November 19, 2013

Bad math? New statin calculator may be too flawed to use, experts say

Daily Briefing

    Experts say an online calculator intended to help doctors assess risks and treatment options under new cholesterol guidelines vastly overestimates the number of people who are candidates for statin drugs.

    The new guidelines from the American Heart Association (AHA) and the American College of Cardiology (ACA) recommend using four specific risk factors to identify patients who should receive statins and patients who should simply make lifestyle changes.

    One of those risk factors is a patient's 10-year risk of heart attack and stroke; patients whose risk exceeds 7.5% should be prescribed a statin, per the new guidelines. Doctors are advised to use an online risk calculator—which factors in blood pressure, age, and total cholesterol levels, among other things—to determine that risk.

    Calculator problems surface

    Problems with the calculator were identified by Harvard Medical School professors Paul Ridker and Nancy Cook, who evaluated the material over a year ago when the NIH's National Heart, Lung, and Blood Institute—which helped to develop the guidelines—provided a draft for their review.

    Ridker and Cook evaluated the calculator using three large studies involving thousands of participants. They were aware of the subjects' ages, whether they smoked, their cholesterol levels, and their blood pressures, as well as how many had heart attacks over the 10-year study period.

    Ridker and Cook found that the calculator overestimated risk by between 75% and 150%—depending on the patient population. Although they reported their findings to the NIH institute, the problems remained in the final guidelines released last week. 

    "Miscalibration to this extent should be reconciled and addressed before these new prediction models are widely implemented," Ridker and Cook wrote this week in The Lancet, adding, "If real, such systematic overestimation of risk will lead to considerable overprescription.

    AHA, ACC respond to study

    According to the Times, the issue created turmoil at the AHA's annual meeting in Dallas this weekend. On Saturday night, the AHA and ACC held an emergency closed-door meeting with Ridker, who presented his data and highlighted the issue.

    Sidney Smith, the executive chair of the guidelines committee, said the concerns raised by Ridker and Cook "merit attention" and pledged to look into the issue to see if "substantive change" is necessary.

    However, Smith noted that the calculator is only one determiner of treatment and that the new guidelines represent the "best efforts of people who have been working for five years" on the issue. For now, he said, "We intend to move forward with the implementation of these guidelines," adding, "It’s really important not to raise unnecessary alarm.”

    Doctors express anger, concern

    Steven Nissen, chief of cardiovascular medicine at the Cleveland Clinic and a former president of the ACC, called for an immediate halt to the implementation of the new guidelines. "Something is terribly wrong" when "your average healthy Joe gets treated [and] virtually every African-American man over 65 gets treated," he said.

    Other physicians voiced concern that the error could shake the public's trust in new heart guidelines. "We're surrounded by a real disaster in credibility," says Peter Libby, chair of cardiovascular medicine at Brigham and Women's Hospital (Kolata, New York Times, 11/17; Armstrong/Cortez, Bloomberg, 11/18; Healy, "Science Now," Los Angeles Times, 11/18).

    The Advisory Board's take

    Megan Tooley,  Consultant

    It's no surprise how much attention this story is receiving, given how widespread the implications are for these guidelines on the U.S. population. Furthermore, it hits on topics that have been top-of-mind for CV administrative and physician leaders alike: scrutiny over the appropriate use of medical services, the pressure to increase CV disease prevention efforts, and the challenges of adopting evidence-based practice.

    More from the Cardiovascular Rountable

    These events also highlight the fine line between earlier intervention and prevention of disease, and delivering potentially unnecessary services. Heart and vascular programs must strike a balance between ensuring at-risk patients receive necessary preventive measures—such as statins—while avoiding the cost and quality implications of applying these strategies to patients not likely to benefit.

    It will be interesting to see how the societies remedy this situation in the coming days, but irrespective of if and how they change the calculator, a key takeaway from this incident will remain the same: risk tools are just that—tools. While they can guide physicians in evaluating the options, risk tools such as these should not supplant physician decision-making, which should also reflect patient input and preferences.

    Ultimately, physicians still must have the final word on whether or not their patients need these drugs—regardless of what a risk calculator says.

    More from today's Daily Briefing
    1. Current ArticleBad math? New statin calculator may be too flawed to use, experts say

    X
    Cookies help us improve your website experience. By using our website, you agree to our use of cookies.