Using AI to Predict Who Gets Readmitted After Heart Surgery

AI may soon help doctors predict who will return to the hospital after coronary artery bypass surgery.

This heart surgery helps restore blood flow to the heart by bypassing blockages in the arteries. More than 12% of people who get bypass surgery are readmitted to the hospital within 30 days. To understand why, a research team led by Rameshbabu Manyam, PhD, assistant research professor of biostatistics at Emory University’s Rollins School of Public Health, developed a machine learning model to predict who was at the highest risk of being readmitted and why. 

The team recently published their results in Journal of Thoracic and Cardiovascular Surgery Open. 

Key Findings

The model identified key factors tied to the risk of readmission after bypass surgery:

• Body mass index
• Kidney function
• Liver health
• Levels of white blood cells, platelets, hemoglobin, and hematocrit in the blood

These factors also interact with each other. That means that risk is not always as straightforward as one measurement being high or low, and a patient’s risk of readmission can depend on the relationship between multiple factors.

Why This Matters

This AI model could support future clinical decision-making tools that help doctors identify who is at a higher risk of readmission after surgery. It could also provide insight into why their risk is higher.
If doctors can identify these patients early, they can provide more targeted care. That could include addressing some risk factors before surgery or more closely monitoring and following up with high-risk patients after surgery. This could help reduce readmissions.

What the Experts Say

“What makes this work particularly interesting is the potential to look beyond predicting who may be readmitted and understand the factors behind that risk,” says Manyam. “For patients and their families, this could ultimately mean care better tailored to their individual needs. When we understand the reasons behind an AI prediction, we can evaluate it more thoughtfully and use that information to provide better care.”