Measurement Agreement and Reliability of the 6-m Gait Speed Test in Adult Population: The ADULT-FIT Project

We aimed to examine: (i) measurement agreement between photoelectric cell system and manual stopwatch timing; (ii) inter-rater agreement between trained and untrained raters using manual stopwatch timing; and (iii) reliability of the 6-m gait speed test to assess motor fitness in adults, according to sex, age, and physical activity levels. 410 healthy adults (203 females) aged 18-64 years participated in this study. The time required to perform the 6-m gait speed test was recorded simultaneously by a trained (R1) and an untrained (R2) rater (inter-rater reliability), and by photoelectric cells (measurement agreement between timing methods). The test was repeated under the identical conditions (test-retest reliability) 48-72 h later. Systematic error between photoelectric cells and R1 was close to zero (0.015 s), whereas a larger difference was observed for R2 (∼0.2 s). The corresponding concordance correlation coefficients (CCC) were 0.966 and 0.861, respectively. There was a systematic error between R2 and R1 of ∼0.2 s, with a CCC of 0.872. Test-retest systematic error was -0.047 s (p < 0.01), with an effect size of 0.067. The intraclass correlation coefficient was 0.984, and all error measurements reported low values, except the coefficient of variation that was 14.386. No heteroscedasticity was observed (p > 0.05). Measurement agreements and reliability were not influenced by sex, age groups and physical activity levels. Manual timing by R1 demonstrated high agreement with photoelectric cells and high test-retest reliability for assessing 6-m gait speed test in adults. Rater training is strongly recommended to minimize measurement error.