Julia Eisen

In a new study led by Yale School of Medicine’s Cardiovascular Data Science (CarDS) Lab, researchers developed an artificial intelligence (AI) tool that can identify individuals at high risk of developing heart failure using electrocardiogram (ECG) images. The new tool enables earlier identification of heart failure, potentially reducing hospitalizations and premature death, the researchers said.

ECGs are noninvasive tests that measure the heart's electrical activity through electrodes placed on the skin. Since the tests are routinely performed and widely available, they provide an ideal platform for broader heart failure screening. Heart failure is a common cardiovascular disorder affecting millions of people worldwide.

Currently, identifying individuals at high risk for heart failure often relies on a series of clinical evaluations, including extensive clinical history, physical examination, and blood testing that may not be accessible in all settings, explains first author 

, a postdoctoral fellow in the CarDS Lab.

The AI-based tool represents a paradigm shift in heart failure risk stratification, he said.

“Using an image or photograph of a 12-lead ECG as its input, our model was able to accurately stratify heart failure risk across diverse populations in the United States, United Kingdom, and Brazil,” Dhingra said. “We can now predict who is at risk of developing heart failure in the future, well before they show overt symptoms.”

, assistant professor of medicine (cardiovascular medicine) and director of the CarDS Lab, emphasized the potential public health impact of the work.

“Every time a clinician obtains an ECG—a test that is already part of standard clinical care—our simple tool now offers an opportunity for screening and risk stratification for cardiovascular disease,” he said. “The broad availability of ECG technology, even where resources are limited, enables early intervention and improved outcomes for patients who might otherwise go undiagnosed.”

As part of CarDS Lab’s global focus, the AI model was validated across multiple international populations, showcasing its potential for large-scale adoption.

“We want to ensure broad and equitable implementation of AI-based health technologies in everyday practice,” Khera said. “That’s our next frontier.”

This study was funded by grants from the National Heart, Lung, and Blood Institute, the National Institute on Aging of the National Institutes of Health (NIH), as well as by the Doris Duke Charitable Foundation.

Heart failure risk stratification using artificial intelligence applied to electrocardiogram images: a multinational study

The Department of Internal Medicine at Yale School of Medicine is among the nation's premier departments, bringing together an elite cadre of clinicians, investigators, educators, and staff in one of the world's top medical schools. To learn more, visit

Dhingra LS, Aminorroaya A, Sangha V, Pedroso AF, Asselbergs FW, Brant LCC, Barreto SM, Ribeiro ALP, Krumholz HM, Oikonomou EK, Khera R

. Heart failure risk stratification using artificial intelligence applied to electrocardiogram images: a multinational study. Eur Heart J. 2025 Jan 13:ehae914. doi: 10.1093/eurheartj/ehae914. Epub ahead of print. PMID: 39804243.

Articles

Researchers at Yale School of Medicine's CarDS Lab have developed an AI tool that uses ECG images to identify individuals at high risk of heart failure. This tool allows for earlier detection, potentially reducing hospitalizations and premature deaths. The AI model was validated across diverse populations in the US, UK, and Brazil, highlighting its potential for widespread adoption. The study emphasizes the public health impact of integrating this tool into routine ECG tests, enabling early intervention and improved outcomes, even in resource-limited settings.

A new tool may enable earlier identification of heart failure, potentially reducing hospitalizations and premature death.

New AI Tool Identifies Risk of Future Heart Failure

Approximately 90,000 people receive hematopoietic stem cell transplants worldwide annually. These patients are immunocompromised and highly vulnerable to infections, including with multi drug-resistant organisms. Because penicillin and related beta-lactam antibiotics are used as preferred therapy for many infections, patients who are labeled as having a penicillin allergy are at risk for inferior treatment and related clinical consequences.

The majority of patients who report penicillin allergy can actually tolerate penicillin, though the labeling of “penicillin allergy” leads to the use of alternative antibiotics, which are often more broad spectrum, less effective, and more toxic. In turn, penicillin allergy labels are associated with more severe infections, multi drug-resistant infections, and increased mortality.

Belmont et al. authored a comprehensive review, advocating for a multidisciplinary approach to evaluating penicillin allergy labels and providing evidence-based guidance on how to evaluate them safely.

Clinical decision tools exist to aid the clinician in assessing a penicillin allergy and can be used to risk stratify patients in order to determine a testing strategy. Patients with low-risk histories can safely undergo direct oral challenges while patients with high-risk reaction histories should undergo skin testing and oral challenge.

“Most patients who report a penicillin allergy can actually tolerate penicillins and other beta-lactam antibiotics. Although penicillin allergy labels are largely erroneous, patients with penicillin allergy labels are more likely to receive alternative antibiotics, suffer from more severe and multi-drug-resistant infections, and have increased healthcare utilization and cost. There is an urgent need to design interventions to expand access to evaluation, which can immediately restore access to preferred narrow spectrum antibiotics and improve outcomes,” said Belmont.

A call to address penicillin allergy labels in patients with hematopoietic stem cell transplants: How to avoid rash decisions

 Trubiano JA. A call to address penicillin allergy labels in patients with hematopoietic stem cell transplants: How to avoid rash decisions. Transpl Infect Dis. 2024 Aug 5:e14350. 

Hematopoietic stem cell transplant recipients, numbering around 90,000 globally each year, are highly susceptible to infections, particularly from multi-drug-resistant organisms. Many patients inaccurately labeled with penicillin allergies face suboptimal treatment due to the use of alternative antibiotics, which are often broader spectrum, less effective, and more toxic. This mislabeling is linked to severe infections and increased mortality. A multidisciplinary approach is recommended to reassess penicillin allergy labels, utilizing clinical decision tools for risk stratification and appropriate testing, ultimately improving patient outcomes and antibiotic stewardship.