Reliable change, sensitivity, and specificity of a multidimensional concussion assessment battery: implications for caution in clinical practice.

OBJECTIVE: To provide reliable change confidence intervals for common clinical concussion measures using a healthy sample of collegiate athletes and to apply these reliable change parameters to a sample of concussed collegiate athletes.
METHODS: Two independent samples were included in the study and evaluated on common clinical measures of concussion. The healthy sample included male, collegiate football student-athletes (n = 38) assessed at 2 time points. The concussed sample included college-aged student-athletes (n = 132) evaluated before and after a concussion. Outcome measures included symptom severity scores, Automated Neuropsychological Assessment Metrics throughput scores, and Sensory Organization Test composite scores.
RESULTS: Application of the reliable change parameters suggests that a small percentage of concussed participants were impaired on each measure. We identified a low sensitivity of the entire battery (all measures combined) of 50% but high specificity of 96%.
CONCLUSIONS: Clinicians should be trained in understanding clinical concussion measures and should be aware of evidence suggesting the multifaceted battery is more sensitive than any single measure. Clinicians should be cautioned that sensitivity to balance and neurocognitive impairments was low for each individual measure. Applying the confidence intervals to our injured sample suggests that these measures do not adequately identify postconcussion impairments when used in isolation.