Researchers at the University of Florida College of Pharmacy and UF Health are turning to prediction algorithms, or risk scores, to prevent adverse drug events and better manage glucose levels of hospitalized patients. The scores are designed to run fully automated in a hospital’s electronic health records and identify patients at risk for hypoglycemia and hyperglycemia.
A research team led by Almut Winterstein, Ph.D., professor of pharmaceutical outcomes and policy, has been awarded a nearly $300,000 grant from the Food and Drug Administration, or FDA, to study the risk score’s effectiveness in drawing attention to patients at highest risk for hypoglycemia and hyperglycemia.
“The prediction models have been developed and validated in the two largest hospitals affiliated with the University of Florida and show excellent predictive performance,” Winterstein said.
UF researchers from the College of Pharmacy’s department of pharmaceutical outcomes and policy are partnering with bioinformatics specialists and hospital pharmacists at UF Health to integrate the statistical model into the electronic health records at UF Health hospitals in Gainesville and Jacksonville.
The computer program pulls data from more than 30 health variables associated with hypo- and hyperglycemia patients such as laboratory values, medications and foods consumed. The data are retrieved from a patient’s medical record of the current and previous hospital stays and clinic visits during the past year. Each variable is weighted and assigned a multiplier to produce a patient’s risk score. Hospital pharmacists use the scores to tailor glucose management strategies based upon the predicted outcome.
“The study predicts that patients with the highest risk score are most likely to suffer from hypoglycemia or hyperglycemia,” said Thomas Johns, Pharm.D., director for pharmacy services at UF Health Shands Hospital. “By providing pharmacists with this information that was previously not available, we are making hospitals safer and directing care activities toward those patients at most risk for preventable adverse drug events.”
Glucose management is one of the most challenging drug issues in hospitals and a focus of a national action plan by the U.S. Department of Health and Human Services to reduce preventable adverse drug events. Insulin’s narrow therapeutic range makes it susceptible to adverse events such as over- or under dosing. As hospitals address new federal quality improvement measures, pharmacists will play a key role in managing drug-related issues and will require more real-time data for effective clinical decision making.
“Today’s electronic medical record systems flag overdoses and dangerous drug interactions when physicians enter a medication order into a patient’s record,” Winterstein said. “Our software addition goes a step further and actually predicts an outcome. The challenge for pharmacists and health care providers is to take the predictive information and figure out how to effectively manage a patient’s blood sugar levels when he or she is at an elevated risk.”
At UF Health, pharmacists work closely with a team of caregivers, including physicians from the Division of Endocrinology, Diabetes and Metabolism, to manage patient glucose levels. In the next year, UF researchers will review if the risk scores presented to pharmacists translate into better management of blood sugar levels, and their potential impact on trends in hypoglycemia and hyperglycemia in hospitalized patients.