Supplementary Materials for Effects of Different Neuromuscular Training Modalities on Balance Performance in Older Adults: A Systematic Review and Network Meta-Analysis

DescriptionThis supplementary package contains all supporting materials for the systematic review and network meta-analysis titled "Effects of Different Neuromuscular Training Modalities on Balance Performance in Older Adults." The files are organized to enhance the methodological transparency and result interpretation across various outcome domains. The included figures, tables, and data are as follows:FiguresFigure S1. Risk of bias summaryVisualizes the methodological quality and item-level judgments across all included randomized controlled trials using the RoB2 tool.Figure S2. Predictive interval plot of TUGTShows the 95% predictive interval and pooled estimate for the Timed Up and Go Test (TUGT), reflecting expected treatment effects in future studies.Figure S3. Predictive interval plot of WTDisplays the predictive interval for Walking Time (WT) as an outcome, accounting for between-study variability.Figure S4. Predictive interval plot of BBSDepicts predictive distributions for the Berg Balance Scale (BBS), improving interpretation of translatability to new populations.Figure S5. Leave-one-out sensitivity analysis of TUGT before exclusion of high-risk studiesIdentifies influential studies contributing to instability in TUGT estimates before bias-based exclusions.Figure S6. Leave-one-out sensitivity analysis of TUGT after exclusion of high-risk studiesDemonstrates improved robustness and reduced heterogeneity in TUGT results after excluding high-risk studies.Figure S7. Subgroup analysis of TUGT by training modalityCompares the effects of different neuromuscular interventions (e.g., ST, NT, WBVT, BT) on TUGT performance.Figure S8. Leave-one-out sensitivity analysis of WT before exclusion of high-risk studiesHighlights individual studies influencing WT outcome estimates prior to bias-related exclusions.Figure S9. Leave-one-out sensitivity analysis of WT after exclusion of high-risk studiesDisplays consistent results and minimized variance following data refinement.Figure S10. Subgroup analysis of WT by age groupsExplores whether the effect of interventions on WT differs across older adult subgroups categorized by age.Figure S11. Subgroup analysis of WT by health statusEvaluates WT outcomes stratified by participants’ baseline health conditions (e.g., healthy, post-stroke, osteoporotic).Figure S12. Leave-one-out sensitivity analysis of BBSExamines the influence of each study on pooled BBS estimates, confirming the robustness of results.Figure S13. Subgroup analysis of BBS by age groupsInvestigates age-specific differences in the effect of interventions on BBS scores.Figure S14. Subgroup analysis of BBS by health statusAssesses BBS improvements across populations with varying health statuses, such as frail or healthy elders.Files and TablesData.xlsx.This Excel file includes the complete dataset extracted from the included studies, detailing sample sizes, means, standard deviations, intervention types, and outcome measurements (TUGT, WT, and BBS). It serves as the foundational input for all statistical analyses, ensuring traceability and reproducibility of effect size calculations and subgroup classifications.File S1. STATA Analysis Code and Dataset for NMAContains the full statistical code and dataset used for the Bayesian network meta-analysis, including model scripts and effect-size inputs.Supplementary Materials A. RoB2Detailed documentation of the risk-of-bias evaluation process using the Cochrane RoB2 tool, including decision paths and justifications.Supplementary Materials B. PRISMA 2020 ChecklistCompleted PRISMA 2020 checklist mapping each reporting item to its location in the manuscript.Table S1. The complete search strategy for the databasesLists all search terms, Boolean operators, filters, and database platforms used during literature retrieval.Table S2. Assessment of Loop Inconsistency in the Network Meta-AnalysisReports the results of inconsistency checks using node-splitting and design-by-treatment interaction models.Table S3. Egger’s Test for Publication Bias Across Outcome MeasuresPresents the results of Egger’s regression test to detect small-study effects in the included literature.This supplementary content is provided to ensure analytical reproducibility, facilitate peer evaluation, and promote transparent synthesis of evidence regarding neuromuscular training effects in older adult populations.