Taiwan Hospital uses AI to detect heart failure risk in kidney patients

One in 10 people on earth suffer from chronic kidney disease. Millions without access to affordable treatment die from the disease each year. Fortunately, many have been able to turn to dialysis.

In Taiwan, about 85,000 people undergo dialysis. The island also happens to have the world’s highest prevalence of such cases based on population density.

Such treatment is not without risk with cardiovascular disease being the leading cause of death for dialysis patients.

To improve the odds of survival for those under dialysis, Taipei Veterans General Hospital (TVGH) has turned to an AI model that predicts heart failure risk in real time during dialysis procedures.

The hospital’s AI tool displays key factors for risk prediction on a dashboard for clinicians, detects abnormal patterns in the streaming data from dialysis machines, and alerts doctors and nurses to intervene immediately.

Early detection can save lives

“In this field, early detection and prompt decision-making can save lives. By deploying NVIDIA Jetson next to each dialyser to perform AI prediction during the procedure, we can achieve real-time insights in a way that’s affordable and effective, even for small-scale dialysis centres,” said Professor Der-Cherng Tarng, Chief of the Department of Medicine at TVGH.

The hospital deploys NVIDIA AI technology, including the NVIDIA Jetson edge AI platform, to analyse patient data in real time. The AI model uses a combination of dialysis machine data, patient medical records, test results and medication information to make predictions with up to 90 percent accuracy.

This dashboard displays the health status of all dialysis patients, showing the patient’s severity and risk category in different colours. On display are each patient’s real-time stream of dialysis machine data and the AI model’s assessment of whether or not the patient’s iron levels are normal.

The AI tool was built to identify abnormal patterns in a patient’s data using multiple AI algorithms including decision trees, gradient boosting and convolutional neural networks. It was trained on a dataset of three million health records.

The team recently added additional predictive indicators to the tool, including hemoglobin level and chest X-ray image analysis.

On top of predicting heart failure risk, the hospital’s AI model has reduced the deviation rate in clinicians’ assessment of a patient’s dry weight by 80 percent, an accuracy boost that helps lessen the risk of complications.

Having experienced success with AI in heart failure prediction, TVGH is working on other AI projects accelerated with the NVIDIA Parabricks genomics software, the NVIDIA Flare federated learning workflow and the NeMo Megatron framework for natural language processing.