Providers and Google Team Up to Use EHR Data in Machine Learning
Google has applied its data collection and analytics to innovations as varied as self-driving cars and virtual home assistants—so it was just a matter of time before they brought their predictive prowess to healthcare.
Google recently announced a partnership with University of Chicago Medicine to use information collected in electronic health records (EHRs) to predict and prevent certain risk factors. Google has established partnerships like this already with Stanford Medicine and the University of California San Francisco.
The Chicago Tribune reported that university researchers are aiming to use machine learning to predict whether a patient will need to be hospitalized, how long their admission might last, and determine whether their health is deteriorating. Data—including a patient’s vital signs, medications, and physician notes—can be included in an algorithm to look for patterns suggesting action should be taken.
Researchers at the University of Chicago have already started using an algorithm called eCART to help predict risk for cardiac arrest and other high-risk conditions. Nurses will do extra checks on patients whose records indicate they are high risk. The Tribune reports that it’s too early to know if Google will develop a product for this kind of data, but eCART has already been commercialized and sold to other healthcare provider networks.