Spss statistical package version 24

Descriptive statistics are presented as medians and interquartile ranges (IQR), and illustrated using Box plots. Most data were not normally distributed and thus the main analyses were non-parametric. The Mann–Whitney U test and the Chi-square test were used for between-group comparisons. Changes from baseline to follow-up at 24 weeks within groups were analyzed with the Wilcoxon Signed Rank test. For all tests p ≤ 0.05 was considered statistically significant.
Supplementary parametric analyses were performed, from which we present means, standard deviations (SD), and p-values from paired and independent samples t-tests.
The IBM SPSS statistical package version 24 was used for all statistical analyses.

Data are median and interquartile range (IQR). Categorical variables are presented as numbers (percentages). Continuous values were compared by the Mann-Whitney or Kruskal-Wallis test and categorical variables by the χ² test. Survival curves were constructed using Kaplan-Meier estimates and were compared by the log-rank test. Multivariate Cox regression analyses were performed with mortality as the outcome variable considering the following baseline parameters: sex, BMI, GFR, NYHA class III or IV, prior cardiac decompensation, DM, COPD, STS score, prior myocardial infarction, atrial fibrillation, prior ICD implantation, EF, LV myocardial infarction, SVI, MPG, mitral regurgitation ≥II°, tricuspid regurgitation ≥II°, SPAP, severe AVCd and the use of a balloon-expandable valve. Univariate predictors with P≤0.1 were entered into multivariate Cox regression analysis. The HF classification was not entered into the analysis because of high co-linearity. All statistical analyses were performed using the SPSS statistical package version 24 (IBM Corp., Armonk, NY, USA).

Descriptive statistical analysis (including means, SDs, and frequencies) and intraclass correlation coefficient (ICC) estimates were calculated using SPSS statistical package version 24 (IBM Corp Released 2016, IBM SPSS Statistics for Windows, Armonk, NY, USA).
To assess the interrater reliability for all reported metric items, ICC estimates and their 95% CIs were calculated for each rater pair based on a one-way random-effects model with absolute agreement.
To assess the interrater reliability for all reported dichotomous items, Gwet AC1 statistic was calculated using formula 4.1 of Gwet [64 ] through the WINPEPI software developed by Abramson [65 (link)]. Gwet AC1 statistic is a chance-corrected measure of the extent of agreement between raters but has been recommended for use because it is less influenced by differences in individual rater tendencies to give positive ratings and by differences in the prevalence of the response categories, thus making it more robust and less biased compared to other metrics [66 (link)-68 ].

The SPSS statistical package version 24 (IBM Corp, Armonk, New York, United States) was used to analyze the data.
To conduct the intention-to-treat analysis by multiple imputations (MIs) of missing values, the data from all 55 participants were used following the Sterne et al. guidelines [58 (link)]. Our missing data were classified as missing at random. The SPSS software package was used for the MIs of data.
Nonparametric tests were conducted because the dataset was not large and the data were not always Gaussian. To explore the effectiveness of the exergame-based intervention, the Mann–Whitney U-test was conducted to examine the differences between groups for each variable. Moreover, within-group comparisons were conducted by the Wilcoxon signed-rank test.
The Benjamini–Hochberg false discovery rate correction for multiple comparisons was applied in each comparison to avoid Type I errors. The partial eta-squared effect size was reported for each statistical test [59 (link)]. According to Cohen [60 ], effect sizes could be classified as small (0.01 ≤ η² < 0.06), medium (0.06 ≤ η² < 0.14), and large (η² ≥ 0.14).
Furthermore, Spearman’s rho correlation analyses were used to evaluate the relationship between pVO2, MMSE, and brain structure volumes.

Estimation of a sample size was based on data from a previous study at MetroHealth [28 ]. A sample size of 20 in each group was sufficient to achieve 80% power to detect a 15% change in FFM from 2.2 to 2.5 (SD 0.3) at 5% significance (Power Analysis and Sample Size Statistical Software 2012). The data are reported using the following descriptive statistics: mean ± standard deviation (SD) for interval, parametric data; media with the interquartile range for nonparametric interval or ordinal data and percentages for nominal data. Parametric interval data were defined as having both the skewness and kurtosis between a −3 and +3. The patients were analyzed according to the group they were randomized to. The following statistical tests were used to show group differences: pooled variance or separate variance t-test depending on the F value of their respective standard deviations, Mann–Whitney U test for nonparametric interval or ordinal data and Chi square or Fishers exact test for nominal data. The data were analyzed using SPSS Statistical Package, Version 24 (IBM Inc., Armonk, NY, USA). Statistical significance was defined a priori as a p < 0.05 (two-tail).