Comparing the Convergent and Concurrent Validity of the Dynamic Gait Index with the Berg Balance Scale in People with Multiple Sclerosis.

Background: Recent clinical guidelines for adults with neurological disabilities suggest the need to assess measures of static and dynamic balance using the Berg Balance Scale (BBS) and Dynamic Gait Index (DGI) as core outcome measures. Given that the BBS measures both static and dynamic balance, it was unclear as to whether either of these instruments was superior in terms of its convergent and concurrent validity, and whether there was value in complementing the BBS with the DGI. Objective: The objective was to evaluate the concurrent and convergent validity of the BBS and DGI by comparing the performance of these two functional balance tests in people with multiple sclerosis (MS). Methods: Baseline cross-sectional data on 75 people with MS were collected for use in this study from 14 physical therapy clinics participating in a large pragmatic cluster-randomized trial. Convergent validity estimates between the DGI and BBS were examined by comparing the partial Spearman correlations of each test to objective lower extremity functional measures (Timed Up and Go (TUG), Six-Minute Walk Test (6MWT), Timed 25-Foot Walk (T25FW) test) and the self-reported outcomes of physical functioning and general health using the 36-Item Short Form Health Survey (SF-36). Concurrent validity was assessed by applying logistic regression with gait disability as the binary outcome (Patient Determined Disease Steps (PDDS) as the criterion measure). The predictive ability of two models, a reduced/parsimonious model including the BBS only and a second model including both the BBS and DGI, were compared using the adjusted coefficient of determinations. Results: Both the DGI and BBS were strongly correlated with lower extremity measures overall as well as across the two PDSS strata with correlations. In PDDS ≤ 2, the difference in the convergence of BBS with TUG and DGI with TUG was -0.123 (95% CI: -0.280, -0.012). While this finding was statistically significant at a type 1 error rate of 0.05, it was not significant (Hommel's adjusted p-value = 0.465) after accounting for multiple testing corrections to control for the family-wise error rate. The BBS⁻SF-36 physical functioning correlation was at least moderate and significant overall and across both PDDS strata. However, the DGI⁻physical functioning score did not have a statistically significant correlation within PDDS ≤ 2. None of the differences in convergent and concurrent validity between the BBS and DGI were significant. The additional variation in 6MWT explained by the DGI when added to a model with the BBS was 7.78% (95% CI: 0.6%, 15%). Conclusions: These exploratory analyses on data collected in pragmatic real-world settings suggest that neither of these measures of balance is profoundly superior to the other in terms of its concurrent and convergent validity. The DGI may not have any utility for people with PDDS ≤ 2, especially if the focus is on mobility, but may be useful if the goal is to provide insight on lower extremity endurance. Further research leveraging longitudinal data from pragmatic trials and quasi-experimental designs may provide more information about the clinical usefulness of the DGI in terms of its predictive validity when compared to the BBS.