R software

Statistical sofrware: All statistical analyses were performed using R software (version 4.04) and IBM SPSS Statistics (version 26).
Regression model: Univariate and multivariate logistics regression models were used to analyze the factors influencing the clinical outcomes of severe HFMD. The “Forward” method was used to select the factors into the logistic regression model. The influencing factors were first analyzed by the univariate analysis, then the variables with differences (p < 0.1) were included in the multivariate logistic regression model. Variables with significant differences (p ≤ 0.05) are considered in the final interpretation. All the comparisons were two-sided, and the p-value of < 0.05 was considered significant. The OR and 95%CI were compared between the survival group and the death groups.

The mean HRQoL questionnaire scores and the prevalence of associated symptoms will also be determined, together with their 95% confidence intervals. HRQoL will be compared with the baseline data using statistical analysis tests for paired data in order to determine the factors associated with a better HRQoL. We will test the normality of the quantitative variables using the Kolmogorov–Smirnov test and will apply Student t-tests or Mann–Whitney U tests as appropriate to compare numerical parameters between two groups. An analysis of variance or the Kruskall–Wallis test will be used to compare more than two groups. The association between qualitative variables will be contrasted using the chi-square statistic and associations between quantitative variables will be studied with Pearson correlation coefficients and Spearman’s rho. Linear regression and multiple logistic models will be used to control for the effect of several variables on HRQoL, perceived quality of care, and patient function. All the statistical analyses will be carried out with SPSS (version 21.0; IBM Corp., Armonk, NY) and R software for Windows.

The clinicopathological characteristics were summarized with descriptive statistics. Survival probability was estimated using the Kaplan–Meier method. The comparisons between variables were performed using the t-test or chi-square test. The association between variables and cumulative incidence of DM was identified with univariable analysis using Fine-Gray competing risk regression, and multivariable analysis was conducted with a Cox proportional hazards model. The cutoff value for the risk score of the high- and low-risk group was the same as the DM risk score model, and a concordance index (c-index) using Cox proportional-hazards regression was conducted to estimate the model. The inverse probability of treatment weighting (IPTW) [18 (link)] was used to minimize the bias of the clinical features between the pairwise groups. The statistical analyses were performed by IBM SPSS Statistic (version 26.0, IBM Corp, Armonk, NY, USA) and R software (version 4.0.5, Vienna, Austria). A two-tailed p-value of <0.05 was considered statistically significant.

Sample size was calculated with AKI as a composite outcome using multivariate Cox Proportional Hazard Analysis. A total of 60 subjects (nearly 30 in each group) were estimated, expecting that 30 out of them meet one or other composite AKI criteria. With an alpha error of 5% and power of 80%, we assumed that the hazards of AKI in Plasma-Lyte will be 0.3 times lower than saline and the hazard ratio is constant throughout the study.
Data was analyzed as per intention-to-treat principle. Unadjusted chi-square test was used to analyze the differences in primary outcome. Absolute and relative risks with 95% confidence intervals were calculated. Survival analysis for time to resolution of AKI and DKA were compared with log-rank test. A Cox proportional hazards regression model was used to evaluate the influence of potential confounders on outcome as age, new onset DKA, and severity of DKA. Quantitative variables with normal and non-normal distribution were expressed as mean (with standard deviation) or median (with inter-quartile range) respectively. Unpaired Student’s t test or Wilcoxon rank-sum test was used for intergroup comparisons. General linear model repeated measure ANCOVA was used to compare the trends of continuous variables over time. A P value (two-tailed) < 0.05 was taken as significant. IBM SPSS Version 21 and R software were used for data analysis.