A team of researchers from Kaiser Permanente has created a risk stratification tool that can identify type 2 diabetes (T2D) patients at greatest risk for hypoglycemia-related emergency department (ED) or hospital use. Findings from this study were published in JAMA Internal Medicine.

Hypoglycemia is a common adverse event in patients with T2D, and is often associated with glucose-lowering medications. Using data from over 200,000 T2D patients, the researchers identified 156 possible risk factors for hypoglycemia, and through “machine-learning analytical techniques” were able to create a model to predict a patient’s 12-month risk of hypoglycemia-related ED or hospital use.

The final model was based on six patient-specific variables: number of prior episodes of hypoglycemia-related ED visits or hospitalizations; insulin use; sulfonylurea use; severe or end-stage kidney disease; number of ED visits for any reason in the past year; and age. 

Based on these variables, the researchers categorized patients into high (>5%), intermediate (1–5%), and low (<1%) risk groups. They then validated the tool using data from over 1.3 million members of the U.S. Veterans Health Administration and around 15,000 Kaiser Permanente members with T2D. The accuracy of the model was assessed in the internal validation sample (n=41,287); 2% were considered high risk, while 10.7% and 87.3% were considered intermediate and low risk, respectively. The observed 12-month hypoglycemia-related utilization rates for these groups were 6.7%, 1.4%, and 0.2%, respectively.

Related Articles

“This tool could be integrated with targeted preventive interventions to facilitate population management, which ultimately could reduce future hypoglycemia risk and improve patient safety,” the authors write. They note that while the two criteria for high risk patients may be easily memorized by clinicians (≥3 previous episodes of hypoglycemia-related utilization; one or two episodes if treated with insulin), “the criteria for intermediate risk are more nuanced and therefore may be less likely to provoke clinical action in primary care without prompting.” 

For more information visit JAMAnetwork.com.