Two architectures, different compromises
Rehabilitation robots can support repeated movement of the arm, hand or lower limbs. An exoskeleton aligns mechanical segments with parts of the body and applies assistance around selected joints. An end-effector system interacts at a more distant contact point, such as a hand grip or footplate. Some gait systems also provide body-weight support. The design influences setup, movement freedom and which impairments a therapist can address.
Exoskeletons may offer more direct control of particular joint motions, but alignment and fit require care. End-effector systems can allow quicker transitions between some users, while providing less direct control over intervening joints. Neither architecture is inherently superior. Evaluate the exact movement task, the intended patient population and how the device supports the therapist’s treatment plan.
Patient selection comes before programming
Suitability depends on the person and the device’s instructions. Assessment may include joint range, skin condition, body dimensions, bone health, cardiovascular tolerance, cognition, pain and the ability to communicate discomfort. Spasticity, contractures or changes in sensation can affect safe use. These considerations need qualified clinical assessment; a list on a website cannot establish eligibility.
Transfers and fitting are part of treatment time, not administrative details. Consider whether hoists, additional staff or wheelchair access are needed. Harnesses and cuffs can create pressure or friction, so therapists need appropriate checks and clear stop criteria. A device that fits a demonstration participant comfortably may still be unsuitable for a substantial part of a service’s caseload.
More repetitions are not the whole goal
A robot can make repeated practice easier to deliver and record. However, movement completed by the machine is not equivalent to movement actively generated by the patient. Assistance should be selected and adjusted within an appropriate clinical protocol. Too much help may reduce effort; too little can produce frustration, compensatory movement or an unsuccessful task.
Look for adjustable assistance, useful feedback and a way to connect practice to meaningful activities. A person working toward reaching a shelf may need more than a score from a screen-based exercise. Robotic sessions may complement conventional therapy, strengthening, balance work and task practice. Evidence should be interpreted for the specific diagnosis, stage of recovery and treatment combination being considered.
Measure progress outside the machine
Built-in dashboards can describe repetitions, range, assistance levels and session duration. These measures are useful operationally but do not necessarily establish functional improvement. Pair device data with suitable clinical assessments and patient-defined goals. Where relevant, consider walking ability, transfers, upper-limb activity, fatigue, participation and whether any gains persist beyond the treatment setting.
When comparing studies or services, check the total therapy exposure. A robotic group receiving more practice time than its comparator raises a different question from equal-duration treatment. Also examine follow-up, participant selection and adverse events. Avoid interpreting a change in an internal game score as proof of recovery or applying findings from one neurological condition to all others.
Choose a service model, not only equipment
For a clinic, useful specifications include patient size limits, movement ranges, transfer requirements, available assistance modes and cleaning instructions. Check floor loading, access routes, power needs and storage for accessories. Ask how software data can be exported and whether a subscription is needed to retain access to treatment records. Accessibility should cover the interface as well as the physical hardware.
Obtain itemised quotations for equipment, accessories, installation, training, service, licences and replacement contact materials. Compare cost per delivered session using realistic fitting and cleaning time. Do not assume that automation removes the need for supervision. Clarify the staffing model permitted by the instructions and local policy, and plan for conventional therapy when the robot is unavailable.
What a sensible trial looks like
Define the patient group and clinical goals before arranging a supervised evaluation. Record exclusions, setup time, staff workload, treatment tolerance and relevant outcomes. Include the people who manage transfers and cleaning, not only the team leading the purchase. Establish how incidents, near misses and patient feedback will be reviewed.
For patients and families, ask what the robot adds to the existing programme, how progress will be assessed and when treatment should be reconsidered. Request the full session price and any assessment charges directly from the provider rather than relying on general market figures. Robotics can be a useful therapy tool; it is not a promise of walking, restored hand function or a particular recovery timeline.