The Inter-Tester and Test-Retest Reliability of the Off-Field SCAT6 Assessment Tool In An Adult Population

No previously published repeatability and reliability data for The Sports Concussion Assessment Tool-6 (SCAT6) exists. We aimed to evaluate inter/intra-tester reliability of the off-field SCAT6 in a non-concussed adult population.

Inter-rater and Intra-rater reliability study design.

Single university site.

Twenty active adults (mean age: 27.55 ± 5.59 years) with no recent history of concussion (Concussive injury within past year).

Participants completed 3 SCAT6 tests on the same day, with 3 testers (Inter-rater testing). The same participants returned at 2 further time points to complete the remaining 2 SCAT6 tests with 1 tester (Intra-rater testing). Participants complete a total of 5 SCAT6 assessments in total across testers and time. Rater Background: Those completing the SCAT6 testing, our study rater team, comprised of 1 senior physiotherapist and PhD candidate, and 2 MSc Physiotherapy students. All raters were from Scotland, and had significant training in completing SCAT6 assessments.

Off-field SCAT6 Domain scores.

ICCs were used to establish inter and intra-rater reliability for continuous, ration and ordinal data components of the SCAT6. For nominal data sets, Fleiss's kappa was calculated. Kendall's W was used for non-parametric data. Percentage error scores were calculated for SCAT6 domains.

Inter-tester : Symptom number, severity, and dual-task scoring demonstrated excellent reliability (ICC = 0.981; 0.984; 0.913, respectively). Total concentration score was found to have good reliability (0.827). Dual-task errors (0.398), Total mBESS (0.199), and Month recall all returned poor scores (k = 0.191). Intra-tester : Dual tasking was the only domain to report excellent reliability (ICC = 0.943). Symptom number (0.868), severity (0.831), total concentration (0.787), total mBESS (0.813), and time tandem gait (0.834) yielded good reliability scores. Dual-task error testing returned poor reliability scores (Kendall's W = 0.001). All remaining domains yielded moderate reliability. Percentage error rates ranges from 3% to 100%, demonstrating the variability between scores yielded for non-concussed individuals completing the same SCAT6 domain tests.

SCAT6 ICC results reported good-excellent reliability for 4 and 6 domains, out of 13 domains, for inter-tester and intra-tester reliability, respectively. Notably, the domains which relied on tester error scoring yielded poor reliability results. Percentage error highlighted the failure of the SCAT6 to provide consistent domain score results in this population.