Dear Editor, Recently, Lee et al1 published a prospective, randomized, single-blind study investigating the effects of electromyography (EMG)-driven robotic training on functional recovery and scar outcomes in patients with post-burn hand skin grafts and peripheral nerve injury. This study represents an important step toward integrating EMG-controlled robotic rehabilitation into the treatment of post-burn hand dysfunction, a field still lacking in high-quality clinical evidence. Nonetheless, several methodological issues merit further discussion. Additionally, we ensured that all content in the article complies with the TITAN Guidelines 20252. Although described as a “randomized, single-blind” trial, the study used group allocation based on “admission order followed by computer assignment,” which raises concerns about allocation concealment and potential selection bias. Moreover, single blinding, where only the assessor was unaware of group assignment, may not fully eliminate expectation bias, since both participants and therapists were aware of the intervention received. Given the motivational and behavioral components inherent to robotic training, the absence of participant blinding could have influenced performance-based outcomes such as the Jebsen-Taylor hand function test (JTT) and Purdue Pegboard test (PPT) scores. The authors reported using G*Power for sample size estimation, yet only 39 participants were included in the final analysis. Considering the multiple outcome variables assessed, range of motion, muscle strength, dexterity, and various scar parameters, this sample size may have been underpowered to detect small to moderate effects, particularly in scar-related measures. Borderline P-values (e.g. scar thickness P = 0.11) may therefore reflect a type II error rather than a true absence of effect. Increasing the sample size or providing an effect size justification would strengthen the reliability of the conclusions. Another limitation is that all participants were male, which restricts the generalizability of the findings to the broader burn survivor population, especially women, who may exhibit different scar physiology and neuromuscular recovery patterns. In addition, although the inclusion criteria required burns involving ≥50% of the dominant hand, the extent and severity of nerve injury varied considerably. Without stratification by injury severity or post-graft duration, within-group variability could have diluted the treatment effect. While both groups received equal total therapy time, the nature of the interventions differed substantially: the experimental group engaged in technology-enhanced, feedback-driven training, whereas the control group performed conventional exercises. This design does not account for possible “novelty” or engagement effects introduced by robot interaction. Including a sham-robot or interactive feedback control group would better isolate the specific efficacy of the robotic system. Furthermore, the follow-up period was limited to 12 weeks, which is insufficient to assess long-term maintenance or the influence of scar maturation. Since hypertrophic scar remodeling may continue for 12–18 months post-burn3,4, short-term biomechanical gains may not necessarily translate into sustained clinical benefits. Future studies should extend follow-up durations and incorporate patient-reported outcome measures (e.g. Disabilities of the Arm, Shoulder and Hand (DASH), Michigan Hand Outcomes Questionnaire (MHQ))5 to better capture functional improvements in real-world rehabilitation settings. In summary, Lee et al1 provide valuable preliminary evidence supporting EMG-driven robotic rehabilitation for post-burn hand dysfunction. However, methodological limitations, particularly in sample size, control design, and potential overinterpretation, warrant cautious interpretation. Future multicenter trials employing stratified randomization, blinded outcome assessment, and long-term follow-up are recommended to confirm both the clinical and physiological benefits of EMG-controlled robotic therapy in this patient population.
Gao et al. (Fri,) studied this question.