Test-Retest Reliability of ImPACT Across a Two-Year Interval in Collegiate Athletes

Abstract

Abstract Purpose To examine the two-year test-retest reliability of Immediate Post-concussion Assessment and Cognitive Testing (ImPACT) in a sample of collegiate athletes. Methods A total of 163 college athletes (63.8% female) completed two pre-season baseline assessments two years apart as mandated by their athletic programs. Participants completed the first baseline prior to their first year of competition (M age=18.38, M edu=12.13). Individuals with baselines flagged as invalid by ImPACT were excluded. No diagnosed concussions occurred between baseline assessments. Results Pearson r correlations for ImPACT composite scores between assessments ranged from .32 to .70. Paired samples t-tests indicated significant differences between verbal memory (t(162)=-4.61, p<.001, MD=-3.6) and visual-motor speed (t(162)=-4.10, p<.001, MD=-1.5) at time 1 and time 2. Intraclass correlation coefficients (ICCs) with a “two-way mixed” model and “consistency” type indicated higher reliability for all composite scores. Visual-motor speed (mean ICC=.83; .76–.87 95% confidence interval [CI]) was the most reliable composite score, followed by visual memory (.66; 95% CI .53-.75), impulse control (.62; 95% CI .48–.72), verbal memory (.58; 95% CI .43–.69), and reaction time (.49; 95% CI .30-.62). The total symptom score showed moderate reliability (.63; 95% CI .50–.73). Conclusion Overall, ImPACT composite and total symptom scores appear to maintain relative long-term stability in a collegiate sample across a two-year period. ImPACT composite scores showed moderate to good reliability, and total symptom scores showed moderate reliability. Results of this study corroborate previous research indicating moderately stable ImPACT scores across a 2-year test-retest period in collegiate athletes (Schatz, 2010).