Test-Retest Reliability and the Effects of Exercise on the King-Devick Test

Objective: To determine the test-retest reliability and the influence of exercise on King-Devick (K-D) test performance. Design: Crossover study design. Setting: Controlled laboratory. Participants: Participants consisted of 63 (39 women and 24 men) healthy, recreationally active college students who were 21.0 + 1.5 years of age. Independent Variables: Participants completed the K-D test using a 2-week, test-retest interval. The K-D test was administered before and after a counterbalanced exercise or rest intervention. Reliability was assessed using testing visits (visit 1 and visit 2) as the independent variables. Exercise or rest and time (baseline, postintervention) were used as independent variables to examine the influence of exercise. Main Outcome Measures: Intraclass correlation (ICC) coefficients with 95% confidence intervals were calculated between visits to assess reliability of K-D test completion time. A repeated-measure 2 x 2 analysis of variance (intervention x time) with post hoc paired t tests was used to assess the influence of exercise on K-D test performance. Results: The K-D test was observed to have strong test-retest reliability [ICC2,1 = 0.90 (0.71, 0.96)] over time. No significant intervention-by-time interaction (P = 0.55) or intervention main effects (P = 0.68) on K-D time were observed. Mean differences of -1.5 and -1.7 seconds (P < 0.001) were observed between baseline and rest and exercise interventions for K-D test performance, respectively. Up to 32% (20/63) of participants were observed to have a false-positive K-D test performance before and after each intervention. Conclusions: Although strong test-retest reliability coefficients were observed using clinically relevant time points, a high false-positive rate warrants caution when interpreting the K-D test.