Dive Brief:
-
The University of Chicago Medicine and Google announced this week they are working together to study whether analyzing EHRs can predict health and can improve the quality of care.
-
The collaboration will use “new machine-learning techniques to create predictive models” that they hope will reduce hospital readmissions and complications, and save patients’ lives.
-
Stanford Medicine and the University of California, San Francisco, are also working with Google on the project.
Dive Insight:
University of Chicago Medicine’s Center for Healthcare Delivery Science and Innovation was created last year “to bring research-quality analytic methods” to face complexity of care and complexity of the system.
Predicting health, while reducing hospital readmissions, could both improve patient outcomes and cut costs. Nearly 20% of Medicare beneficiaries have an unplanned hospital, which costs about $26 billion a year.
The question is — are EHRs the answer? EHR utilization has increased over the past decade, including in smaller hospitals. Nearly 80% of EHR purchases in the U.S. last year were purchased by community hospitals with fewer than 200 beds. Plus, EHRs already have critical patient data.
The fact that hospitals and health systems have such a high EHR adoption rate would make the project valuable if the researchers can pull off and find valuable analytics to help drive better insights into patients' health. Still, many physicians find EHR use to be burdensome so there is the potential to have insights become buried under the dearth of information that floods a physician already. Therefore, user-centric design and information presentation will likely be an area of exploration to consider.