HQ BaselineInterpreting results
Reliable Change Indices: The Statistical Tool That Determines Whether Your Post-Injury Score Is Actually Worse
A statistical tool that separates real cognitive decline from normal noise.
When a clinician compares a post-injury ImPACT test to a baseline, they don’t simply look for any decline. A small drop in score might be meaningless — just normal test-retest variability. Or it might represent genuine concussion-related cognitive impairment. The statistical tool that distinguishes between these two scenarios is the Reliable Change Index (RCI).
What RCIs do
RCIs were developed from psychometric research on test-retest reliability. The foundational work for ImPACT’s RCIs was published by Iverson et al. (2003) in the Archives of Clinical Neuropsychology, using a one-week test-retest design. Here’s the concept in plain language: when the same person takes the same test twice, their scores will vary somewhat — even if nothing has changed about their brain function. This variation comes from normal factors like attention fluctuations, fatigue, motivation, and random noise in cognitive performance.
The RCI calculates how much variation is expected by chance alone, accounting for the test’s standard error of measurement and any expected practice effects (the tendency for scores to improve slightly on retesting due to familiarity). If a post-injury score declines more than the RCI threshold, the clinician can be statistically confident (typically at the 80% or 90% confidence level) that the decline represents genuine impairment rather than chance.
Where RCIs fall short
ImPACT automatically applies RCIs when comparing post-injury to baseline scores and flags composites where the decline exceeds the reliable change threshold. However, clinicians should be aware that the published RCIs were derived from short-interval test-retest data — and as noted by researchers in PMC (Bailey et al., 2020), these indices may be less appropriate when the baseline was administered months or years before the post-injury test. See also the practice effect problem.
Clinical judgment still required
At Headquarters, our clinicians understand the strengths and limitations of reliable change statistics and interpret post-injury data within the full clinical context — never relying on a single statistical threshold to make clearance decisions.