Mount Sinai researchers have developed a groundbreaking AI model that provides individualized treatment recommendations for patients with atrial fibrillation (AF), helping clinicians decide whether anticoagulant therapy—commonly known as blood thinners—is necessary to prevent stroke. This innovation marks a significant shift from current population-based guidelines to a personalized, data-driven approach. AF, the most common heart rhythm disorder, affects around 59 million people worldwide. In AF, the heart’s upper chambers quiver instead of contracting properly, increasing the risk of blood clots forming and traveling to the brain, causing a stroke. While blood thinners are standard treatment to reduce this risk, they also carry a risk of major bleeding, making the decision to treat complex. The new AI model analyzes a patient’s complete electronic health record—including 21 million doctor visits, 82 million clinical notes, and 1.2 billion data points from 1.8 million patients—to calculate individualized risks of both stroke and major bleeding. Unlike traditional risk scores that offer population-level estimates, this model computes net benefit at the individual level, weighing the potential harm of treatment against its protective benefits. In validation studies, the model was tested on 38,642 AF patients within the Mount Sinai Health System and externally on 12,817 patients from public datasets at Stanford. The results showed that the model recommended against anticoagulant treatment for up to half of the patients who would have received it under current guidelines. These patients were reclassified as low risk for stroke and high risk for bleeding, suggesting that avoiding blood thinners could prevent unnecessary harm. The researchers emphasize that this approach moves beyond the limitations of one-size-fits-all risk scores. By providing clinicians with personalized probabilities for stroke and bleeding, the model reduces cognitive burden, supports shared decision-making, and enables more accurate patient counseling. “This study represents a profound modernization of how we manage anticoagulation for patients with atrial fibrillation and may change the paradigm of how clinical decisions are made,” said Dr. Joshua Lampert, Director of Machine Learning at Mount Sinai Fuster Heart Hospital. Dr. Girish Nadkarni, Chair of the Windreich Department of Artificial Intelligence and Human Health at the Icahn School of Medicine at Mount Sinai, added that the model synthesizes vast amounts of data to deliver precision medicine tailored to each patient’s unique medical history. Dr. Vivek Reddy, Director of Cardiac Electrophysiology at Mount Sinai Fuster Heart Hospital, noted that preventing stroke is the primary goal in AF management, with the condition affecting an estimated one in three adults over their lifetime. The team believes that if future clinical trials confirm the model’s effectiveness, it could dramatically improve patient outcomes by reducing both strokes and bleeding events. “Patients often ask, ‘What does this mean for me?’” said co-first author Justin Kauffman, a data scientist at Mount Sinai. “Our system answers that by showing not just what might happen, but how likely it is to happen to them personally—using their full medical history to guide care.”