Objective: To derive a target range of optimal sedation for the COMFORT Scale and to prospectively test that target range against intensivist assessment of adequacy of sedation.
Design: Serial prospective agreement cohort studies.
Setting: Twelve-bed pediatric intensive care unit in an urban academic teaching hospital.
Patients: Eighty-five mechanically ventilated children (aged 0 to 102 months).
Interventions: Three serial prospective studies comparing simultaneous, independent ratings conducted by trained observers using an objective scale and intensive care physicians using global assessment. The initial study was designed to derive the target range. The second study was conducted to verify that target range in a second population. The third study was added to evaluate relative variability in methods used in the second study.
Measurements and main results: Adequacy of sedation using visual analog scale and descriptive ratings or the COMFORT Scale (a previously validated behaviorally anchored scale to rate eight behavioral or physiologic dimensions of distress). The first study comprised 100 observations. Groups of patients described by the intensivist as inadequately sedated, optimally sedated, and excessively sedated had different mean COMFORT scores (30.5 +/- 0.7 vs. 22.9 +/- 5.8 vs. 14.3 +/- 0.7, respectively, p < .05). The target range of optimal sedation was defined as COMFORT scores of 17 to 26. The second study verified the target range prospectively in a second group of 96 observations. The COMFORT score was strongly associated with the sedation adequacy rating by the intensivist (p < .001; r2 = .662). COMFORT scores accurately predicted the patient assignment to adequacy of sedation categories by the intensivist in 66.1% of observations. Discrepancy between physicians occurred in 38.5% of 16 paired physician ratings in the second study. In the third study, 120 observations comparing paired COMFORT scores to paired physician ratings of the same subjects demonstrated significantly less variability in COMFORT assessment of adequacy of sedation. COMFORT scores were similarly unbiased (1.1% vs. 0.22%) but more precise (8.0% vs. 16.7%) than intensivist ratings (p < .025).
Conclusion: Adequacy of sedation is measured more consistently by observers using the COMFORT Scale than by intensivist global assessment.