Predictive Tool Could Target Patients at Risk of Hypoglycemia-Related Utilization

Researchers have outlined a 6-item risk stratification tool that successfully predicted the likelihood that patients with diabetes will have a hypoglycemia-related emergency department (ED) visit or hospital admission.

According to the study, published in JAMA Internal Medicine, interventions to prevent costly episodes of hypoglycemia have been developed, but are difficult to target and apply to appropriate populations with type 2 diabetes. Using a prospective model, researchers developed and tested a risk prediction tool using hypoglycemia risk factors that could be found in electronic health records (EHRs).

Beginning with 156 potential hypoglycemia predictors, the study authors narrowed their tool to use 6 of the factors: total number of prior hypoglycemia-related ED or hospital utilization events, number of total ED encounters in prior 12 months, insulin use, sulfonylurea use, severe or end-stage kidney disease, and age younger than 77 years. The responses for these variables were then combined to yield an estimate of the risk of hypoglycemia-related utilization in the next year, classified as high (>5%), intermediate (1% to 5%), or low (<1%).

The tool was then tested for its discrimination ability in 3 samples of patients with diabetes (1 internal within the Kaiser Permanente Northern California health plan and 2 external samples). There were no significant differences between predicted and observed hypoglycemia risk among the internal validation sample. Specifically, it categorized 2% of these patients as high risk, 10.7% as intermediate, and 87.3% as low; the 12-month rates of hypoglycemia-related utilization in these groups were 6.7%, 1.4%, and 0.2%, respectively.

The tool also showed good discrimination in the external validation samples. Additionally, patients classified as high risk were 5 times more likely to report a severe hypoglycemic episode where they necessitated assistance compared with those considered low risk (49.7% vs 9.2%; P <.001).

Because this tool uses data commonly found in EHRs, it does not require patient input and thus “offers an efficient, low-cost approach for identifying patients for targeted interventions to reduce their risk of hypoglycemia,” according to the study authors. Such interventions could include medication monitoring and conversations about behavioral hypoglycemia risk factors.

Less rigorous interventions could be utilized to target the patients considered to be at intermediate risk. Besides its potential uses in clinical settings, researchers could also use such a tool to include or exclude potential study participants based on their risk for hypoglycemia-related utilization.

The study authors wrote that further studies will be “needed to evaluate whether and how implementation of this hypoglycemia risk stratification tool may influence clinician behavior, patient decision making, drug safety, and hypoglycemia incidence.”

According to a press release from Kaiser Permanente, which provided the internal sample of patients and employs some of the study authors, public and private healthcare stakeholders like CMS and the Mayo Clinic are interested in using the risk stratification tool to raise hypoglycemia awareness and potentially prevent future hypoglycemia episodes for at-risk patients.

“This work is an example of how federal agencies can work with private researchers to reduce preventable adverse drug events,” said John Whyte, MD, MPH, director of Professional Affairs and Stakeholder Engagement for the FDA, in the press statement. The FDA had funded the tool’s development under its Safe Use Initiative.