New Tool Helps Identify Patients With Type 2 Diabetes at High Risk for Hypoglycemic Events

Diabetes - Evidence and Advice

New Tool Helps Identify Patients With Type 2 Diabetes at High Risk for Hypoglycemic Events

Before your group commits to this or any other tech initiative, consider starting more simply. And be aware of alert overload.

Author: Renee Cocchi

Clinical Advisor: Harold H. Katz, MD, FACP, FACE

You may hear about a tool recently tested by researchers1 that can help categorize 12-month risk of hypoglycemia-related ER/hospital use in patients with type 2 diabetes. While it potentially has value, you might be better off taking another approach, says our clinical advisor.

A classification tree based on potential predictors of hypoglycemia-related emergency department (ED) or hospital use was created using these 6 electronic health record (EHR) inputs: hypoglycemia-related utilization, insulin use, sulfonylurea use, ED use, chronic kidney disease, and age. Investigators tested it in >1.5 million people with type 2 diabetes and found that it showed good discrimination and has the potential to help primary care clinicians offer better targeted interventions. But the study did not reveal cost to install in an EHR system.

“It seems that a tool like this would be of moderate value to large healthcare systems,” says Harold H. Katz, MD, FACP, FACE, lead physician at Allina Health United Medical Specialties, St. Paul, MN. The way he sees it, if you are a small or mid-sized group, rather than taking the time to “… put in some sophisticated software to flag a small number of individuals per year, your technical time could be better spent trying to flag other higher-risk patients, such as those at high risk for heart failure.”

Short of that, not knowing the cost makes it difficult to evaluate the tool’s potential cost-effectiveness, adds Dr Katz. Once that information is known —perhaps after a prospective study — an organization can determine how many events it is preventing and at what net savings.

Until then, you could also start out more simply, he suggests. “For instance, look at those individuals who had two or more ER visits or hospital admissions in the past 12 months.” Start there, and get them the help they presumably need, whether that is a visit with the diabetes nurse educator or a prompt to consider an alternative agent. “We already know that being on insulin, sulfonylureas or being of elderly age puts the patient at greater risk for severe hypoglycemia without a tool telling us so.”

Finally, no matter how you proceed, Dr Katz warned of alert overload. “When every little thing in the EMR is flagged, you tend to start ignoring the flags, kind of like the boy who cried wolf,” he explains. This is one danger of adding too many flags to EMR systems.

1Karter AJ, Warton EM, et al. Development and validation of a tool to identify patients with type 2 diabetes at high risk of hypoglycemia-related emergency department or hospital use. JAMA Internal Medicine. 2017; August 21: epub ahead of print. Doi:10.1101/jamainternmed.2017.3844.