Control strategies used in lower limb exoskeletons for gait rehabilitation after brain injury: a systematic review and analysis of clinical effectiveness.

This review did not treat new participants, so its conclusions depend on published studies rather than direct patient data. The clinical comparison included only 73 studies, and the stroke conclusions rested on 57 studies, so many controller comparisons were based on few studies. From the 57 stroke studies, 5 included acute stroke, 13 included subacute stroke, and 42 included chronic stroke; only two studies focused on traumatic brain injury. Cerebral palsy studies could not be compared across controllers because participants were often pooled across GMFCS levels and the few studies used the same control strategy. Outcome measures were highly variable; no outcome was used in more than 50% of studies, and no study reported exoskeleton setup time. Only 12.58% of studies did follow-up, on average four months after the last intervention, so long-term effects are unclear. The average level of evidence was low, because 66.04% of studies were observational. Training intensity varied widely, and higher improvements sometimes occurred with higher min/week, so it is hard to separate controller effect from dose.

The supplied text lists funding from Agència de Gestió d’Ajuts Universitaris i de Recerca, Agencia Estatal de Investigación, Swiss National Science Foundation and Dutch Research Council Talent Program. It does not provide an author conflict-of-interest declaration.

Do not read the reported percentage improvements as proof that one exoskeleton controller is clinically superior. The review included few studies per strategy, varied training intensity and outcome measures, and could not compare controllers in cerebral palsy or traumatic brain injury.