AI outperforms traditional statistical models at predicting a range of clinical outcomes from a patient’s entire raw electronic health record (EHR). Alvin Rajkomar MD, Research Scientist and Eyal Oren PhD, Product Manager, Google AI, published on the Google AI blog the results of a study that they did together with researchers from UC San Francisco, Stanford Medicine, and The University of Chicago Medicine. The study “Scalable and Accurate Deep Learning with Electronic Health Records” was in Nature Partner Journals: Digital Medicine.
“When patients get admitted to a hospital, they have many questions about what will happen next. When will I be able to go home? Will I get better? Will I have to come back to the hospital? Having precise answers to those questions helps doctors and nurses make care better, safer, and faster — if a patient’s health is deteriorating, doctors could be sent proactively to act before things get worse. Predicting what will happen next is a natural application of machine learning,” they wrote on the blog.
They went on to explain that for predictions to be useful in practice they should be scalable and acurate: “Scalable: Predictions should be straightforward to create for any important outcome and for different hospital systems. Since healthcare data is very complicated and requires much data wrangling, this requirement is not straightforward to satisfy. Accurate: Predictions should alert clinicians to problems but not distract them with false alarms. With the widespread adoption of electronic health records, we set out to use that data to create more accurate prediction models.”
The research team developed a data processing pipeline for transforming EHR files into a standardized format and then applied deep learning models to data from more than 200,000 patients which were hospitalized for more than 24 hours each at two academic medical centers. Their results show that their algorithm could accurately diagnose diseases, from cardiovascular illnesses to cancer, and predict related things such as the length of hospital stay, hospital readmission and even the likelihood of death (with over 90 percent accuracy).
Source: Google AI Blog – See the full text here