AI-Powered Alerts Reduce Kidney Complications After Heart Surgery

Acute kidney injury (AKI) is one of the most serious complications following heart surgery, raising mortality fivefold and tripling hospital costs. Diagnosis currently depends on declines in urine output or rises in creatinine — markers that appear only after the ideal treatment window has passed. Now, researchers are developing an artificial intelligence (AI) system that analyzes real-time clinical data to detect AKI early, giving clinicians critical time to intervene.

This four-year project combines the expertise of researchers in machine learning and statistics at Rice University (Houston, TX, USA) with the extensive clinical dataset of more than 9,000 cardiac surgery patients and 68 million data points at Baylor College of Medicine (Houston, TX, USA). The system uses ensemble machine-learning models trained on minute-by-minute electronic medical record (EMR) data — including vital signs, lab results, medications, and fluid balance — to identify early signals of kidney stress.

The aim is to detect AKI up to 24 hours before standard diagnostic markers appear, recommend personalized interventions, and provide clinicians with interpretable insights via symbolic regression–derived “digital biomarkers” and a simple bedside scoring system. The platform will operate in a secure clinical deployment environment that streams EMR data every 15 minutes to generate rolling risk profiles and suggested actions.

The study includes prospective ICU validation to assess accuracy, alignment between clinicians’ decisions and model recommendations, and the system’s impact on AKI incidence. The approach emphasizes interpretability to support trust and adoption, ensuring clinicians can understand which factors drive predictions and how specific actions may reduce risk. By translating complex real-world data into actionable guidance, the system aims to improve outcomes after heart surgery and establish a blueprint for trustworthy clinical AI deployment.

“Early prediction would enable targeted interventions that can improve outcomes, but prior risk tools are static and have limited value in the dynamic post-operative environment,” said Meng Li, site principal investigator for the project. “Our central hypothesis is that dynamic machine-learning models can accurately predict AKI in real time from routinely collected EMR data and augment clinical decision-making by quantifying risk reduction of therapeutic interventions.”

“Presently, AKI is identified via clinical parameters, but these represent late findings often manifesting after the ideal treatment window,” said Dr. Ravi Ghanta, principal investigator for the project. “The electronic medical record provides complex, multidimensional data that clinicians incorporate into decision-making, but which remains underutilized in clinical decision support. We aim to change that.”

Related Links:
Rice University
Baylor College of Medicine