Development and implementation of evidence-based guidelines for IV insulin: A statewide collaborative approach

Purpose: Recent studies have shown that the outcomes of hospitalized patients are greatly enhanced when steps are taken to improve control of their blood glucose levels. The Georgia Hospital Association Research and Education Foundation's Partnership for Health Accountability established a Diabetes Special Interest Group (D-SIG) in February 2003. Goals of the D-SIG were to enlighten health care professionals in Georgia hospitals about the benefits of controlling hyperglycemia in hospitalized patients and to develop processes to assist hospitals in the adoption of an IV insulin dosing algorithm, development of an IV insulin standing order set, and implementation of a hyperglycemia management plan. Methods: The D-SIG created an assessment tool titled "Key Elements of IV Insulin Guidelines" and evaluated numerous published IV insulin administration algorithms and protocols. After an extensive literature review, including international protocols and guidelines, user-friendly guidelines for subcutaneous and IV insulin were developed by a multidisciplinary work group, with members representing hospitals and other stakeholders from throughout the state. The group chose a well-researched method that was available in both computerized and hand-calculated formats and developed a Columnar Insulin Dosing Chart to assist with IV insulin infusions. This insulin-infusion table stems from mathematical formulas published by multiple investigators since the 1980s. The D-SIG guidelines and dosing chart were evaluated for ease of use, effectiveness, and safety in 3 settings: a small, rural critical-access hospital (CAH); an intensive care unit (ICU) in the trauma center of a large Georgia teaching hospital; and a surgical ICU in a midsize metropolitan hospital. Results: After implementation of the guidelines, the incidence of hypoglycemia (blood glucose level <60 mg/dL) was 0.9% in the trauma center ICU and 0.6% in the surgical ICU. All hypoglycemic patients in these 2 settings were asymptomatic, remained hypoglycemic only for a short time, and experienced no complications attributable to hypoglycemia. Using a moderate insulin sensitivity level for dosing initiations resulted in a time to target blood glucose level (80-110 mg/dL) of 6.4 hours, whereas using the most conservative approach required 12.8 hours to attain target range. At the CAH, time to reach the target blood glucose level (90-140 mg/dL) was 5.8 hours, and no episodes of hypoglycemia were reported. Although not part of the pilot initiative, the surgical ICU also reported a 5-fold reduction in surgical infection rates. The success of the dosing chart and standing order set paralleled that of the computerized formula when similar initiation doses were used. Conclusions: The Columnar Insulin Dosing Chart and sample clinical guidelines were piloted at 3 different settings and found to be safe and effective. Furthermore, by including the treatment for hypoglycemia in the guidelines, nurses in all patient care areas were able to manage blood glucose levels below the target range in a safe and timely manner. Use of the dosing chart and guidelines reduced blood glucose levels to the target range with no clinically significant hypoglycemia.