Table 1_Optimal exercise prescription for balance after stroke: a Bayesian dose–response n

A Bayesian dose-response network meta-analysis of 42 randomized controlled trials investigating the relationship between exercise dose and balance improvement in stroke survivors. The dataset, published by Chunbai Xue Yang in April 2026, quantifies effects using the Berg Balance Scale and models dose-response curves for various exercise modalities. Results indicate a non-linear relationship with a minimum effective dose and modality-specific peak benefits.

Use Cases

• Modeling non-linear dose-response curves for exercise interventions based on MET-min/week data.
• Comparing the efficacy of different exercise modalities (e.g., resistance, aerobic, water-based) for post-stroke balance.
• Estimating minimum effective doses and peak benefit ranges for clinical program design.
• Conducting secondary Bayesian meta-analyses on pooled effect sizes (Hedges' g) from included trials.

Strengths

• Analysis is based on 42 randomized controlled trials, providing a substantial evidence base.
• Employs Bayesian model-based methods for dose-response estimation, a sophisticated statistical approach.
• Includes specific numerical findings, such as a minimum effective dose of ~270 MET-min/week and peak doses for modalities like resistance training (~666 MET-min/week).

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• The dataset is very small (34.1 KB), indicating it likely contains summary or analysis results rather than raw trial data.

Provenance

Source

Chunbai Xue Yang via figshare

Collection Method

Systematic review and Bayesian meta-analysis of trials from PubMed/MEDLINE, Embase, Web of Science, Scopus, CENTRAL, and Chinese databases.

Time Range

Trials from database inception to December 31, 2025.

Freshness

Last updated 2026-04-13 04:31:17

Geography

Likely global, based on inclusion of international and Chinese databases.