Where: Medical University of South Carolina Medical Center, a 709-bed tertiary care referral center in Charleston, S.C.
The issue: Developing a nurse-driven Web-based insulin infusion protocol to manage blood glucose in the intensive care unit, as well as for patients in the rest of the hospital.
“We're in the diabetes belt of the country,” said endocrinologist Kathie L. Hermayer, ACP Member. “At any one time, we have about 25% to 26% of our hospitalized patients with diabetes and then another 12% or so have new-onset hyperglycemia.”
In 2003, Dr. Hermayer and her colleagues at Medical University of South Carolina (MUSC) Medical Center launched a multi-disciplinary diabetes task force to try to improve care for their many hyperglycemic patients. The task force completed a number of projects, including a new hypoglycemia protocol and a move from sliding-scale to basal insulin. But they were still searching for a better way to manage intravenous (IV) insulin.
First, the task force analyzed what other hospitals had done through a literature search. They didn't find anything that quite met their needs. “Every time we tried one of the published algorithms, they just didn't fit our patient population. Physicians and nurses were not happy with it,” explained diabetes educator Pamela C. Arnold, MSN.
Some options were too expensive. “If we had chosen a commercially available product for our institution, it would have cost over $100,000 annually. We realized up front it would be difficult to justify this type of expense,” said Ms. Arnold. The team was also looking for a protocol that could be shared with community hospitals in other parts of South Carolina.
Unable to find an existing solution to their problem, the MUSC team decided to build one. Conveniently, they had some in-house talent to apply to the challenge, a computer whiz named Michael Irving who collaborated on the statistics, said Dr. Hermayer. With his help, the diabetes task force developed a Web-based algorithm that nurses could use to calculate the appropriate rate of IV insulin for patients.
How it works
Under the protocol, the IV insulin rate is calculated using a multiplier as a surrogate for an insulin sensitivity factor. The multiplier starts at 0.03 and changes as the patient's glucose level responds to treatment. The formula combines the multiplier with a patient's current blood glucose level and the target glucose to determine how many units of insulin per hour should be delivered. So the starting formula for the infusion rate is 0.03 × (current blood glucose − 60).
The math behind the protocol may be complicated, but in practice, it's hidden within a simple online tool. To use the protocol, a nurse goes to the Web page for the IV Insulin Infusion Calculator. He or she inputs the patient's current glucose reading, previous glucose level, and the current multiplier. Using those inputs, the calculator computes a new multiplier and new drip rate, measured in units of insulin per hour.
“It's very, very easy for the nurses to use and follow,” said Dr. Hermayer. Research confirmed that the MUSC nurses had success with the tool. In a study published in the Journal of Diabetes Science and Technology (2008;2(3):376-83), more than 80% of 103 surveyed nurses (most of whom were using the protocol for the first time) found it to be easy to implement, easy to interpret and successful in controlling blood glucose values.
The project's objective success in controlling blood glucose was measured in another study in Diabetes Technology and Therapeutics (2007:9(6):523-34). Dr. Hermayer, Ms. Arnold and colleagues used retrospective data from the hospital's cardiothoracic intensive care unit to compare glucose control before and after the calculator was adopted.
In patients with type 2 diabetes, mean blood glucose dropped from 153.8 to 117.6 mg/dL and the percentage of patients not reaching target levels within 48 hours decreased from 26% to 5% after implementation of the protocol. The percentages of patients with glucose over 180 mg/dL dropped significantly under the protocol while there was no statistically significant increase in episodes of hypoglycemia (blood glucose between 40 and 70 mg/dL).
Despite the many successes of the infusion tool, the project leaders have so far been unable to share the resource with community hospitals because it's not FDA approved. “We're looking into [approval] but we have not totally gotten an answer,” said Dr. Hermayer.
After the tool had been in practice for a while, some upgrades were needed. The protocol was modified to offer a choice of blood glucose target ranges. Now, the calculator can be set differently depending on whether a patient is in the ICU, labor and delivery, or the medical/surgical floors. There are also customized targets for treating diabetic ketoacidosis, hyperosmolar non-ketotic coma and neurosurgery conditions.
The computer program also now has pop-up warnings that appear when a multiplier greater than 0.2 is entered. The warnings reduce the likelihood of a dispensing error due to a misplaced decimal point.
How patients benefit
Although the effect of the tool on patient outcomes has not been studied, the MUSC researchers did recently publish an analysis of the collective impact of their glycemic control improvement efforts in the Journal of Hospital Medicine (2009:4(6):331-339). The study, which compared patients admitted to the hospital during the month of June in 2004, 2005, 2006 and 2007, found that overall patients spent 10% more time in the target blood glucose range.
The diabetes task force is continuing the effort to improve glycemic control at MUSC. They've recently submitted phase I of the application to The Joint Commission for advanced certification in inpatient diabetes, and they've started tracking high (over 200 mg/dL) and low (under 50 mg/dL) blood sugars throughout the hospital. “Everybody gets monthly reports, all the service lines,” said Dr. Hermayer. “We're trying to get the word out about how bad, in particular, hypoglycemia can be for patients.”
Words of wisdom
It doesn't take specific expertise to complete a project like this, according to Dr. Hermayer, but it does take a team. “You really can't do this unilaterally. I'm an endocrinologist, but a hospitalist can take this on,” she said. “You need to form something like a task force so you can talk to other members of the staff.”