Knowledge Centre
category guide

How to choose Rehabilitation Systems

April 30, 2026· 11 min read· AI-generated

How to choose Rehabilitation Systems

From parallel bars and e-stim to robotic gait trainers: what procurement teams and biomed engineers need to know before signing a capital purchase order.


What this is and who buys it

Rehabilitation systems is one of the broadest capital categories in clinical equipment. At the low end it includes the familiar workhorses — parallel bars, Hi-Lo treatment tables, TENS/e-stim and therapeutic ultrasound units, recumbent steppers, and upper-body ergometers. In the middle tier sit body-weight support treadmill systems, isokinetic dynamometers, shockwave and laser therapy platforms. At the premium end are robotic gait trainers, ceiling-mounted dynamic body-weight support (BWS) tracks, powered upper-limb exoskeletons, and fully instrumented gait labs with integrated EMG and virtual reality. These are not interchangeable product lines — they serve different clinical missions, carry different regulatory footprints, and generate very different total cost of ownership profiles.

The buyers are equally varied. Inpatient rehabilitation facilities (IRFs) and Level I trauma centers tend to drive demand for the high-acuity robotic and BWS end. Skilled nursing facilities and long-term acute care hospitals (LTACHs) lean heavily on recumbent trainers and sit-to-stand devices suited to deconditioned geriatric patients. Outpatient orthopedic and sports PT practices are the core market for isokinetic testing systems, CPM, and modality carts. VA polytrauma centers and academic medical centers are disproportionately represented among early adopters of perturbation-capable overground BWS and EMG-integrated exoskeleton platforms. Pediatric specialty hospitals represent a separate sizing and harness challenge altogether.

Timing matters. The U.S. outpatient rehabilitation market is projected to grow at roughly 6% CAGR between 2022 and 2027, driven by aging demographics, post-COVID deconditioning referrals, and expanding post-acute service lines [S15]. Capital planning cycles are commonly tied to seven-year depreciation schedules or service-line expansion projects, which means that procurement teams rarely have the luxury of a relaxed decision window — the clinical program is usually already operational or nearly so when the equipment RFP drops.


Key decision factors

Patient population and clinical mission should be the first filter, not the last. A Lokomat-class robotic gait trainer is designed for survivors of stroke, spinal cord injury, and traumatic brain injury, and can extend to cerebral palsy, MS, and Parkinson's disease [S13, S14] — but it adds little value in an orthopedic sports clinic where the caseload is ACL reconstructions and rotator cuff repairs. Before issuing an RFP, map your actual diagnosis mix from the past 12 months and confirm that the device's indications align with your top five patient populations.

Therapist labor model and supervision ratio are where the business case is actually built for high-cost platforms. A technology-based mixed rehabilitation approach, structured so that one physiotherapist supervises up to four patients concurrently, has demonstrated cost savings over conventional one-on-one therapy [S9]. In upper-limb robotic studies, active therapist involvement occupied only about 25% of total session time [S9], meaning three-quarters of the session the therapist can be attending to other patients. If your operational model cannot achieve at least a 2:1 supervision ratio with the device, the ROI case for a $200,000+ system weakens substantially.

Reimbursement and payer policy risk is a deal-breaker that is frequently underweighted in the capital planning stage. Some major payers — Aetna's Clinical Policy Bulletin 0778 is an explicit example — classify robotic-assisted rehabilitation of the upper and lower limb as experimental or investigational for indications including traumatic brain injury, and specifically list named devices as not covered [S11]. Before a purchase order is signed, your reimbursement team needs to verify CPT code coverage with your actual payer mix in your state, not the vendor's national reference list.

Footprint and facility readiness deserves a line item in the project budget long before equipment arrives. Ceiling-mounted dynamic BWS systems require structural engineering review of load ratings; the ZeroG system, for example, requires a ceiling-mounted overhead track (100 ft standard configuration) with a minimum ceiling height of 8 feet 6 inches [S12], plus appropriate electrical capacity. Robotic platforms commonly require dedicated 208V/30A circuits. Facilities that underestimate civil and electrical prep routinely absorb $20,000–$80,000 in unplanned costs after the capital purchase is already approved.

Patient sizing envelope is a specific technical parameter that should appear in every RFP for any device that physically interfaces with the patient. The Lokomat's exoskeleton accommodates femur lengths of 35–47 cm [S13]; patients outside that range require a pediatric module or cannot use the device at all. Bariatric caseloads introduce separate weight and joint-torque limits. Confirm anthropometric envelopes for your 10th and 90th percentile patients before narrowing your shortlist.

Clinical evidence base deserves honest scrutiny. A 2025 meta-analysis of 54 RCTs involving 2,744 participants found that robotic upper-limb rehabilitation produced only a small, statistically significant positive effect on upper-limb capacity (standardized mean difference 0.14) that was not maintained at follow-up, with no significant differences for activities of daily living outcomes [S10]. That does not mean robotic systems have no value, but it does mean procurement teams should push vendors past glossy outcome testimonials and ask for effect sizes, follow-up duration, and comparator conditions.

Software, data integration, and outcome tracking have become first-tier decision factors as value-based care contracts require documented functional improvement. Verify that the system exports session data in standard formats (HL7/FHIR or at minimum CSV), that gait metrics and exercise parameters are structured and exportable to your EMR, and that any cloud-hosted analytics platform will be covered by a signed HIPAA Business Associate Agreement.


What it costs

Rehabilitation equipment pricing spans nearly three orders of magnitude, which is why "what does it cost?" is genuinely unanswerable without knowing the clinical tier. The ranges below reflect publicly verifiable distributor pricing at the entry level; mid-tier pricing is based on published commercial listings; premium robotic system pricing is not publicly listed by most manufacturers and should be treated as a request-for-quote category with significant variability based on configuration, country, and negotiated volume.

  • Entry: $500–$25,000. Parallel bars, mat and Hi-Lo treatment tables, TENS/e-stim/ultrasound combination units, recumbent steppers, and upper-body ergometers. Refurbished modalities dominate the low end — mobility assist devices from major manufacturers appear on distributor sites starting in the low four figures [S15].
  • Mid-tier: $25,000–$150,000. Medical treadmills with partial body-weight support, isokinetic dynamometers, shockwave and high-power laser platforms, basic overground harness systems, and gravity-supported upper-limb exoskeleton platforms such as Armeo Spring–class devices.
  • Premium: $150,000–$500,000+. Robotic gait trainers, ceiling-mounted dynamic BWS systems, fully instrumented gait labs with VR integration, and powered upper-limb exoskeletons. Manufacturers in this tier typically publish pricing on request only; exact figures for systems like the Lokomat are not publicly verifiable [S13], and quoted prices exclude delivery, customs duties, installation, and software activation options.

Common use cases

The device category that makes sense depends entirely on who you are treating and where. The most common procurement contexts in this space are:

  • IRFs treating stroke, SCI, and TBI — robotic gait trainers and overhead BWS systems for high-repetition locomotor therapy, where patient volume and acuity justify the capital.
  • SNFs and LTACHs — recumbent total-body trainers and sit-to-stand transfer devices suited to deconditioned geriatric patients who cannot tolerate conventional gait training at program entry.
  • Outpatient orthopedic and sports PT — isokinetic testing and dynamometry, CPM, traction tables, and modality carts (ultrasound, e-stim, shockwave, laser) for post-surgical and musculoskeletal rehabilitation.
  • Cardiac and pulmonary rehabilitation programs — multi-use ergometers including upright and recumbent stationary bikes, ellipticals, treadmills, and rowing machines that support both cardiac and orthopaedic rehab indications.

Regulatory and compliance

Most powered rehabilitation systems are regulated under 21 CFR Part 890 (Physical Medicine Devices) as FDA Class II devices subject to special controls [S1]. Many require 510(k) premarket clearance demonstrating substantial equivalence to a predicate device; a 2020 Federal Register final order clarified a limited exemption for measuring exercisers and interactive rehabilitation exercise devices when prescription-use only, subject to the limitations in §890.9 [S4]. Regardless of exemption status, verify any device's status independently in the FDA 510(k) database at accessdata.fda.gov — and note that it is legally incorrect to describe a 510(k)-cleared device as "FDA approved." The correct term is "FDA cleared" [S2].

Electrical safety and electromagnetic compatibility are governed by the IEC 60601-1 series. IEC 60601-1:2005+A1:2012+A2:2020 (Edition 3.2) covers basic safety and essential performance for medical electrical equipment generally [S5], and December 17, 2023 was the mandatory implementation date for Amendment 2 for new FDA submissions. Relevant collateral standards include IEC 60601-1-2 (EMC), 60601-1-6 (usability engineering), and IEC 60601-1-11 for home healthcare environment use. Any system that captures or transmits identifiable patient performance data — cloud-hosted analytics, remote diagnostics — triggers HIPAA obligations; require a signed BAA before go-live. Maintenance programs should align with ANSI/AAMI EQ56:2013 and the updated ANSI/AAMI EQ103:2024 Alternate Equipment Management standard, which establishes minimum AEM requirements that satisfy CMS Conditions of Participation [S6].


Service, training, and total cost of ownership

Installation for robotic and ceiling-track systems is not a plug-in event. Structural engineering review, NRTL inspection, and commissioning typically consume one to four weeks, and therapist certification training — commonly 16–40 hours depending on platform complexity — must be completed before clinical go-live. Budget these labor and downtime costs explicitly; they rarely appear in the capital line but routinely surface as budget variances.

Annual preventive maintenance contracts on robotic systems generally run 8–12% of capital cost. OEM contracts typically bundle software updates, harness replacement, and load-cell calibration — all of which are safety-critical on dynamic BWS platforms. For non-robotic Class II modalities, third-party service can yield meaningful savings, with some independent service organizations pricing solutions at up to 40% below OEM rates; this is defensible under an AAMI EQ103-compliant AEM program as long as safety is not compromised [S6, S7].

Parts availability and the vendor's published end-of-support date are more consequential than they appear at contract signing. Today's rehabilitation robotics are software-defined platforms: when the vendor ends support for a hardware generation, the clinical program ends with it, regardless of the physical condition of the hardware [S8]. For depreciation planning, most accountants use a seven-year recovery period for medical and dental equipment [S8]; in practice, mechanical rehabilitation equipment (parallel bars, mat tables) tends to remain serviceable for 10–15 years, while software-driven robotic platforms realistically carry a 7–10 year useful life bounded by vendor support cycles.


Red flags to watch for

A vendor who supports clinical claims with only single-arm pilot studies or unpublished white papers should be treated with skepticism — the peer-reviewed literature shows that robotic rehabilitation produces only modest pooled effect sizes and that benefits often do not persist at follow-up [S10]. If an RFP response cannot produce a 510(k) K-number and product code that verify in the FDA database, do not proceed. Marketing materials that describe a Class II device as "FDA approved" (rather than "FDA cleared") indicate either carelessness or deliberate misrepresentation — neither is reassuring in a safety-critical product [S2]. Similarly, any system without a visible NRTL mark (UL, ETL, or CSA) or whose manufacturer declines to share IEC 60601-1 test reports is an unacceptable electrical safety risk in a clinical environment. Finally, software licensing structured around steep annual subscription escalators for basic exercise libraries, or session data locked in a proprietary format with no standard export, are contractual risks that will compound over the system's entire life — negotiate data portability and subscription caps before signing.


Questions to ask vendors

  1. Provide the FDA 510(k) K-number, product code, and 21 CFR regulation citation; also provide current IEC 60601-1 (Edition 3.2 / Amd. 2:2020) and IEC 60601-1-2 EMC test reports.
  2. What is the documented mean time between failures (MTBF), the current hardware/software end-of-support date, and your policy for migrating customers to successor platforms?
  3. What is the full patient anthropometric envelope — height, weight, femur length, joint range of motion — and which accessories are required for pediatric or bariatric populations?
  4. Provide a fully loaded 5-year total cost of ownership including installation, structural reinforcement, therapist certification training, annual PM, software subscriptions, consumables (harnesses, electrodes, cuffs), and load-cell calibration.
  5. What staffing ratio is achievable in steady-state clinical operations — can one therapist supervise two or four patients concurrently — and what certification is required to reach that ratio?
  6. List peer-reviewed RCTs (not vendor case reports) supporting the indications you market; include reported effect sizes, follow-up duration, and payer coverage status for the relevant CPT codes in our state.

Alternatives

The new-versus-refurbished calculus depends heavily on device type. Mechanical equipment — parallel bars, mat tables, ergometers — is well-suited to certified pre-owned channels, where reputable distributors conduct safety verification and component inspection before resale. Robotic systems are a different story: load-cell recalibration, harness pull-test certification, and software entitlement transfer are safety-critical steps that generally require OEM involvement, making non-OEM-refurbished robotics a meaningful risk.

On the lease-versus-purchase question, operating leases of 36–60 months make particular sense for software-heavy robotic platforms, where technology generations turn over quickly and the lessor absorbs obsolescence risk. Capital purchase remains the more economical structure for mechanical equipment with a 10–15 year useful life and stable technology. For high-cost robotics, it is also worth modeling a portfolio approach rather than concentrating an entire program's investment in a single vendor's platform — since most commercial devices act on a limited number of joints and planes, comprehensive upper-limb or lower-limb rehabilitation programs realistically require a set of complementary devices, each addressing different joints or movement planes [S10]. Single-vendor lock across an entire rehabilitation program creates both clinical coverage gaps and unfavorable contract leverage at renewal.


Sources

Sources

Browse vendors in

MedSource publishes neutral guidance. We do not accept payment from vendors to influence the content of articles. AI-generated articles are reviewed for factual accuracy but cited sources should be the primary reference for procurement decisions.