PURPOSE: Guidelines and protocols (both paper-based and computerized) are used to stabilize clinical care process. Clinicians manage blood glucose in ICU patients because patient survival was reported to increase 6-9% absolute/10 mg/dl reductions in blood glucose.
METHODS: We compared three 80-110 mg/dl blood glucose target decision-support strategies with varying detail and process control: a simple guideline, (without a bedside tool), a bedside paper-based protocol, and a bedside computerized protocol. We also evaluated the bedside computerized protocol in three distinct environments (LDS, UVA and NUH-Singapore) for clinician acceptance of instructions and glucose control.
RESULTS: Glucose control with different protocols [mean (mg/dl), % values between 80-110 mg/dl, % values < 40 mg/dl]: Computer- 113, 46%, 08% Paper - 134, 28%, .04%, Guideline - 141, 22%, 0.28%. Glucose control in different environments [mean (mg/dl), % values between 80-110 mg/dl, % values < 40 mg/d]: LDS - 113, 46%, .08% UVa - 116, 42, .18% NUH-Singapore 111, 41%, .66%. Clinician acceptance rates of computer genertated instructions were 95, 95 and 98%, respectively.
CONCLUSION: Glucose control is superior with a bedside computers protocol when compared to approaches relying upon explicit a paper protocol or guideline. The computerized protocol enabled a replicable means of process control in three sites in two cultural environments with high clinician acceptance rates of computer generated instructions.
CLINICAL IMPLICATIONS: Bedside computers permit exportation of complicated clnical protocls to multiple sites with high reproducibility of results and process. This could prove invaluable in multicentered clinical trials or in translating evidence based medicine to the bedside.
DISCLOSURE: Jonathon Truwit, None.