Abstract

Dear Editor:
I
Table 4 of Rubik's study reports the results of an independent sample t-test comparing LF/HF after intervention with the Heart+ App versus after intervention with a sham app. However, the degrees of freedom of the t-test was reported as 118, which indicates that each of the 60 Heart+ App observations, and each of the 60 control observations, were treated as independent (i.e., sample size was treated as 60 per intervention). This is incorrect, as measurements were made in triplicate and in reality, there were only 20 independent observations (20 subjects) per intervention. The difference between replicate measurements and independent repeats and the statistical implications of this has been covered in detail by Vaux et al. 3
Hence, using the reported means and variances, as well as the corrected sample size of 20 in each group, an independent t-test of the data demonstrates no significant difference in LR/HR after intervention with the Heart+ App compared with the sham app (mean difference = −1.29, 95% confidence interval = −2.96 to 0.38, p = 0.126). 4 Given the error in table 4, we also question whether the paired t-tests reported in table 2 have been performed correctly. It should further be noted that since this is a crossover trial, a paired analysis for table 4 would have been more apt than the independent t-test chosen by Rubik.
Contrary to the conclusions made by Rubik, the reported results fail to prove that the tested “subtle energy technology” affects autonomic nervous system balance.
Footnotes
Authors' Contributions
A.C.Q.L. and C.C.W.L. conceived the study and wrote/reviewed the article. A.C.Q.L. performed the statistical analysis.
