Continuous glucose monitoring-guided insulin infusion in critically ill patients promotes safety, improves time efficiency, and enhances provider satisfaction

Abstract

OBJECTIVE: Evaluate the integration of real-time CGM (rtCGM) into an insulin infusion computer calculator (IICC) to improve glycemic control, time efficiency, safety, and clinician workflow in the intensive care unit (ICU). METHODS: A retrospective analysis was conducted on 35 critically ill adult patients requiring insulin infusion in the surgical and medical ICUs. Dexcom G7 rtCGM values were integrated into an institution-developed IICC using an ongoing validation protocol, allowing for nonadjunctive CGM use. The accuracy of rtCGM was assessed by comparing matched CGM and point-of-care (POC) glucose values using mean absolute relative difference (MARD), surveillance error grid, and Parkes Error Grid analyses. CGM time-in-range metrics, clinician turnaround time for glucose monitoring, and nurse satisfaction were also evaluated. RESULTS: A total of 1291 matched glucose pairs were analyzed. The rtCGM system demonstrated a MARD of 12.5%, with 99.6% of the values falling within clinically acceptable error zones (A+B) on the Parkes Error Grid. Patients in the rtCGM-IICC protocol had mean glucose 141.9 mg/dL, with mean time in range (70-180 mg/dL) 82.8%, time above range (> 180 mg/dL) 14.5%, and time below range (< 70 mg/dL) 0.5%. Clinician time efficiency improved significantly, with POC testing requiring a mean turnaround time of nearly 5 minutes compared to 3-second CGM retrieval. All surveyed nurses (n = 20) reported rtCGM increased efficiency and improved safety and preferred rtCGM with POC over POC testing alone. CONCLUSION: Integrating rtCGM with an IICC protocol in the ICU enhances glycemic control, improves workflow efficiency, and reduces clinician workload while maintaining high accuracy.