Predictive analytics helps Fresenius anticipate dialysis complications
Hemodialysis is a life-saving treatment for those suffering from kidney failure. The procedure, often called kidney dialysis, cleansing a patient’s blood, substituting for the function of the kidneys, and is not without risk, however.
Hemodialysis is a life-saving treatment for those suffering from kidney failure. The procedure, often called kidney dialysis, cleansing a patient’s blood, substituting for the function of the kidneys, and is not without risk, however. German healthcare company Fresenius Medical Care, which specializes in providing kidney dialysis services, is using a combination of near real-time IoT data and clinical data to predict one of the most common complications of the procedure.
Fresenius operates a network of more than 4,000 outpatient dialysis centers globally, primarily treating patients with end-stage renal disease (ESRD), which requires those patients to receive dialysis three times a week for the rest of their lives. About 10% of hemodialysis treatments result in intradialytic hypotension (IDH) — low blood pressure.
“IDH holds a potentially severe immediate risk for patients during dialysis and therefore requires immediate attention from staff,” says Hanjie Zhang, director of computational statistics and artificial intelligence at the Renal Research Institute, a joint venture of Fresenius North America and Beth Israel Medical Center. “As such, IDH not only reduces patients’ quality of life and is associated with morbidity and mortality, but also results in lower clinical efficiency and effectiveness.”
Dr. Peter Kotanko, research director of the Renal Research Institute, adds, “Whenever a patient’s blood pressure drops and IDH ensues, healthcare providers need to intervene, and the operations of a clinic can be disrupted.”
In September 2021, Fresenius set out to use machine learning and cloud computing to develop a model that could predict IDH 15 to 75 minutes in advance, enabling personalized care of patients with proactive intervention at the point of care. To do so, the team had to overcome three major challenges: scalability, quality and proactive monitoring, and accuracy. The project, dubbed Real-Time Prediction of Intradialytic Hypotension Using Machine Learning and Cloud Computing Infrastructure, has earned Fresenius Medical Care a 2023 CIO 100 Award in IT Excellence.
Putting data to work to improve health outcomes
“Predicting IDH in hemodialysis patients is challenging due to the numerous patient- and treatment-related factors that affect IDH risk,” says Pete Waguespack, director of data and analytics architecture and engineering for Fresenius Medical Care North America. “Clinically, prediction is more useful if it predicts an IDH event for a given patient during an ongoing dialysis treatment. It was imperative to bring a cross-functional team of clinical, operational, and technology experts together to define the needs of the near real-time prediction and response.”
About Author
