Stata ver 15

Data were analyzed using Excel ver. 16.52 (Microsoft Corp.) and STATA ver. 15.1 (Stata Corp.). Descriptive statistics summarized the distribution, central tendency, and dispersion of responses. For continuous variables, the distribution of data was first evaluated for normality using the Shapiro-Wilk test. However, all the variables did not pass the Shapiro-Wilk test, we additionally checked a q-q plot, which did not show marked deviation from linearity. We therefore assumed that normal distribution assumption for parametric test was not violated, and decided to apply linear mixed effect model, which was made with times (T1, T2, T3, and T4) as independent fixed factors, and individual patients as random effects. Significance was set at α=0.05.

Data analysis was carried out in Stata, ver.15.1 (StataCorp LLC, 4905 Lakeway Drive, College Station, Texas 77,845, USA) and AgreeStat 2015.1 for Excel Windows/Mac (Advanced Analytics, LLC. PO Box 2696, Gaithersburg, MD 20,886–2696, USA).
In the statistical analysis, inter-rater reliability was determined for nominal data by calculating percent agreement, and a change corrected agreement coefficient: Gwet’s AC₁ (unweighted) and AC₂ (weighted) for respectively pair-wise raters and for three raters overall [29 ]. Percent agreement and chance-corrected agreement coefficients (except for marginal totals) were reported with 95% confidence intervals. Proportions of absolute agreement were calculated to evaluate the precision of the strength of reliability. Finally, an additional probabilistic method for benchmarking to an interpretation scale was used and presented as the cumulative probability exceeding 95% for the coefficient to fall into one of the following intervals using the benchmark scale of Landis and Koch: < 0.00 poor; 0.00 to 0.20 slight; 0.21 to 0.40 fair; 0.41 to 0.60 moderate; 0.61 to 0.80 substantial, and 0.81 to 1.00 almost perfect [30 (link)]. This method allows for a direct comparison between different agreement coefficients and to what extent they are paradox-resistant, i.e., subject to instability if ratings had very low or very high prevalence.

Variation in miRNA expression data were analyzed by ANOVA with variance partitioned between trimester and clinical status. Two statistical approaches to analyse the miRNA sequencing data were used. First, data was analyzed by two-way ANOVA (with LSD post-hoc testing to discriminate among the means) to determine the interaction between gestational age and GDM. Second, data was also categorized into NGT and GDM, and a one-way ANOVA with post-hoc test on each was analyzed to determine the overlap. Non-normally distributed data were logarithmically transformed before analysis. All analyses were performed using the R statistical software. Statistical significance was denoted by False Discovery Rate (FDR)-adjusted p < 0.05.
Quantitative miRNA expression data was analyzed using commercially available software (STATA ver 15.1, STATA Corp. College Station, TX, USA). Shaprio-Wilk [27 (link)] tests were used to assess the normality of data distributions. Non-parametric statistical tests were used where data distributions significantly deviated from normality. Between group differences were assessed by Two-sample Wilcoxon Rank-Sum tests. [28 (link)] Variation in miRNA expression within case and control cohorts was assessed using a Panel Data Analysis and Random Effects Generalized Least Squares (RE-GLS) models. [29 ] Statistical significance was ascribed when p < 0.05.

Kolmogorov-Smirnov test was used for the assessment of the normality of distribution. Continuous variables were presented as mean with standard deviation for normal distribution or median with interquartile ranges (IQR) for normal distribution. Student’s t-test or Wilcoxon rank-sum test were used as appropriate. Categorical variables were presented as counts (percentages) and compared using the chi-square test. We used a locally weighted smoothing (Lowess Smoothing) technique to explore the crude relationship between normalized CVP/ECVP and mortality. We divided the overall population into the high and low normalized ECVP load groups base on the median of normalized ECVP load. Kaplan-Meier survival analysis was performed and tested by the Log-Rank test. Multivariate logistic regression analysis was performed using foreword procedures with factors showing p < 0.20 in univariate analysis. STATA (ver. 15.1, StataCorp., TX, USA) was used for data analysis. All reported p-values were two-sided, and a p < 0.05 was considered significant.