Computer Algorithms Managing Insulin Dosages Result in Better Glycemic Control

MAY 31, 2016

Computer Algorithms Managing Insulin Dosages Result in Better Glycemic Control

When insulin dosages are calculated using computer algorithms to reach target blood glucose levels, hospitalized patients are less likely to have episodes of low blood sugar and recover sooner from diabetic ketoacidosis than when insulin dosages are calculated without an algorithm.

Those were key findings from two multisite studies presented at the American Association of Clinical Endocrinologists meeting, held May 25-29 in Orlando, Florida.

“Computer algorithms facilitate doing the right thing,” said Dr. Joseph A. Aloi, professor of endocrinology and metabolism at Wake Forest Baptist Medical Center in Winston-Salem, North Carolina, and the studies’ principal investigator.

In the first study, Aloi’s research team analyzed nearly 60,000 patient health records from four health systems. They determined that patients whose subcutaneous insulin dosages were calculated by an electronic glycemic management system embedded into electronic health records were less likely to develop hypoglycemia than patients whose insulin dosages were calculated without the algorithm.

Only 13.8 percent of patients whose dosages were determined with the computer algorithm experienced hypoglycemia, compared with 21.7 percent of patients in the other group. That’s an important finding, Aloi pointed out, because hypoglycemia episodes are associated with longer hospital stays and increased mortality.

In the second study, Aloi’s researchers reviewed patient records of about 2,600 patients treated at 34 medical sites to track incidents of diabetic ketoacidosis, or dangerously high blood glucose levels. Patients whose intravenous insulin dosages were calculated with Glytec Systems Glucommander computer algorithm recovered in an average of 9.15 hours, which was nearly two hours more quickly than the 10.9 hours it took for patients to recover whose insulin dosages were calculated using a paper-based system. What’s more, the average hospital stay of the Glucommander patients was about a day shorter than that of the patients in the other group.

While glycemic management computer algorithms are commonly built into EHRs, “physician inertia” means that clinicians often bypass them, Aloi explained. He said he hopes the studies will encourage more doctors to use software embedded into EHRs to calculate insulin dosages for hospitalized patients.

“It saves money, prevents hypoglycemia, and may prevent mortality,” Aloi said.

View on MedCity News