Evaluation of a Computerized Insulin Dosing Tool for the Treatment of Diabetic Ketoacidosis

Computerized insulin dosing tools (CIDT) have been shown to improve the care of critically ill patients with hyperglycemia. Application of a CIDT in addition to a diabetic ketoacidosis (DKA) order set for the treatment of DKA has not been evaluated. Our goal was to determine the effects the CIDT would have on the treatment of a patient with DKA.

In this retrospective, pre-post chart review, a provider-driven insulin dosing strategy (pregroup) was compared to the CIDT (postgroup) with 24-hour pharmacist monitoring. The CIDT utilized an equation that incorporated a patient's most recent blood glucose (BG) value and recommended a rate of insulin (units/hour) every hour.

All baseline characterizes were similar between the 2 groups. There were no significant differences in average time to anion gap closure (≤ 12 mEq/L) or intensive care unit length of stay between the pregroup and postgroup (12.5 [6] hours vs 10.5 [7] hours, P = 0.235; 40.6 [24] hours vs 40.8 [24] hours, P = 0.945). Although not statistically significant, 17 hypoglycemic events (BG < 70 mg/dL) occurred in the pregroup with 4 being severe (BG < 50 mg/dL) while 5 hypoglycemic events occurred in the postgroup, none of which were severe.

This study suggests, when compared to a provider-driven insulin dosing strategy, the CIDT with 24-hour pharmacist monitoring is efficacious and safe for treatment of patients with a primary diagnosis of DKA.