Author + information
- Robert A. Harrington, MD∗ ()
- ↵∗Reprint requests and correspondence:
Dr. Robert A. Harrington, Department of Medicine, Stanford University, 300 Pasteur Drive, S-102, Stanford, California 94305.
We are in the era of “big data” in cardiovascular medicine (1). This brings both excitement and challenges to the clinician who is making dozens, if not hundreds, of clinical decisions every day in the care of individual patients. Enormous amounts of data are available on each patient, much of which is increasingly available in electronic health records (EHR), but there are also data from clinical trials and observational studies on population-level outcomes associated with certain diagnostic and therapeutic options. Providers are expected to know about all of this information, assess its quality, aggregate it, analyze it, and interpret it in a way that increases the likelihood of excellent outcomes for each patient. This is a daunting task to accomplish for a single patient, and yet it must be repeated almost continuously in the context of a busy clinical practice.
The American College of Cardiology (ACC) and the American Heart Association (AHA) have partnered to create a series of clinical practice guidelines (CPGs) to help clinicians review the available evidence and make treatment decisions by building on evidence-based recommendations (2). The CPG efforts have been supplemented by resources that assist in clinical decision making, such as appropriate use criteria, which can serve as tools for informed conversations with patients about treatment options (3). Payers deciding on reimbursement suitability and whether to incentivize healthcare systems to provide high-quality care also have used such tools. Clearly, there is a societal desire to get the most complete quantitative information to patients and their providers, so that the best evidence-based decisions can be made in the quest for optimal patient outcomes.
The care of patients with coronary artery disease involves making multiple decisions about diagnosis and risk stratification, medical treatments, and revascularization options. Risk scores have been developed to aid in that decision-making process by providing tools that quantify a patient’s likelihood of having the disease, contribute insight into the risk of subsequent clinical events (such as cardiac mortality or ischemic events like myocardial infarction), estimate the likelihood of gaining incremental benefit from certain medical therapies, and aid in determining the appropriateness of 1 type of revascularization procedure over another (percutaneous coronary intervention [PCI] vs. coronary artery bypass grafting [CABG]). For those patients undergoing revascularization procedures, risk scores have been developed that provide an estimate of procedural or surgical adverse risks (4,5). All of these scores have the ultimate goal of being useful to the clinician in discussions with their patient, so decisions regarding their care can be thoughtful, evidence-based, and potentially tailored to the individual’s characteristics. This is the essence of informed clinical decision making. Recent controversy about the risk calculator released as part of the ACC/AHA Prevention Guidelines reflects the intense interest from multiple stakeholder groups in the construction of useful risk scores.
The SYNTAX score, a risk assessment score built from coronary angiographic variables, has been shown to be useful in separating patients with coronary artery disease into low-, intermediate-, and high-risk groups (6). Additionally, the SYNTAX score suggests cutoffs, indicating which patients gain more benefit (i.e., have better clinical outcomes) with 1 form of revascularization versus another. Most importantly, the analyses from SYNTAX support the notion that patients with the most complex coronary anatomy (high SYNTAX risk score) preferentially benefit from CABG more than PCI. It is this ability of the SYNTAX score to discriminate reasonably well the benefits of a therapy for a group of patients that has resulted in its inclusion in both the American and European revascularization guidelines (7,8). In this issue of the Journal, Zhang et al. (9) extend the findings from the original SYNTAX trial by reporting the results of a series of analyses comparing site investigator–reported SYNTAX scores (sSS) with core laboratory–assessed SYNTAX scores (cSS). In these analyses, the sSS performs less well than the cSS in discriminating among the 3 risk groups (low, intermediate, and high). Of note, when clinical variables are combined with the angiographic variables into a new risk score, SSII, the concordance index is similar whether using the sSS or the cSS as the angiographic piece of the risk equation. The authors conclude that appropriate training is necessary to reduce variability and improve performance in the assessment of the SS. Also, they deduce that inclusion of the clinical variables as part of the patient assessment is critical for optimal risk scoring and for guiding the revascularization strategy.
Clinical trials frequently use core laboratory assessment or central review for interpreting images and laboratory measures as well as for adjudicating clinical events, such as myocardial infarction, stroke, or bleeding (10). Analyses from clinical studies, including modeling for outcomes, typically use centrally reviewed and adjudicated data, as these data have been systematically evaluated with less bias and more attention to protocol definitions. But once these published analyses are to be used in routine clinical practice (i.e., outside the controlled environment of a clinical study), locally derived and defined data must be used in the risk scoring systems by practical necessity. In making clinical decisions, practitioners heuristically incorporate all sorts of data (clinical, imaging, laboratory, and so on) from the individual patient (11). Formal, validated risk scores can aid the clinician (and the patient) in the decision-making process, but such scores need to be statistically robust (12) while also having face validity and being easy to use.
The assessment of the SS score (sSS) at the site suffers from weakness in its discriminatory ability, although the authors note that performance will likely improve through appropriate training. Formal incorporation of the clinical variables (something the practitioner does as a matter of course) strengthens the score’s performance (c-index = 0.744). For the sSS-derived SSII to be useful at the point of care, it needs the addition of tools and applications that will facilitate its incorporation into the clinical workflow. Increasingly, clinicians want to use smartphone applications that can be readily accessed at the point of care delivery. A good example is the recently released atherosclerosis risk calculator (13).
Ultimately, the most value will be in having EHR data integrated seamlessly with specialized data on genotyping, imaging, other laboratory data, and eventually, even wearable sensor data. Risk assessment and treatment decision algorithms will be updated and recalculated continuously and fed back to providers at the point of care, so that they can integrate each patient’s values and preferences to place them in their individual social and environmental context to arrive at decisions that best fit that individual. At that point, “big data” will be more than a buzz phrase; it will have turned into useful information.
↵∗ Editorials published in the Journal of the American College of Cardiology reflect the views of the authors and do not necessarily represent the views of JACC or the American College of Cardiology.
Dr. Harrington is on the Scientific Advisory Board for Adverse Events (Santa Rosa, California).
- American College of Cardiology Foundation
- Topol E.
- ↵ACCF/AHA Task Force on Practice Guidelines. Methodology Manual and Policies From the ACCF/AHA Task Force on Practice Guidelines. American College of Cardiology Foundation and American Heart Association, Inc. 2010. Available at: http://assets.cardiosource.com/Methodology_Manual_for_ACC_AHA_Writing_Committees.pdf and http://my.americanheart.org/idc/groups/ahamah-public/@wcm/@sop/documents/downloadable/ucm_319826.pdf. Accessed June 3, 2014.
- Patel M.R.,
- Spertus J.A.,
- Brindis R.G.,
- et al.,
- American College of Cardiology Foundation
- The Society of Thoracic Surgeons. Explanation of Quality Rating, Composite Score, and Star Ratings. Available at: http://www.sts.org/quality-research-patient-safety/sts-public-reporting-online/explanation-quality-rating-composite-sco. Accessed June 3, 2014.
- Task Force Members,
- Montalescot G.,
- Sechtem U.,
- Achenbach S.,
- et al.
- Fihn S.D.,
- Gardin J.M.,
- Abrams J.,
- et al.,
- American College of Cardiology Foundation; American Heart Association Task Force on Practice Guidelines; American College of Physicians; American Association for Thoracic Surgery; Preventive Cardiovascular Nurses Association; Society for Cardiovascular Angiography and Interventions; Society of Thoracic Surgeons
- Zhang Y.J.,
- Iqbal J.,
- Campos C.M.,
- et al.
- ↵Brush JE Jr. The Science of the Art of Medicine. Available at: https://itunes.apple.com/us/book/science-art-medicine/id643948555?mt=11. Accessed June 3, 2014.
- ↵American Heart Association. 2013 Prevention Guidelines Tools: CV Risk Calculator. Available at: http://my.americanheart.org/professional/StatementsGuidelines/PreventionGuidelines/Prevention-Guidelines_UCM_457698_SubHomePage.jsp. Accessed June 3, 2014.