Photo credit: www.sciencedaily.com
Revolutionizing Hypertension Detection Through Electronic Health Records
A recent investigation by researchers at Mass General Brigham reveals that valuable information about hypertension may be hidden within electronic health records (EHR). Utilizing natural language processing, a branch of artificial intelligence, the team successfully identified patients whose heart ultrasounds indicated thickening of the heart muscle, a common result of high blood pressure. After physicians were informed of these findings, their likelihood of diagnosing hypertension and prescribing treatments significantly increased, nearly quadrupling in comparison to previous rates. This study emphasizes the potential of automated, innovative strategies to leverage existing health data for improved management of heart conditions. The findings are published in JAMA Cardiology and were also presented during the 2025 American College of Cardiology’s Annual Scientific Session & Expo.
“Hypertension is often referred to as the silent killer because individuals may experience elevated blood pressure without any visible symptoms,” stated Jason H. Wasfy, MD, MPhil, senior author and a cardiologist at Massachusetts General Hospital (MGH). As part of the Mongan Institute at MGH, he emphasized the importance of early detection in preventing serious cardiovascular damage over time.
In the U.S., it is estimated that nearly half of those with hypertension remain unaware of their condition or do not receive appropriate treatment.
“Routine clinical encounters generate vast amounts of information, and often there are subtle indicators within this data that suggest a patient might have hypertension. Due to the extensive nature of medical records, it’s challenging for clinicians to catch these nuances. Our study sought to explore how we could uncover these overlooked details to enhance patient care,” explained lead author Adam Berman, MD, MPH. Berman, previously affiliated with Brigham and Women’s Hospital, is currently an assistant professor at NYU Grossman School of Medicine.
The research involved developing a natural language processing algorithm that examined echocardiogram data to identify instances of left ventricular hypertrophy, a thickening of the heart muscle linked to hypertension. The algorithm flagged 648 patients within the Mass General Brigham system who had not been previously diagnosed with heart issues or treated for high blood pressure. The average age of these patients was 59, with women making up 38% of the group. Half of these patients were included in the intervention, where a health coordinator informed their doctors of the findings and facilitated additional testing or specialist consultations. The remaining patients continued to receive standard care without this intervention.
Results showed that the intervention group had a nearly fourfold increase in hypertension diagnoses (15.6% compared to 4.0%) and were more likely to receive antihypertensive medication (16.3% versus 5.0%) compared to those in the control group. Follow-up appointments with primary care physicians remained consistent across both groups, and feedback from clinicians was predominantly positive, with 72% expressing favorable views towards the notifications they received.
“Our team was committed to ensuring the intervention would be beneficial for both physicians and patients,” noted Wasfy. “Clinicians often face alert fatigue and burnout, so we designed our notification approach to be more personal.”
Future research will investigate enhancing the notification process to ensure broader integration and applicability across various healthcare settings while maintaining its effectiveness.
“Our aim is to supplement conventional care by tapping into the wealth of data already available,” Berman stated. “These patients have already undergone testing, leaving their information underutilized in a digital database. Our study illustrates how these data can be harnessed to enhance healthcare delivery and improve patient outcomes.”
Authorship for this study also includes contributions from Michael K. Hidrue, Curtis Ginder, Linnea Shirkey, Japneet Kwatra, Anna C. O’Kelly, Sean P. Murphy, Jennifer M. Searl Como, Yee-Ping Sun, William T. Curry, Marcela G. del Carmen, Ron Blankstein, David A. Morrow, Benjamin M. Scirica, Niteesh K. Choudhry, and James L. Januzzi, alongside John A. Dodson and Danielle Daly.
For further details regarding authors’ disclosures, please refer to the paper published in JAMA Cardiology.
The research intervention was funded by the Massachusetts General Physicians Organization to support innovative cardiovascular care delivery.
Source
www.sciencedaily.com