Wang et al. (1) recently questioned one aspect of our statistical analysis and commented on conclusions drawn in our recent study exploring the acute effects of sole vibration on gait and balance in people with diabetic peripheral neuropathy (DPN) (2). We value their feedback and intend to clarify several points.
Following their concerns, the data were meticulously reanalyzed and verified using MATLAB and SPSS, and we confirmed that the original results were correct. Wang et al. (1) asserted that larger group average differences and smaller SD often indicate a higher likelihood of significant differences. However, for a paired t test, the critical values are the mean of paired differences within each participant (Xdiff) and SD of differences (Sdiff) rather than the group means and SD. The formula for the paired samples Student t test is t = Xdiff/(Sdiff/√n). Wang et al. focused on the group means and SD, which do not directly reflect the variability of paired responses to each condition. The vibratory conditions (Vcs) identified as significantly different for stair ascent were Vc2, Vc4, and Vc6, for which the Sdiff for gait speeds were 0.17, 0.07, and 0.15 m/s, respectively. These Sdiff values were lower than those of nonsignificant conditions Vc1, Vc3, Vc5, and Vc7, with Sdiff values of 0.19, 0.21, 0.23, and 0.19 m/s, respectively. Therefore, despite higher Xdiff values in some nonsignificant conditions, higher Sdiff values (indicating greater variability in differences within the group) resulted in lower t values, demonstrating a lower probability of significant differences among groups (this finding is also consistent with the descent data). This higher variability in paired differences, despite lower overall group variability, explains the observations of Wang et al. and validates our statistical outcomes as initially reported. The differing responses among individuals remain consistent with the interpretation of our findings in the article that sole vibration improves several gait variables.
Additionally, Wang et al. (1) commented on some conclusions on gait speed improvement following vibration. We reemphasize here (as was made clear in the article) that our findings do not identify a relationship between acute gait speed improvement with vibration and the risk of falling. Within the discussion, we state that prior studies have identified gait speed as a well-established predictor of fall risk. Later, we discussed the need for future clinical trials to investigate the long-term effects of vibrating insoles on fall risk.
While we establish the acute effects of vibration in people with DPN, we agree that the long-term impacts on peripheral sensation and physical function remain unknown, as does the mechanism(s) by which vibration benefits postural control, including the common finding that subsensory frequencies are beneficial, highlighting it is not perception of the vibration itself that is beneficial. As stated, we agree with Wang et al. that further studies are required. Our research group is conducting a clinical trial to test the long-term effects of vibration on neuropathic signs and symptoms and physical function in people with DPN, and we will report on this study in due course.
Article Information
Funding. This research was funded by an EU small- and medium-sized enterprises instrument grant (ID 879948).
Duality of Interest. No potential conflicts of interest relevant to this article were reported.
Handling Editors. The journal editor responsible for overseeing the review of the manuscript was Steven E. Kahn.
