Doctors from Brigham and Women’s Hospital (Boston, MA) have proposed that limiting the overuse of costly resources through predictive analytics could help to reduce rising health care costs for high-risk patient groups.
The United States spends twice as much on health care as most industrialized nations. To mitigate this, some systems have begun to use electronic health records (EHRs) to look at retrospective claims data to better identify high-risk individuals. However, these systems often do little in the way of risk stratification, making it difficult to apply the information obtained to the care of an individual patient.
In a Viewpoint published in The Journal of the American Medical Association, authors, led by Ravi B Parikh, MD, MPP (Brigham and Women’s Hospital, Boston, MA), explained how some health systems are starting to better implement analytics into clinical care. Citing examples of how other industries have successfully used predicative analysis to better tailor their services—such as Amazon’s product recommendation system—the authors propose that these systems could be adapted to health care in order to improve health outcomes for patients and populations based on data derived from historical patient records. Some organizations have already begun to implement these systems and, as more data becomes available, more institutions may be able to use them to develop better predictive models for different clinical issues.
Kaiser Permanente of North California (KPNC), an integrated health service organization, has begun to use predictive analytics to manage the use of antibiotics for neonates. KPNC used maternal data from over 600,000 births to determine the likelihood of early-onset neonatal sepsis in nonpremature infants prior to birth. The newborns were classified as low-, medium-, or high-risk. Physicians used this data to determine whether they should administer antibiotics.
In another example, Parkland Health and Hospital System (Dallas, TX) used electronic health data based on 29 clinical, social, behavior, and utilization factors available within 24 hours of hospital admission to better determine which patients were most at risk for 30-day readmission, which accounts for over $41 billion in health care spending annually.
The Veterans Health Administration (VHA) created the Corporate Data Warehouse, which houses patient-level data gathered from across the VHA. More than 12,000 physicians now use this data to guide treatment strategies, including palliative care for high-risk individuals, and are demonstrating improved outcomes including reduced hospitalizations and emergency department visits for their institutions.
The authors suggest that, for the future use of EHRs, data will need to be better integrated at the care level so that physicians can access data quickly and use it to make decisions in real time. Also, health care systems will need to work toward integrating EHRs as a part of their full care continuum.
The authors are confident that the proper integration of EHRs could improve predictive analytics and bring greater value to the care that patients receive. The most important step in this transformation will be implementing these programs, termed "precision delivery," as part of routine care.
