Watson-like tech tackles readmissions
October 11, 2013 in Medical Technology
UNC Health Care is using IBM big data analytics to help hospital workers reduce costly and preventable readmissions, decrease mortality rates and improve patient care.
UNCHC is employing an IBM Smarter Care solution, which enables clinicians to quickly access and analyze critical patient information using natural language processing similar to what’s used in IBM Watson technology. With the ability to see and interpret both structured and unstructured data, UNCHC can now identify high-risk patients, understand in context what is causing them to be hospitalized, and then take preventative action.
More than 80 percent of an institution’s data today is unstructured – meaning it cannot readily be collected and analyzed using standard methods, IBM executives note in a news release. In healthcare, this is in the form of physician notes, registration forms, discharge summaries, phone calls, documents and more. In addition, medical literature is doubling every five years, making it difficult for clinicians to remain up-to-date with the latest scientific information.
[See also: Big data sets sights on heart disease.]
“IBM Content Analytics allows us to quickly transform raw information into healthcare insights,” said Carlton Moore, MD, Associate Professor of Medicine at UNCHC. “It can reveal trends, patterns and deviations while predicting the probability of outcomes so that we can make decisions in minutes versus weeks or months. This is a game changer for us.”
Previously, UNCHC used IBM Content Analytics to mine clinical data to improve the accuracy of its 2012 Physician Quality Reporting System measures, achieving double digit quality improvements in the areas of mammogram, cancer and pneumonia screening.
[See also: UNC Health Care, IBM launch HIE.]
UNCHC is focusing the new technology in three areas:
Timely follow-up of abnormal cancer screening results. Follow-up care for patients with abnormal tests is often delayed because the results are buried in electronic medical records. Using IBM Content Analytics, UNCHC can extract abnormal results from cancer screening reports such as mammograms and colonoscopies and store the results as structured data. The structured results are used to generate alerts immediately for physicians to proactively follow-up with patients that have abnormal cancer screening results.
Reducing costly 30-day readmissions. Preventable readmissions impact one in five U.S. patients, which adds unnecessary costs to the already strained health system, according to a 2012 article in the New England Journal of Medicine. As of last year, hospitals are also penalized for high readmission rates, with reductions in Medicare discharge payments. UNCHC the analytics technology to extract predictors of readmission risk from free-text clinical notes to find more effective ways to care for high-risk patients and provide safer patient care.
Engaging more patients. Getting patients involved in healthcare management is key to improving health outcomes, but clinical data in today’s patient portals is often unfiltered and hard for patients to understand. UNCHC will use the analytics tool to transform clinical data from electronic medical records into a simpler format so that patients can better understand their health information and participate in their care management plan.
Based on the success of these programs, UNCHC is planning to apply IBM Content Analytics to additional use cases such as helping patients with diabetes or patients with other chronic illnesses.