Researchers at UT Southwestern in Dallas are using machine learning to predict which patients will benefit most from intensive high blood pressure treatment.
The decision tree algorithm they developed combines three simple variables routinely collected during clinic visits to identify adults with hypertension who are at the highest risk for early major adverse cardiovascular events, such as death, heart attack or stroke.
Patient data from two large clinical trials that tested intensive vs. standard blood pressure lowering—Systolic Blood Pressure Intervention Trial (SPRINT) and Action to Control Cardiovascular Risk in Diabetes (ACCORD)—were leveraged to identify patients for whom the benefits of intensive therapy outweigh the risks.
While long-term intensive high blood pressure drug therapy can reduce the risk of heart failure and death, it carries an increased risk of side effects. However, what researchers found is that the risk prediction model consisting of age, urinary albumin-creatinine ratio (UACR) and clinical cardiovascular disease history successfully identified some hypertensive patients who have a better chance of deriving benefits from aggressive blood pressure lowering.
The machine learning method determined that three simple criteria—an age of 74 or older, a UACR of 34 or higher and a history of clinical cardiovascular disease—predicted those patients among a high-risk group who were more likely to benefit from intensive blood pressure-lowering treatment, while those patients younger than age 74 who had a UACR less than 34 and no history of cardiovascular disease may do equally as well with less aggressive treatment.
“Large randomized trials have provided inconsistent evidence regarding the benefit of intensive blood pressure lowering in hypertensive patients,” says Yang Xie, director of UT Southwestern’s Quantitative Biomedical Research Center and the Bioinformatics Core Facility. “To the best of our knowledge, this is the first study to identify a subgroup of patients who derive a higher net benefit from intensive blood pressure treatment.”
Researchers published the results of their study in the July 15 issue of the American Journal of Cardiology.
“We feel that our findings have major clinical implications, since in addition to its predictive effects, the model generated here is simple and easy to implement in clinical practice without additional lab tests or computational tools,” adds Xie, who is also an associate professor of clinical sciences and bioinformatics at UT Southwestern. “We hope that clinicians can someday use this algorithm to identify which patients should receive standard vs. intensive treatment, and we hope to design a prospective clinical trial to further validate this algorithm.”
Register or login for access to this item and much more
All Health Data Management content is archived after seven days.
Community members receive:
- All recent and archived articles
- Conference offers and updates
- A full menu of enewsletter options
- Web seminars, white papers, ebooks
Already have an account? Log In
Don't have an account? Register for Free Unlimited Access