.0b013e31817893dc
health spectrum - a predictive, preventive, personalized and participatory 27. Gunter N, Moore L, Odom P. Cooperative cardiovascular project. JSC Med
continuum for promoting healthspan. Prog Cardiovasc Dis. (2017) 59:506–21. Assoc. (1997) 93:177–9.
doi: 10.1016/j.pcad.2016.08.002 28. Ramunno LD, Dodds TA, Traven ND. Cooperative cardiovascular project
9. Trachana K, Bargaje R, Glusman G, Price ND, Huang S, Hood LE. (CCP) quality improvement in Maine, New Hampshire, and Vermont. Eval
Taking systems medicine to heart. Circ Res. (2018) 122:1276–89. Health Prof. (1998) 21:442–60. doi: 10.1177/016327879802100404
doi: 10.1161/CIRCRESAHA.117.310999 29. Kelsey SF, Miller DP, Holubkov R, Lu AS, Cowley MJ, Faxon DP, et al. Results
10. Bhavnani SP, Parakh K, Atreja A, Druz R, Graham GN, Hayek SS, et al. of percutaneous transluminal coronary angioplasty in patients greater than or
2017 Roadmap for Innovation-ACC health policy statement on healthcare equal to 65 years of age (from the 1985 to 1986 national heart, lung, and blood
transformation in the era of digital health, big data, and precision health: institute’s coronary angioplasty registry). Am J Cardiol. (1990) 66:1033–8.
a report of the american college of cardiology task force on health policy doi: 10.1016/0002-9149(90)90500-z
statements and systems of care. J Am Coll Cardiol. (2017) 70:2696–718. 30. Ayalon N, Jacobs AK. Incomplete revascularization in patients treated
doi: 10.1016/j.jacc.2017.10.018 with percutaneous coronary intervention: when enough is enough. JACC
11. Weintraub WS. Role of big data in cardiovascular research. J Am Heart Assoc. Cardiovasc Interv. (2016) 9:216–8. doi: 10.1016/j.jcin.2015.12.001
(2019) 8:e012791. doi: 10.1161/JAHA.119.012791 31. Hansen JW, Foy A, Yadav P, Gilchrist IC, Kozak M, Stebbins A, et al.
12. Hemkens LG. How routinely collected data for randomized trials provide Death and dialysis after transcatheter aortic valve replacement: an analysis
long-term randomized real-world evidence. JAMA Netw Open. (2018) of the STS/ACC TVT registry. JACC Cardiovasc Interv. (2017) 10:2064–75.
1:e186014. doi: 10.1001/jamanetworkopen.2018.6014 doi: 10.1016/j.jcin.2017.09.001
13. Hernandez-Boussard T, Monda KL, Crespo BC, Riskin D. Real world 32. Moussa I, Hermann A, Messenger JC, Dehmer GJ, Weaver WD, Rumsfeld
evidence in cardiovascular medicine: ensuring data validity in electronic JS, et al. The NCDR CathPCI registry: a US national perspective on care and
health record-based studies. J Am Med Inform Assoc. (2019) 26:1189–94. outcomes for percutaneous coronary intervention. Heart. (2013) 99:297–303.
doi: 10.1093/jamia/ocz119 doi: 10.1136/heartjnl-2012-303379
14. Stapff MP. Using real world data to assess cardiovascular outcomes of 33. Kramer CM, Appelbaum E, Desai MY, Desvigne-Nickens P, DiMarco
two antidiabetic treatment classes. World J Diabetes. (2018) 9:252–7. JP, Friedrich MG, et al. Hypertrophic cardiomyopathy registry:
doi: 10.4239/wjd.v9.i12.252 the rationale and design of an international, observational study
15. Stapff M, Hilderbrand S. First-line treatment of essential hypertension: a real- of hypertrophic cardiomyopathy. Am Heart J. (2015) 170:223–30.
world analysis across four antihypertensive treatment classes. J Clin Hypertens doi: 10.1016/j.ahj.2015.05.013
(Greenwich). (2019) 21:627–34. doi: 10.1111/jch.13531 34. Xiao H, Ali S, Zhang Z, Sarfraz MS, Zhang F, Faisal M. Big data,
16. Buckley BJR, Harrison SL, Fazio-Eynullayeva E, Underhill P, Lane DA, extracting insights, comprehension, and analytics in cardiology: an
Thijssen DHJ, et al. Exercise-based cardiac rehabilitation and all-cause overview. J Healthc Eng. (2021) 2021:6635463. doi: 10.1155/2021/66
mortality among patients with atrial fibrillation. J Am Heart Assoc. (2021) 35463
10:e020804. doi: 10.1161/JAHA.121.020804 35. Shah RU, Rumsfeld JS. Big Data in Cardiology. Eur Heart J. (2017) 38:1865–7.
17. Buckley BJR, Harrison SL, Fazio-Eynullayeva E, Underhill P, doi: 10.1093/eurheartj/ehx284
Sankaranarayanan R, Wright DJ, et al. Cardiac rehabilitation and all- 36. Barabási AL. Network medicine–from obesity to the “diseasome”. N Engl J
cause mortality in patients with heart failure: a retrospective cohort study. Med. (2007) 357:404–7. doi: 10.1056/NEJMe078114
Eur J Prev Cardiol. (2021) 28:1704–10. doi: 10.1093/eurjpc/zwab035 37. Meireles MA, Golçalves J, Neves J. Acute heart failure comorbidome:
18. Alkhouli M, Nanjundappa A, Annie F, Bates MC, Bhatt DL. Sex differences in the impact of everything else. Acta Med Port. (2020) 33:109–15.
case fatality rate of COVID-19: insights from a multinational registry. Mayo doi: 10.20344/amp.11051
Clin Proc. (2020) 95:1613–20. doi: 10.1016/j.mayocp.2020.05.014 38. Cohn SL, Fernandez Ros N. Comparison of 4 cardiac risk calculators in
19. Harrison SL, Fazio-Eynullayeva E, Lane DA, Underhill P, Lip GYH. Atrial predicting postoperative cardiac complications after non-cardiac operations.
fibrillation and the risk of 30-day incident thromboembolic events, and Am J Cardiol. (2018) 121:125–30. doi: 10.1016/j.amjcard.2017.09.031
mortality in adults ≥50 years with COVID-19. J Arrhythm. (2020) 37:231–7. 39. Lee KT, Hour AL, Shia BC, Chu PH. The application and future of big database
doi: 10.1002/joa3.12458 studies in cardiology: a single-center experience. Acta Cardiol Sin. (2017)
20. Meltzer SN, Weintraub WS. The role of national registries in improving 33:581–7. doi: 10.6515/ACS20170331B
quality of care and outcomes for cardiovascular disease. Methodist Debakey 40. Van den Eynde J, Manlhiot C, Van De Bruaene A, Diller GP, Frangi
Cardiovasc J. (2020) 16:205–11. doi: 10.14797/mdcj-16-3-205 AF, Budts W, et al. Medicine-based evidence in congenital heart disease:
21. Heart Rhythm Society, Food and Drug Administration. Proceedings how artificial intelligence can guide treatment decisions for individual
Document from the Policy Conference on Pacemaker and ICD Performance. patients. Front Cardiovasc Med. (2021) 8:798215. doi: 10.3389/fcvm.2021.7
Kappenberger L, panelist. (2005). Washington, DC. 98215
22. Myers WO, Blackstone EH, Davis K, Foster ED, Kaiser GC, CASS. Registry 41. Van den Eynde J, Kutty S, Danford DA, Manlhiot C. Artificial intelligence in
long term surgical survival. Coronary artery surgery study. J Am Coll Cardiol. pediatric cardiology: taking baby steps in the big world of data. Curr Opin
(1999) 33:488–98. doi: 10.1016/s0735-1097(98)00563-4 Cardiol. (2022) 37:130–6. doi: 10.1097/HCO.0000000000000927
23. Jacobs JP, Shahian DM, Prager RL, Edwards FH, McDonald D, Han JM, et al. 42. de Marvao A, Dawes TJ, Howard JP, O’Regan DP. Artificial intelligence and
Introduction to the STS national database series: outcomes analysis, quality the cardiologist: what you need to know for 2020. Heart. (2020) 106:399–400.
improvement, and patient safety. Ann Thorac Surg. (2015) 100:1992–2000. doi: 10.1136/heartjnl-2019-316033
doi: 10.1016/j.athoracsur.2015.10.060 43. Ahn I, Na W, Kwon O, Yang DH, Park GM, Gwon H, et al. CardioNet:
24. Smaha LA. American Heart Association. The American Heart Association a manually curated database for artificial intelligence-based research on
Get With The Guidelines program. Am Heart J. (2004) 148(5 Suppl):S46–8. cardiovascular diseases. BMC Med Inform Decis Mak. (2021) 21:29.
doi: 10.1016/j.ahj.2004.09.015 doi: 10.1186/s12911-021-01392-2

Frontiers in Cardiovascular Medicine | www.frontiersin.org 11 April 2022 | Volume 9 | Article 844296

Dai et al. Big Data in Cardiology

44. Barbieri S, Mehta S, Wu B, Bharat C, Poppe K, Jorm L, et al. Predicting 65. Davis AJ, Parker HM, Gallagher R. Gamified applications for secondary
cardiovascular risk from national administrative databases using a combined prevention in patients with high cardiovascular disease risk: a systematic
survival analysis and deep learning approach. Int J Epidemiol. (2021):dyab258. review of effectiveness and acceptability. J Clin Nurs. (2021) 30:3001–10.
doi: 10.1093/ije/dyab258 doi: 10.1111/jocn.15808
45. Rumsfeld JS, Joynt KE, Maddox TM. Big data analytics to improve 66. Prasitlumkum N, Cheungpasitporn W, Chokesuwattanaskul A, Thangjui
cardiovascular care: promise and challenges. Nat Rev Cardiol. (2016) 13:350– S, Thongprayoon C, Bathini T, et al. Diagnostic accuracy of smart
9. doi: 10.1038/nrcardio.2016.42 gadgets/wearable devices in detecting atrial fibrillation: a systematic
46. Silverio A, Cavallo P, De Rosa R, Galasso G. Big health data and cardiovascular review and meta-analysis. Arch Cardiovasc Dis. (2021) 114:4–16.
diseases: a challenge for research, an opportunity for clinical care. Front Med. doi: 10.1016/j.acvd.2020.05.015
(2019) 6:36. doi: 10.3389/fmed.2019.00036 67. Ben Ali W, Pesaranghader A, Avram R, Overtchouk P, Perrin N, Laffite
47. Kalantar-Zadeh K, Block G, Horwich T, Fonarow GC. Reverse epidemiology S, et al. Implementing machine learning in interventional cardiology: the
of conventional cardiovascular risk factors in patients with chronic heart benefits are worth the trouble. Front Cardiovasc Med. (2021) 8:711401.
failure. J Am Coll Cardiol. (2004) 43:1439–44. doi: 10.1016/j.jacc.2003. doi: 10.3389/fcvm.2021.711401
11.039 68. Seetharam K, Shrestha S, Sengupta PP. Cardiovascular Imaging and
48. Slagle CE, Conlon FL. Emerging field of cardiomics: high-throughput Intervention Through the Lens of Artificial Intelligence. Interv Cardiol. (2021)
investigations into transcriptional regulation of cardiovascular development 16:e31. doi: 10.15420/icr.2020.04
and disease. Trends Genet. (2016) 32:707–16. doi: 10.1016/j.tig.2016.09.002 69. Gonzales RA, Zhang Q, Papiez BW, Werys K, Lukaschuk E, Popescu
49. Lau E, Wu JC. Omics, Big data, and precision medicine in cardiovascular IA, et al. MOCOnet: robust motion correction of cardiovascular
sciences. Circ Res. (2018) 122:1165–8. doi: 10.1161/CIRCRESAHA.118.313161 magnetic resonance T1 mapping using convolutional neural networks.
50. V A, Nayar PG, Murugesan R, Mary B, P D, Ahmed SS. cardiogenbase: a Front Cardiovasc Med. (2021) 8:768245. doi: 10.3389/fcvm.2021.7
literature based multi-omics database for major cardiovascular diseases. PLoS 68245
ONE. (2015) 10:e0143188. doi: 10.1371/journal.pone.0143188 70. Infante T, Cavaliere C, Punzo B, Grimaldi V, Salvatore M, Napoli
51. Vakili D, Radenkovic D, Chawla S, Bhatt DL. Panomics: new databases C. Radiogenomics and artificial intelligence approaches applied to
for advancing cardiology. Front Cardiovasc Med. (2021) 8:587768. cardiac computed tomography angiography and cardiac magnetic
doi: 10.3389/fcvm.2021.587768 resonance for precision medicine in coronary heart disease: a
52. Sarajlić A, Janjić V, Stojković N, Radak D, PrŽulj N. Network topology systematic review. Circ Cardiovasc Imaging. (2021) 14:1133–46.
reveals key cardiovascular disease genes. PLoS ONE. (2013) 8:e71537. doi: 10.1161/CIRCIMAGING.121.013025
doi: 10.1371/journal.pone.0071537 71. Antonopoulos AS, Angelopoulos A, Tsioufis K, Antoniades C, Tousoulis
53. Zerhouni E. Medicine. The NIH roadmap. Science. (2003) 302:63–72. D. Cardiovascular risk stratification by coronary computed tomography
doi: 10.1126/science.1091867 angiography imaging: current state-of-the-art. Eur J Prev Cardiol. (2021)
54. Zerhouni EA. Clinical research at a crossroads: the NIH roadmap. J Investig :zwab067. doi: 10.1093/eurjpc/zwab067
Med. (2006) 54:171–3. doi: 10.2310/6650.2006.X0016 72. Antonopoulos AS, Boutsikou M, Simantiris S, Angelopoulos A, Lazaros G,
55. Collins FS, Wilder EL, Zerhouni E. Funding transdisciplinary research. Panagiotopoulos I, et al. Machine learning of native T1 mapping radiomics
NIH roadmap/common fund at 10 years. Science. (2014) 345:274–6. for classification of hypertrophic cardiomyopathy phenotypes. Sci Rep. (2021)
doi: 10.1126/science.1255860 11:23596. doi: 10.1038/s41598-021-02971-z
56. Wang H, Song H, Yang Y, Cao Q, Hu Y, Chen J, et al. Three- 73. Rauseo E, Izquierdo Morcillo C, Raisi-Estabragh Z, Gkontra P, Aung
dimensional printing for cardiovascular diseases: from anatomical N, Lekadir K, et al. New imaging signatures of cardiac alterations
modeling to dynamic functionality. Biomed Eng Online. (2020) 19:76. in ischaemic heart disease and cerebrovascular disease using CMR
doi: 10.1186/s12938-020-00822-y radiomics. Front Cardiovasc Med. (2021) 8:716577. doi: 10.3389/fcvm.2021.7
57. Sammani A, Baas AF, Asselbergs FW, Te Riele ASJM. Diagnosis and risk 16577
prediction of dilated cardiomyopathy in the era of big data and genomics. J 74. Eysenbach G. Infodemiology: the epidemiology of (mis)information. Am J
Clin Med. (2021) 10:921. doi: 10.3390/jcm10050921 Med. (2002) 113:763–5. doi: 10.1016/s0002-9343(02)01473-0
58. Lanzer JD, Leuschner F, Kramann R, Levinson RT, Saez-Rodriguez J. Big data 75. Eysenbach G. Websites on screening for breast cancer: “infodemiology”
approaches in heart failure research. Curr Heart Fail Rep. (2020) 17:213–24. studies have surely had their day. BMJ. (2004) 328:769.
doi: 10.1007/s11897-020-00469-9 doi: 10.1136/bmj.328.7442.769-b
59. Chen S, Qi J, Fan S, Qiao Z, Yeo JC, Lim CT. Flexible wearable sensors for 76. Kling JM, Miller VM, Mankad R, Wilansky S, Wu Q, Zais TG, et al. Go Red for
cardiovascular health monitoring. Adv Healthc Mater. (2021) 10:e2100116. Women cardiovascular health-screening evaluation: the dichotomy between
doi: 10.1002/adhm.202100116 awareness and perception of cardiovascular risk in the community. J Womens
60. Direito A, Rawstorn J, Mair J, Daryabeygi-Khotbehsara R, Maddison R, Tai Health. (2013) 22:210–8. doi: 10.1089/jwh.2012.3744
ES. Multifactorial e- and mHealth interventions for cardiovascular disease 77. Suero-Abreu GA, Barajas-Ochoa A, Perez-Peralta A, Rojas E, Berkowitz R.
primary prevention: protocol for a systematic review and meta-analysis Assessment of the effect of the go red for women campaign on search engine
of randomised controlled trials. Digit Health. (2019) 5:2055207619890480. queries for cardiovascular disease in women. Cardiol Res. (2020) 11:348–52.
doi: 10.1177/2055207619890480 doi: 10.14740/cr1107
61. Gandhi S, Chen S, Hong L, Sun K, Gong E, Li C, et al. Effect of mobile 78. Dzaye O, Adelhoefer S, Boakye E, Blaha MJ. Cardiovascular-related health
health interventions on the secondary prevention of cardiovascular disease: behaviors and lifestyle during the COVID-19 pandemic: an infodemiology
systematic review and meta-analysis. Can J Cardiol. (2017) 33:219–31. study. Am J Prev Cardiol. (2021) 5:100148. doi: 10.1016/j.ajpc.2021.100148
doi: 10.1016/j.cjca.2016.08.017 79. Dzaye O, Duebgen M, Berning P, Graham G, Martin SS, Blaha MJ.
62. Akinosun AS, Polson R, Diaz-Skeete Y, De Kock JH, Carragher L, Leslie S, et al. Understanding myocardial infarction trends during the early COVID-
Digital technology interventions for risk factor modification in patients with 19 pandemic: an infodemiology study. Intern Med J. (2021) 51:1328–31.
cardiovascular disease: systematic review and meta-analysis. JMIR Mhealth doi: 10.1111/imj.15399
Uhealth. (2021) 9:e21061. doi: 10.2196/21061 80. Azer SA, AlSwaidan NM, Alshwairikh LA, AlShammari JM. Accuracy
63. Spaulding EM, Marvel FA, Piasecki RJ, Martin SS, Allen JK. User engagement and readability of cardiovascular entries on Wikipedia: are they reliable
with smartphone apps and cardiovascular disease risk factor outcomes: learning resources for medical students? BMJ Open. (2015) 5:e008187.
systematic review. JMIR Cardio. (2021) 5:e18834. doi: 10.2196/18834 doi: 10.1136/bmjopen-2015-008187
64. Wali S, Hussain-Shamsy N, Ross H, Cafazzo J. Investigating the use of 81. Jones J, Cassie S, Thompson M, Atherton I, Leslie SJ. Delivering healthcare
mobile health interventions in vulnerable populations for cardiovascular information via the internet: cardiac patients’ access, usage, perceptions of
disease management: scoping review. JMIR Mhealth Uhealth. (2019) 7:e14275. usefulness, and web site content preferences. Telemed J E Health. (2014)
doi: 10.2196/14275 20:223–7. doi: 10.1089/tmj.2013.0182

Frontiers in Cardiovascular Medicine | www.frontiersin.org 12 April 2022 | Volume 9 | Article 844296

Dai et al. Big Data in Cardiology

82. Okop KJ, Murphy K, Lambert EV, Kedir K, Getachew H, Howe R, 88. Meystre SM, Friedlin FJ, South BR, Shen S, Samore MH. Automatic
et al. Community-driven citizen science approach to explore cardiovascular de-identification of textual documents in the electronic health record:
disease risk perception, and develop prevention advocacy strategies in sub- a review of recent research. BMC Med Res Methodol. (2010) 10:70.
Saharan Africa: a programme protocol. Res Involv Engagem. (2021) 7:11. doi: 10.1186/1471-2288-10-70
doi: 10.1186/s40900-020-00246-x 89. Ren AH, Fiala CA, Diamandis EP, Kulasingam V. Pitfalls in cancer biomarker
83. Mensah GA, Cooper RS, Siega-Riz AM, Cooper LA, Smith JD, Brown discovery and validation with emphasis on circulating tumor DNA. Cancer
CH, et al. Reducing cardiovascular disparities through community-engaged Epidemiol Biomarkers Prev. (2020) 29:2568–74. doi: 10.1158/1055-9965
implementation research: a national heart, lung, and blood institute workshop
report. Circ Res. (2018) 122:213–30. doi: 10.1161/CIRCRESAHA.117.3 Conflict of Interest: The authors declare that the research was conducted in the
12243 absence of any commercial or financial relationships that could be construed as a
84. Anoushiravani AA, Patton J, Sayeed Z, El-Othmani MM, Saleh KJ. Big data, potential conflict of interest.
big research: implementing population health-based research models and
integrating care to reduce cost and improve outcomes. Orthop Clin North Am. Publisher’s Note: All claims expressed in this article are solely those of the authors
(2016) 47:717–24. doi: 10.1016/j.ocl.2016.05.008 and do not necessarily represent those of their affiliated organizations, or those of
85. Anker SD, Agewall S, Borggrefe M, Calvert M, Jaime Caro J, Cowie MR, et al.
the publisher, the editors and the reviewers. Any product that may be evaluated in
The importance of patient-reported outcomes: a call for their comprehensive
this article, or claim that may be made by its manufacturer, is not guaranteed or
integration in cardiovascular clinical trials. Eur Heart J. (2014) 35:2001–9.
86. Chen Y, Nagendran M, Gomes M, Wharton PV, Raine R, Lambiase PD. Gaps endorsed by the publisher.
in patient reported outcome measures in randomised clinical trials of cardiac
catheter ablation: a systematic review. Eur Heart J Qual Care Clin Outcomes. Copyright © 2022 Dai, Younis, Kong, Puce, Jabbour, Yuan and Bragazzi. This is an
(2020) 6:234–42. doi: 10.1093/ehjqcco/qcaa022 open-access article distributed under the terms of the Creative Commons Attribution
87. Burns DJ, Arora J, Okunade O, Beltrame JF, Bernardez-Pereira S, License (CC BY). The use, distribution or reproduction in other forums is permitted,
Crespo-Leiro MG, et al. International consortium for health outcomes provided the original author(s) and the copyright owner(s) are credited and that the
measurement (ichom): standardized patient-centered outcomes measurement original publication in this journal is cited, in accordance with accepted academic
set for heart failure patients. JACC: Heart Failure. (2020) 8:212–22. practice. No use, distribution or reproduction is permitted which does not comply
doi: 10.1016/j.jchf.2019.09.007 with these terms.

Frontiers in Cardiovascular Medicine | www.frontiersin.org 13 April 2022 | Volume 9 | Article 844296