Inter-rater Reliability And Construct Validity Of The Dynamic Gait Index In Older Adults

Abstract

The Dynamic Gait Index (DGI) is one of the most challenging gait mobility tests available for the older adult population. Research demonstrates that scores on the DGI discriminate community-dwelling elderly who report falls from those who do not report a fall history. Reports indicate that the DGI has good inter-rater reliability within a group of instrument-specific trained physical therapists. No studies have been performed on interrater reliability of the DGI as a single test in the population of community-dwelling older adults or among untrained raters of varying clinical experiences. PURPOSE The purpose of this investigation is to examine the interrater reliability of the DGI as applied to community-dwelling elderly by raters with backgrounds in either physical therapy or Kinesiology. METHODS Fourteen community-dwelling adults (age range: 51–93; mean: 75.4 years) performed the DGI and the reduced Continuous Scale Physical Function Performance Test (CS-PFP10). The DGI tests were videotaped and subsequently analyzed by five raters. Three raters were licensed physical therapists, one rater was a PhD in Kinesiology, and one rater was a first year graduate student with a BS in Kinesiology. A repeated-measures ANOVA was used to derive intraclass correlation coefficients (ICC) for the purpose of assessing inter-rater reliability of the DGI scores. Pearson correlation was used to compare the DGI scores from the five raters to the subscales and total score of the CS-PFP10. RESULTS The estimated inter-rater reliability within the individual eight tasks ranged from r = .69 to .97 and the intraclass r for the total DGI score was .90. The DGI scores also correlated well with the lower body strength, endurance, and balance and coordination subscales, as well as the PFP10 total score (r = .52 −.78, P<0.10 across all testers). In addition, the magnitudes of the associations were marginally stronger in cases when physical therapists provided the DGI rating as compared to the raters without a physical therapy background. CONCLUSIONS The strong ICC of the total DGI score implies that the variation in measurements on the total score is more likely due to the participants' abilities to execute the tasks versus conflicting scoring ability of the testers. While these findings generally provide evidence that the DGI merits attention as an appropriate assessment tool that can be applied in various clinical settings, the stronger degree of association between DGI and the ADL-based CS-PFP10 scores among physical therapists suggests that test-specific training may further improve the inter-rater reliability of this instrument.