Ochsner Medical Center is using an innovative approach to improve patient safety, care delivery and health outcomes. As one example, the New Orleans health system applies machine learning to reduce health care-acquired Clostridioides difficile infections, creating a neural network model that was trained by using 250,000 hospital admissions over three years. That work revealed a surprise: The use of gastric acid-suppression medications was known to be associated with increased risk of C. difficile infection, but not to the extent seen in other Ochsner patients
“Those medications were by far the most impactful and principal predictor for our patients,” said Armin Schubert, M.D., vice president of medical affairs, quality and patient safety at Ochsner. “This did not correspond to what had previously been reported in the literature, which shows that you really have to study your own population and devise an action plan based on that.”
At Ochsner, patients at high risk for C. difficile infection who are identified with machine learning are reviewed by a dedicated pharmacist who contacts the attending physician for risk review and medication. The pharmacist advises physicians on the minimal use of acid-suppressing medications and, if necessary, discontinuation of use.
The result: Monthly health care-associated C. difficile infections fell by 49%, avoiding 166 infections and an estimated $4 million in treatment costs over a two-year period.
Read more about Ochsner’s work on models that help predict patients that are at a higher risk for health care-associated infections.