Table 2.

Summary of Inpatient Glucometrics for Intensive Care Unit (ICU) and Non-ICU Settings^*

Glucometric Model	Numerator	Denominator	Interpretation	Advantages	Disadvantages
Population	All glucose measurements over a given time period	None	Mean or median of all BGs % of all readings within a given glucose category	Largest sample size Easiest to generate Useful for evaluating IV insulin protocols	Doesn’t account for clustering of repeated measures Skews data toward patients with long hospital stays Limited utility outside of the ICU
Patient-day	All glucose measurements over a given time period	Average number of blood glucose readings collected per patient-hospital day	Mean or median patient-day–weighted blood glucose % of patient-days with mean BG within a given glucose category % of patient-days with any episode of adverse glucose outcomes.	Most clinically relevant Appropriately accounts for differences in glucose testing frequency	Skewed somewhat toward patients with long hospital stays
Patient-stay	All glucose measurements over a given time period	Average number of blood glucose readings collected per patient-hospital stay	Mean patient-day– weighted BG for a patient’s hospital stay % of patient stays with day-weighted mean BG falling within a given glucose category.	Best single measure of long-term glycemic control in a given patient	Least powered to detect meaningful changes Doesn’t account for differences in lengths of stay Cannot be used to accurately gauge the frequency of adverse glycemic events

^*BG, blood glucose; IV, intravenous.