What you need to know about the forces reshaping our industry.


May 23, 2017

How 3 hospitals are partnering with Google to predict patients' futures

Daily Briefing

    The University of Chicago (UChicago) Medicine on Wednesday became the latest hospital to partner with Google to explore whether machine learning can predict a patients' likelihood of being readmitted to a hospital to improve health outcomes.

    Get your readmissions reduction toolkit

    Stanford Medicine and the University of California-San Francisco also have paired with Google on the initiative. According to CNBC, the effort comes as hospitals face increasing pressure to keep patients healthy and avoid readmissions as the U.S. health system shifts away from traditional fee-for-service reimbursement models toward value-based payments.

    Partnership details

    The partnership will leverage experts in predictive modeling from UChicago Medicine's Center for Healthcare Delivery Science and Innovation and Center for Research Informatics and Google's machine-learning tools, which researchers developed under the company's Google Brain initiative.

    The researchers will look to see whether there are patterns in individuals' electronic health records that indicate whether the patients will be hospitalized, how long hospitalizations could last, or whether an individual's health is worsening, the Chicago Tribune reports.


    Katherine Chou, head of products at Google Brain, said, "We can improve predictions for medical events that might happen to you." She added, "There's so much health care data, especially with [EHRs] being adopted over the last 10 years. The potential for using that data for predictions, people haven't really figured out how to harness it."

    Michael Howell, director of UChicago Medicine's Center for Healthcare Delivery Science and Innovation, said, "Prediction helps make patient care better. It's a core component of prevention, and it can also make complex care safer." However, Howell said, "Traditional tools of epidemiology and statistics simply can't use free text or images to create predictive algorithms that could alert physicians and nurses about patients' risks for problems." He added, "But together with Google, we can" (Kim Cohen, Becker's Health IT & CIO Review, 5/19; Schencker, Chicago Tribune, 5/17; Farr, CNBC, 5/17; Wood, "ScienceLife," University of Chicago Medicine & Biological Sciences, 5/17).

    Get your readmission reduction toolkit

    Reducing your hospital's readmission rates can seem like an overwhelming task—but it doesn't have to be.

    Knowing where to focus is half the battle. We've found that the best strategies target four stages of care with significant potential to influence patient outcomes. The other half is knowing what improvements to make.

    That's where our Readmission Reduction Toolkit comes in. We've compiled resources from across Advisory Board that will help you isolate and correct patient and systemic issues in the four critical stages of care:

      Stage 1:  Transition planning during the inpatient stay

      Stage 2:  Discharge education

      Stage 3:  Post-acute care coordination

      Stage 4:  Transitional care support

    Get the Toolkit

    Have a Question?


    Ask our experts a question on any topic in health care by visiting our member portal, AskAdvisory.