Interrater reliability of neurologic soft signs in an acquired brain injury population.

Abstract

To determine the interrater reliability of a battery of tests that assess for neurologic soft signs in an acquired brain injury population. Correlation study of the battery's interrater reliability. Inpatient and outpatient acquired brain injury rehabilitation program. Fourteen patients with an acquired brain injury and two control subjects. The battery of tests for neurologic soft signs was administered to all 16 subjects. Each subject was videotaped. Fifteen "nonexpert" allied health professionals independently rated the subjects' performances. Two scoring systems were used. All neurologic soft signs were rated on an ordinal scale, and, where possible, a quantitative measure of performance was recorded. Using the ordinal scale scoring system, all neurologic soft signs had interclass correlation coefficients (ICCs) of > or =.60. The signs "Two-Point Discrimination" and "Go-No-Go part 1 and part 2" achieved ICCs of > or = .75. For the scoring system requiring a quantitative measure of performance, ICCs ranged from .18 to .97. Neurologic soft signs administered to an acquired brain injury population can be reliably rated by "nonexpert" clinicians. The signs should be rated on an ordinal scale.