Estimating Longitudinal Risks and Benefits from Cardiovascular Preventive Therapies Among Medicare Patients: The Million Hearts Longitudinal ASCVD Risk Assessment Tool

In 2015 the Center for Medicare and Medicaid Innovation (CMMI) announced plans to perform a large cluster randomized payment model to test a value-based payment program designed to determine whether financially rewarding reductions in 10-year predicted risk for atherosclerotic cardiovascular disease (ASCVD) across a physician's patient population is an effective model to reduce the burden of heart attack and stroke.1 Such a program necessitates the ability to accurately predict changes in ASCVD risk expected from individual or combinations of multiple interventions.

All current risk-factor-based ASCVD risk prediction models are designed to predict the natural course of ASCVD in the absence of therapy and, with the exception of anti-hypertensive therapy, are unable to estimate the impact of therapies on ASCVD risk. In response, Lloyd-Jones et al.2 have proposed adjusting the pooled cohort equation (PCE) to predict the impact of four ASCVD prevention therapies/interventions (statin, anti-hypertensive, aspirin, and smoking cessation), individually or in combination, on an individual's ASCVD risk. The proposed model includes adjustment of baseline PCE risk score based on the initiation and an individual's response to these therapies (e.g., change in blood pressure) observed during follow-up visits. This modified PCE is introduced as The Million Hearts® Longitudinal ASCVD Risk Assessment Tool.

The authors should be commended for this first of its kind tool for estimating the impact of ASCVD preventive therapies. If accurate, it has the potential to improve risk prediction of patients "on treatment" and advance physicians' and patients' ability to understand the impact of four important risk prevention strategies on ASCVD risk. This objective assessment may result in improved balancing of the risk and benefit of these preventive therapies in individual patients. Potential improvement in patient understanding of how individual or combinations of therapies impact their personal ASCVD risk may result in improved adherence to such therapies. This new risk tool was rigorously developed by a thorough examination of the available trial data on the expected impact of individual and combinations of therapies on ASCVD risk. However, several issues should be considered.

The Million Hearts® Longitudinal ASCVD Risk Assessment Tool has not been externally validated and therefore use in clinical practice is premature. At a minimum, the accuracy of The Million Hearts Longitudinal ASCVD Risk Assessment Tool to predict future events among individuals using the preventive therapies included in the proposed model should be validated in independent cohort(s). The cohorts should be reflective of the target population (gender balanced, ethnically diverse, contemporary US patients) and include appropriate follow-up. Ideally, this novel methodology could be applied to other CVD risk scores (e.g., Framingham, Reynolds) and compared to The Million Hearts® Longitudinal ASCVD Risk Assessment Tool. Publication of the complete Million Hearts® Longitudinal ASCVD Risk Assessment Tool is necessary for external validation and efficient development of other longitudinal risk estimators from existing risk prediction scores.

The methodology used to develop The Million Hearts Longitudinal ASCVD Risk Assessment Tool is unique from traditional risk score development because it considers how risk factors change over time and are modified by treatment. The new score may be limited, however, because it assumes the baseline PCE accurately reflects ASCVD risk in the absence of the four interventions (new risk variables) –aspirin, statin, smoking cessation, and follow-up blood pressure. Performance of the PCE has been shown to overestimate risk in some cohorts and other CVD risk scores have performed better in head to head comparison.3,4

The PCE may overestimate risk because the model was derived from cohorts beginning in the 1960s and 1970s when incident ASCVD was higher and the impact of individual risk factors on incident ASCVD was likely different from today's target population.3,4 Although it is logical to assume that risk overestimation by the PCE is due to the therapies for which The Million Hearts® Longitudinal ASCVD Risk Assessment Tool is designed to account for, data suggests that these preventive therapies may only explain a portion of the overestimation.4-7 Finally, ASCVD preventive therapies may impact different individuals (e.g., those with diabetes or of different race/ethnicity) differently. This issue should be addressed with assessment of the new prediction score in multiple and/or diverse cohorts.

The relatively high treatment adherence rates achieved in randomized control trials, which are used to estimate the effect of preventive therapies in the Million Hearts® Longitudinal ASCVD Risk Assessment Tool, should not be expected in clinical practice. While follow-up risk factor changes included in the model will be reflective of adherence for some medication (statin and serum cholesterol, anti-hypertensive medications, and blood pressure), measures of response are not included for other therapies in the model (aspirin use or smoking cessation). Estimates of how adherence impacts risk would be an outstanding addition to this model if data are sufficient to do so.

Lloyd-Jones et al. correctly emphasize the importance of lifestyle interventions for all patients with respect to ASCVD risk. Unfortunately, no measure of the potential impact of lifestyle intervention was included in the model with the authors citing a lack of data to allow for an accurate quantitative estimate of how lifestyle interventions impacts ASCVD risk. It is important to consider that the Center for Medicare and Medicaid Innovation proposal to "financially reward reductions in 10-year predicted risk for ASCVD across a physician's patient population" may have many unintended consequences. Specifically, important lifestyle interventions (diet, exercise) may receive less emphasis because such efforts are not accounted for in the financial reward system. Additionally, the proposed system could result in overuse of medication based on the possibility of reward despite competing risk that is not accounted for in the current model. Overtreatment is partially addressed by Lloyd-Jones et al. via a "floor effect" which does not allow the risk prediction to fall below a minimum event rate expected from individuals with untreated optimal levels of all risk factors of the same age, sex and race, to reflect futility of treatment in low risk individuals. Harm from therapy, such as bleeding with aspirin use or incident diabetes with statin use, is not considered by the model and may be important outcomes to explore. This could be particularly important if CMS rewards predicted ASCVD risk reduction without penalties for increasing risk of harm.

The current US guideline paradigm for decision-making in the primary prevention of ASCVD is that the intensity of prevention efforts should be calibrated to the absolute ASCVD risk.8.This paradigm and utilization of an incentive program that rewards reductions of a population's calculated ASCVD risk, like that proposed by CMS, is best suited for individuals with a high short-term risk—mostly older individuals. However, ASCVD develops over a lifetime and is modifiable over a lifetime. Young individuals with several risk factors have significantly more risk than their age matched peers (high lifetime relative risk) but low 10-year absolute risk. The low absolute 10-year risk would be reflected in the Million Hearts® Longitudinal ASCVD Risk Assessment Tool potentially providing physicians and patients with a false sense of wellness leading to missed opportunities to negate this risk. The 2016 European Prevention Guidelines include relative risk reduction to guide individual patient therapeutic strategies.9 The proposed risk tool could be modified to include relative risk –a concept that deserves further investigation.

The Million Hearts® Longitudinal ASCVD Risk Assessment Tool introduces a novel approach to ASCVD risk prediction. This approach models the expected impact of four important ASCVD preventive therapies in individual patients, offering the potential to improve the efficacy of utilization of these therapies by care providers and patients alike. As is true of all new diagnostics, care must be taken to evaluate the accuracy of this new tool in independent cohorts, the impact it has on care provider and patient behavior and how this will impact ASCVD events. The proposed utilization of The Million Hearts Longitudinal ASCVD Risk Assessment Tool for CMS based financial reward programs must be carefully evaluated for both intended and unintended consequences.


  1. Shanghavi DM, Conway PH. Paying for prevention: a novel test of Mediare value-based payment for cardiovascular risk reduction. JAMA 2015; 314:123-4.
  2. Lloyd-Jones DM, Sanghavi DM, Wright J, et al. Estimating longitudinal risks and benefits from cardiovascular preventive therapies among Medicare patients: the Million Hearts® Longitudinal ASCVD Risk Assessment Tool. J Am Coll Cardiol 2016.
  3. Cook NR, Ridker PM. Calibration of the pooled cohort equations for atherosclerotic cardiovascular disease: an update. Ann Intern Med 2016. [Epub ahead of print]
  4. DeFilippis AP, Young R, Carrubba CJ, et al. An analysis of calibration and discrimination among multiple cardiovascular risk scores in a modern multiethnic cohort. Ann Intern Med 2015;162:266-75.
  5. Cook NR, Ridker PM. Further insight into the cardiovascular risk calculator: the roles of statins, revascularizations, and underascertainment in the Women's Health Study. JAMA Intern Med 2014;174:1964-71.
  6. DeFilippis AP, Young R, McEvoy JW, et al. Risk score overestimation: the impact of individual cardiovascular risk factors and preventive therapies on the performance of the American Heart Association-American College of Cardiology-Atherosclerotic Cardiovascular Disease risk score in a modern multi-ethnic cohort. Eur Heart J 2016. [Epub ahead of print]
  7. Muntner P, Colantonio LD, Cushman M, et al. Validation of the atherosclerotic cardiovascular disease pooled cohort risk equations. JAMA 2014;311:1406-15.
  8. Goff DC, Lloyd-Jones DM, Bennett G, et al. 2013 ACC/AHA guideline on the assessment of cardiovascular risk: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines. J Am Coll Cardiol 2014;63:2935-59.
  9. Piepoli MF, Hoes AW, Agewall S, et al. 2016 European Guidelines on cardiovascular disease prevention in clinical practice: The Sixth Joint Task Force of the European Society of Cardiology and Other Societies on Cardiovascular Disease Prevention in Clinical Practice (constituted by representatives of 10 societies and by invited experts)Developed with the special contribution of the European Association for Cardiovascular Prevention & Rehabilitation (EACPR). Eur Heart J 2016;37:2315-81.

Keywords: Antihypertensive Agents, Aspirin, Atherosclerosis, Blood Pressure, Cholesterol, Diabetes Mellitus, Diet, Hydroxymethylglutaryl-CoA Reductase Inhibitors, Life Style, Medical Futility, Myocardial Infarction, Primary Prevention, Risk Assessment, Risk Factors, Risk Reduction Behavior, Smoking Cessation, Stroke

< Back to Listings