PREVENT equations accurately predict heart disease in diverse Asian and Hispanic groups

Cardiovascular disease (CVD) risk prediction models have improved the ability to stratify adults across the CVD risk spectrum. Researchers at Sutter Health and colleagues at Stanford University tested the performance of the American Heart Association's Predicting Risk of CVD Events (PREVENT) equations in the six largest Asian subgroups as well as in Mexican and Puerto Rican Hispanic subgroups.

The findings, published June 25 in JAMA Cardiology, showed the PREVENT equations accurately predicted CVD, atherosclerotic CVD (ASCVD) and heart failure events across broad, self-identified Asian and Hispanic patient populations.

The burden of heart disease, one of the biggest health problems in the U.S., is likely to grow as the population gets older. Making sure risk tools work for all our patients is key to giving the best care possible.

Advancing cardiovascular health for all

"We led this study to uncover findings that may inform better cardiovascular healthcare for all. In the original validation, PREVENT equations demonstrated good discernment and calibration among racial and ethnic groups, but the model performance among Asian and Hispanic subgroups had not been previously described," said Xiaowei (Sherry) Yan, Ph.D., senior scientist at Sutter Health's Center for Health Systems Research and first author of the study. "Our findings showed the PREVENT models performed well across multiple subgroups, with modest variation across individual Asian and Hispanic populations."

Study overview

Dr. Yan and colleagues conducted an electronic health record–based retrospective cohort study of 361,778 primary care patients aged 30 to 79 years across Sutter Health - a large integrated health system in Northern California caring for more than 3.5 million patients across urban, suburban and rural communities - from January 2010 to September 2023. Patients who had at least two primary care visits during the study period were eligible for the study. Among patients who met the inclusion criteria, the mean age was 54.6 years; 53% were female; 22% were non-Hispanic Asian and 11% were Hispanic.

Key findings

Results showed the PREVENT equations appropriately predicted risk in contemporary diverse Asian and Hispanic subgroups with modest variation in performance across the subgroups studied.

• Over a mean follow-up of 8.1 years, there were 22,648 (6.3%) CVD events among study subjects.
• The predictive accuracy, as demonstrated through use of statistical measures such as the C-statistic and calibration slope, was high for the outcome of total CVD among the Asian and Hispanic populations.
• When examined among the subgroups studied, predictive accuracy for total CVD remained high.
• There were small differences in the performance of atherosclerotic CVD and heart failure PREVENT equations among racial and ethnic groups and subgroups.

The path forward

The burden of CVD and its risk factors is projected to increase across the U.S. in the coming decades, so the ability to more accurately estimate CVD risk in all the communities we serve is more critical now than ever. Although best practices for clinical implementation of the PREVENT CVD risk prediction models should be further investigated, we believe they are an important step forward for communities that have been historically underrepresented in CVD research."

Powell Jose, M.D., FACC, Sutter Medical Group cardiologist and section chief of general cardiology for Sutter's heart and vascular service line

Dr. Yan, Dr. Jose and colleagues shared study limitations: they were unable to fully examine all Asian and Hispanic subgroups. Moreover, comparisons of predictive utility of PREVENT and pooled cohort equations were limited by small sample sizes in the Mexican and Puerto Rican Hispanic subgroups.

To better inform clinical application of the study findings, future work may focus on providing clinically sensible cutoffs based on PREVENT ASCVD risk to guide treatment decisions.