CHICAGO – A centralized glucose management program at Johns Hopkins Hospital had a significant impact on hospital costs, and reduced the in-hospital mortality rate and length of stay for patients with diabetes or hyperglycemia, a 3-year analysis showed.

"This policy is important ... to make the hospital more efficient and less costly," said Dr. Elias K. Spanakis of Johns Hopkins University, Baltimore, who presented his abstract at the annual scientific sessions of the American Diabetes Association.

Studies have shown that it is associated with increased mortality, higher risk of infections, and prolonged length of stay (J. Clin. Endocrinol. Metab. 2002;87:978-82 and Diabetes Care 1999;22:1408-14).

Many hospitals now have hypoglycemia policies in place, said Dr. Spanakis, and there have been other studies on the topic. But what sets this study apart from previous research is the fact that it is long term and also considers multiple glycemia-related policies. Most of the previous studies have been done on isolated policies, and over a short period of time, he said.

In 2006, Johns Hopkins Hospital, a 1,000-bed tertiary referral center, developed an institutional glucose management committee led by an endocrinologist and a diabetes nurse practitioner.

The program was implemented between January 2006 and December 2009 and included ongoing staff education. Its individual elements were developed and implemented gradually: hospital-wide hypoglycemia policies, diabetes nursing super-user program, hospital-wide hyperglycemia order set and policy, and hyperglycemia order sets.

Specifically, the policies included a subcutaneous insulin order set, guidelines for calculation and management of insulin doses, and a transition protocol from intravenous to subcutaneous insulin, Dr. Spanakis said. The computer-based smart hyperglycemia order set gave dosing recommendations based on four clinical questions: including the patient’s type of diabetes, type of insulin, total daily dose of insulin, and the nutrition source.

By the end of 2009, the hospital showed a sustained reduction in hyperglycemic episodes. However, to know whether the program had any impact on in-hospital mortality rate, length of stay and total and daily hospital costs required a closer look.

Dr. Spanakis and colleagues conducted a retrospective cohort study of 16,537 unique admissions for patients who had a diagnosis code for diabetes or hyperglycemia during the program’s 4-year stint. Patients who were nonpregnant adults in non–critical care units were excluded.

Researchers compared the results of four consecutive intervention periods with the baseline (January to June 2006). The intervention periods included the implementation of hypoglycemia policy at the hospital (July to Dec. 2006), the diabetes nursing super-user program (January to November 2007), the hyperglycemia order set and policy (December 2007 to November 2008), and finally the smart hyperglycemia order set (December 2008 to December 2009).

They used three models:

• Model 1 adjusted for age, sex, and race.

• Model 2 adjusted for model 1 plus hyperglycemia.

• Model 3 adjusted for model 2 plus mortality risk index, severity of illness index, and hospital unit.

In all three models, researchers found more than $3,000 in reductions in hospital costs per admission in the final period compared with baseline, which was statistically significant in all models. The researchers are still collecting data in this area, Dr. Spanakis said.

For mortality rate ratio, model 1 showed a significant 68% reduction in mortality in the final period of the study versus baseline. The significance persisted in model 2.

However, in model 3, the mortality rate ratio became statistically nonsignificant (0.53; P = .08), showing a 47% reduction in the final period, compared with baseline.

Length of stay was also significantly reduced in models 1 and 2, but lost its significance in model 3 (0.54 fewer days; P = .15).

"People are trying to get better at safety within hospitals, and this seems to be an area where they’re making progress."

Dr. Spanakis said that although the mortality rate and length of stay did not reach statistical significance after the full adjustment in model 3, the policies did have some effect.

The study had some limitations, including its retrospective design. Also, the interventions were not randomized and "secular trends in other inpatient practices may have impacted the outcomes," Dr. Spanakis said.

"People are trying to get better at safety within hospitals, and this seems to be an area where they’re making progress," Thomas J. Hoerger, Ph.D., a health economist and senior fellow at RTI International, Research Triangle Park, N.C., said in an interview. He was not involved in the study.

Dr. Spanakis said further studies could determine which specific program components and policies are most effective in reducing hospital costs, in-patient mortality, and length of stay in patients with diabetes and hyperglycemia.

Dr. Spanakis had no disclosures.

nmiller@frontlinemedcom.com

On Twitter @NaseemSMiller

Glucose management program reduced hospital costs