Statistical quality in RCTs of technology-based rehabilitation after spinal cord injury: A CHAMP-Based Meta-Research Study

Abstract

STUDY DESIGN: Cross-sectional meta-research study. OBJECTIVES: Technology-based rehabilitation (TBR) is increasingly used in spinal cord injury (SCI), with a growing number of randomised controlled trials (RCTs) evaluating its effectiveness. However, inadequate statistical reporting may limit the interpretation, reproducibility, and clinical applicability of findings. This study aimed to evaluate the statistical quality of RCTs investigating TBR interventions in individuals with SCI. METHODS: MEDLINE, CINAHL Complete, Cochrane CENTRAL, PEDro, and Web of Science were searched from 1996 to October 2025. Eligible studies were superiority parallel-arm RCTs involving adults with traumatic or non-traumatic SCI evaluating TBR interventions and explicitly reporting adherence to the Consolidated Standards of Reporting Trials (CONSORT). Statistical quality was assessed using the Checklist for Statistical Assessment of Medical Papers (CHAMP), while methodological quality was evaluated using the Risk of Bias (RoB) 2 tool. RESULTS: Eight RCTs were included. Overall, CHAMP compliance was 57.9%, with over 40% of items rated as unclear or not fulfilled. Compliance was highest in the design domain (64.6%) and lower in the analysis (58.8%), interpretation (58.9%), and reporting (50.0%) domains. CHAMP scores ranged from 0.30-0.83. According to RoB 2, one study (12.5%) was judged at low risk of bias, five studies (62.5%) raised some concerns, and two studies (25.0%) were judged at high risk of bias. CONCLUSIONS: Statistical quality in RCTs of TBR in SCI is frequently suboptimal despite reported adherence to CONSORT guidelines. Greater transparency in statistical reporting is needed to improve the interpretation, reproducibility, and clinical applicability of rehabilitation research.