Answers are grounded in this paper and cite their sources.
This letter highlights methodological and statistical flaws in a trial comparing yoga and hydrotherapy in heart failure patients, emphasizing the need for appropriate non-parametric testing and per-protocol baseline comparisons.
To the Editor We read with great interest a recent article titled1 Effects of yoga versus hydrotherapy training on health-related quality of life and exercise capacity in patients with heart failure: A randomized controlled study. The authors determined whether yoga and hydrotherapy training had an equal effect on the health-related quality of life in patients with heart failure and compared the effects on exercise capacity, clinical outcomes, and symptoms of anxiety and depression between and within the two groups. Besides the intrinsic interest of the manuscript itself we would like to make the following comments that might not affect the overall conclusions of the paper but could affect the data presented and hence the accuracy of the provided reference tables. A total of 40 patients were randomized in the study and 30 patients (18 in the yoga group and 12 in hydrotherapy group). A major concern is about the date presented in Tables 2 and 3. The authors compare between and within the yoga group and hydrotherapy interventions effects on health-related quality of life, symptoms of depression, exercise capacity, lower limb muscle strength and clinical variables at baseline and after 12 weeks of follow-up. We think that the results shown in these tables fail to correctly interpret the results of their study. First, they compared both baseline (yoga: n=20 vs. hydrotherapy: n= 20) and follow-up (yoga: n=18 vs. hydrotherapy: n=12) groups. Then, they performed within-group comparisons. Although the data presented seem to be well analyzed, the authors should have compared between the baseline groups with the population that terminated both therapies, being that there could have been significant differences that could argue the results obtained in the follow-up. In the Methods section in statistical analyses it is indicated that univariate statistics on differences between baseline variables were calculated by the χ2 test, but some data shown in Table 1, based on observed frequencies less than 5, suggest that exact Fisher test is the more appropriate. In addition, some data in Tables 2 and 3 do not seem to have a normal distribution, so non-parametric tests should be used in the comparison (Mann–Whitney test in independent sample and Wilcoxon test in within-group). Finally, in the limitations section it is indicated that the sample size was small, which makes it difficult to generalize the findings. We think that the authors should have calculated the effect size through Cohen’s d. We cannot calculate the effect size because the baseline data from the yoga and hydrotherapy groups correspond to the subjects who were included in the study (n=20 in each group) and not those who completed the treatment. In conclusion, the authors’ valuable results could better be used as citable evidence if analyzed with suitable statistical analysis methods. The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
No takes yet. Share an insight, caveat, or question.
Rico‐Martín et al. (2017) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: