Discrimination and other statistical intervals for the interpretation ofin vivopatient monitoring data

摘要

AbstractA calibration line is used to define the relationship between a new clinical technique and a standardin vitrolaboratory methodology. Discrimination intervals quantify the reliability of inverse estimates obtained from the calibration line. Applied to transcutaneousP CO 2monitoring, a newin vivomeasurement, discrimination intervals for inverse estimates of arterialP CO 2range in size from 10 to 50 torr (1.33–6.67 kPa), or more, depending on the values of statistical parameters selected. Discrimination interval analysis shows that the inverse estimates are more reliable near the means. Also, it indicates that the primary value of transcutaneousP CO 2monitoring lies not in providing accurate inverse estimates of arterialP CO 2, but instead, in monitoring short‐t