Responses to the Comments on “Effects of Cognitive Versus Mind–Motor Training on Cognition and Functional Skills in the Community-Dwelling Older Adults”

The authors sincerely acknowledge the experts’ comments and appreciate the insights and explanation of sample size estimation in two-group randomized controlled trials (RCTs). We agree that the estimated sample size for the study objective was incorrect. It is noteworthy that the sample size for the mean change cannot be estimated without standard deviation (SD), and the SD cannot be identical for two different outcome measures.

We decided that, among older adults, if a novel mind motor training (MMT) strategy influences a one-point change in the General Practitioner Assessment of Cognitive Scale (GPACOG) or any of the eight functional skills in the Lawton Instrumental Activities of Daily Living, compared to another rehab training, this absolute difference is clinically relevant. Only the GPACOG, the primary outcome measure, was considered for the initial sample size calculation. Based on the pilot clinical data in older adults, the mean change(SD) of the GPACOG post-MMT was 1.46 (0.78) and usual care (control) was 0.45 (0.69) points. Since many participants scored 10–14 out of a maximum possible score of 15 in GPACOG at baseline, we concluded that a one-point change, that is, the least possible change, is clinically significant. The institutional research committee recommended cognitive training (CT) as a comparative arm instead of usual care and also suggested including 20 subjects per arm, due to the time-bound research nature of the postgraduate dissertation. In this line, the difference of GPACOG between groups (MMT and control) was erroneously reported in the paper and was written in a vague way.

Nevertheless, this is the first preliminary RCT comparing the efficacy of MMT and CT on cognition and functioning in older adults (n = 40); the study results shall be deliberated for the sample size estimation and analysis. It is true that when the detection of a small change is fixed as the target, the required sample size is large. The GPACOG showed a mean change (SD) of 1.75 (1.29) and 1.5 (1.36) after MMT and CT, respectively, with an effect size (d) of 0.2. Setting a 5% level of significance and 80% power in the formula n = [2 (Zα + Z1−β)² σ²] / d²,1 the necessary sample size would be 349 per arm. With a 10% dropout rate, a study would require 385 participants per arm to report with any amount of confidence whether a change occurs between the MMT and CT. Besides the effect size and SD critical for the sample size calculation, the clinicians should consider the severity of older adults’ cognitive decline, practice intensity, and responsiveness of outcome measurements. The clinical inference of this study is that a novel MMT strategy shall impact the cognitive and functional skills as similar to a well-established CT regime in healthy older adults, warranting well-powered RCTs with robust outcome measures.