Psychometric Properties of the Coma Near-Coma Scale for Adults in Disordered States of Consciousness: A Rasch Analysis.

Abstract

To examine the construct validity and measurement precision of the Coma Near-Coma scale (CNC) in measuring neurobehavioral function (NBF) in patients with disorders of consciousness receiving postacute care rehabilitation. Rasch analysis of retrospective data. Participants (N=48) with disordered consciousness who were admitted to postacute care rehabilitation. Not applicable. CNC. Assessment with CNC repeated weekly until the participant was conscious or discharged from the postacute care facility (451 participant records). Rating scale steps were ordered for all items. Eight of the 10 CNC items evaluated in this study fit the measurement model (χ²=5332.58; df=11; P=.17); pain items formed a distinct construct. The ordering of the 8 items from most to least challenging makes clinical sense and compares favorably with other published hierarchies of NBF. Tactile items are more easily responded to. Visual and auditory items requiring higher cognitive processing were more challenging. In the full sample, the CNC achieved good measurement precision, with a person separation reliability of 0.87. The items of the CNC reflect good construct validity and acceptable interrater reliability. The measurement precision achieved indicates that the CNC may be used to make decisions about groups of individuals but that these items may not be sufficiently precise for individual patient treatment decision-making.