Effects of computerized decision support systems on blood glucose regulation in critically ill surgical patients

The use of computerized decision support systems (CDSS) in glucose control for critically ill surgical patients has been reported in both diabetic and nondiabetic patients. Prospective studies evaluating its effect on glucose control are, however, lacking. The objective of this study was to evaluate patient-specific computerized IV insulin dosing on blood glucose levels (BGLs) by comparing patients treated pre-CDSS with those treated post-CDSS.

A prospective study was performed in 4 surgical ICUs and 1 progressive care unit comparing patient data pre- and post-implementation of CDSS. The primary outcomes measures were the impact of the CDSS on glycemic control in this population and on reducing the incidence of severe hypoglycemia.

Data on 1,682 patient admissions were evaluated, which corresponded to 73,290 BGLs post-CDSS compared with 44,972 BGLs pre-CDSS. The percentage of hyperglycemic events improved, with BGLs of >150 mg/dL decreasing by 50% compared with 6-month historical controls during the 18-month study period from July 2010 through December 2011. This was true for all 5 units individually (p < 0.0001, by one sample sign test). In addition, severe hypoglycemia (defined as BGL <40 mg/dL) decreased from 1% to 0.05% after implementing CDSS (p < 0.0001 by 2-sided binomial test).

Patients whose BGLs were managed using CDSS were statistically significantly more likely to have a glucose reading under control (<150 mg/dL) than in the 6-month historical controls and to avoid serious hypoglycemia (p < 0.0001).