Using basic information like age, gender and clinical history, Weill Cornell Medical College researchers have developed a simple method to more accurately predict whether a stable patient is likely to suffer from coronary artery disease or die of a heart attack in the next three years.
Cardiologists often use algorithms to determine whether or not patients should undergo invasive testing. With this easy-to-use, accurate method to determine risk, they can intervene when patients need it, skip invasive testing if they don't, and ultimately save time and money.
"We do 10 million stress tests a year in the United States, and so many people don't need them and don't have the disease," said lead author Dr. James K. Min, the director of the Dalio Institute of Cardiovascular Imaging at NewYork-Presbyterian Hospital and Weill Cornell Medical College and a professor of radiology at Weill Cornell. "We're wasting a lot of money, and wasting it on the wrong people. This method will allow us to better define who we need to evaluate."
While other algorithms to predict risk are already in use, Dr. Min said that the status quo "severely overestimates the probability of disease by almost three-fold." In a study published April 10 in the American Journal of Medicine, Dr. Min's team set out to create an updated, contemporary method to reach a risk score that quantifies the probability of disease, and also the likelihood that a patient will die of a heart attack, a figure that previously didn't exist.
To reach these figures, Dr. Min and his team followed 14,004 adults who were symptomatically stable but had suspected coronary artery disease for periods of 1.6 to five years between 2004 and 2011. They collected patient data, and with it, used advanced statistical methods to develop the algorithm, in which doctors input simple digits correlated to clinical information. For example: If a patient is 60 years old they get a six; if they're 50 years old they get a five; if they have diabetes they get a one; if they don't have diabetes they get a zero. When all of the data is entered, the algorithm computes an integer-based score, which corresponds to a percent likelihood of having disease and a percent likelihood of having a heart attack in the next three years. If a patient's likelihood for either is greater than 30 percent, their clinician will likely send them in for more testing.
"Our intention was to create something really, really easy to use," Dr. Min said. "In the clinic when doctors are seeing a patient, it's not hard to apply this score because it's just so foolproof."