Adopting a Centralized Glucose Management Program Improves Clinical Outcomes in General Hospital Patients with Diabetes

By adopting a comprehensive set of case management policies for the care of diabetic and hyperglycemic patients, Johns Hopkins Hospital improved in-hospital mortality rate, reduced length of stay, and reduced hyperglycemia, according to an oral presentation given at the American Diabetes Association 73^rd Scientific Sessions in Chicago, IL.

The management policies included the ADA policy, which was adopted totally, as well as the institution of a supervisor nurse, computerized order sets and insulin dosing, according to Elias Spanakis, MD, a physician in the division of endocrinology and metabolism at Johns Hopkins.

“In the general ward setting in the hospital, we use insulin to control hyperglycemia, we avoid giving pill medications, and sometimes advise patients not to eat,” explained Spanakis.

The changes over four years were associated with a 47% reduction in the mortality rate, and a $3,200 decrease in the cost of patient care, said Spanakis, who presented abstract 261-OR, “Impact of a Hospital-Wide Inpatient Glucose Management Program on Economic and Clinical Outcomes in the Non-Critical Care Setting.”

“In 2006, one of the patients at our hospital had a sentinel hypoglycemic event, so the hospital decided to adopt a number of management policies,” explained Spanakis.” In addition to adopting the ADA management policy, the institution also implemented a supervising nurse program, uniform computerized order sets, targeted staff education, and clinical decision support.

A retrospective review of a cohort study of non-pregnant adults in the non-critical care units of the hospital with diabetes or hyperglycemia confirmed the improvements. The researchers tracked 16,537 patients who had been admitted across four intervention periods between 2006 and 2009, and found that the adjusted mortality rate after the most recent policy had been implemented was 0.43 (p = 0.01).

After adjusting for severity of illness and other factors, however, the finding just missed statistical significance (p = 0.08), Spanakis said.

In addition, the abstract said that the average length-of-stay was reduced by 0.93 days (p = 0.026), compared to the period prior to the institution of any of the policies.

“Under the model which was fully-adjusted for mortality, however, the decrease was deemed to be half a day, which was not statistically significant, with a p value of 0.15,” Spanakis added.

The improvements were partially explained by improved glycemic control, according to the abstract, which noted that the glucose management program significantly improved the mortality rate and length-of-stay, but did not reduce hospital charges. The findings regarding cost had a p value of 0.04, Spanakis said.

“We didn’t analyze what impact these policies may have had on in-hospital, hyperglycemia-related infections,” added Spanakis, who noted that some researchers have associated hyperglycemia with an increased rate of infection due to a diminished immunological response.