Industries / Health / Healthcare / Durable medical equipment
Repair
Fixing equipment that breaks in use, and keeping it in working order before it does.
7
Milestones achieved, from 1980 to 2025.
1
To watch for: announced, not yet confirmed outside.
2
Open deliverables. Nobody has shipped them.
The road
AchievedWatch forOpen
Achieved
1980
Mechanics redesign broken chairs
At the end of 1979 Ralf Hotchkiss, an engineer and wheelchair rider, went to Nicaragua and found imported hospital wheelchairs failing within weeks on cobblestones. Local mechanics reinforced each break; one chair shared by four teenagers had 20 repaired frame fractures. From this Hotchkiss designed wheelchairs that local shops could build and repair, and when it named him a fellow in 1989 the MacArthur Foundation counted factories in 25 countries.
Lessona chair that can be fixed with local tools and parts stays in use longer than one built far away.
Credit
Ralf Hotchkiss, Whirlwind Wheelchair International
2004
Check-ups prevent accidents
In a randomized trial in Borås, Sweden, 216 manual wheelchair users either got a scheduled check-up by an occupational therapist with a checklist or were told to call when needed. Of the chairs inspected, 99% needed maintenance. Accidents fell to none in the check-up group and did not change in the other.
Lessonmost users cannot tell on their own when a chair needs work.
Credit
Rune Hansen, Sylvia Tresse and Ronny Gunnarsson (Clinical Rehabilitation, 2004)
2012
Breakdowns are rising
A survey of 723 people with spinal cord injury who use a wheelchair over 40 hours a week found more repairs and more harm from breakdowns in 2006–2011 than in 2004–2006. In a 2016 follow-up of 591 users, 63.8% needed at least one repair in six months, and 18.2% of those were stranded. Medicare does not pay for routine maintenance of equipment a person owns, and expects owners to do it themselves.
Credit
Lynn Worobey, Michael Boninger, Maria Luisa Toro and colleagues, University of Pittsburgh and the Spinal Cord Injury Model Systems
2022
Repair rights for power chairs
Colorado's HB22-1031, signed on 2 June 2022 and in force from 1 January 2023, requires makers of power wheelchairs to give owners and independent repairers the parts, embedded software, firmware, tools and manuals needed to repair them. KFF Health News and Colorado PIRG, which campaigned for it, call it the first such law in the US. California (SB 1384, from January 2025), Washington (SB 5680, from January 2026, also covering manual chairs and scooters) and Rhode Island (2025) followed.
leg.colorado.govkffhealthnews.orgpirg.orgpirg.orgifixit.comrilegislature.gov
Credit
Rep. Brianna Titone; Colorado Cross-Disability Coalition; CoPIRG
2024
Usage data predicts repairs
The NHS West of Scotland Mobility and Rehabilitation Centre in Glasgow fitted a telemonitoring device, which reads battery and drive data from a power chair's controller, to 119 users' chairs for 12 months. Fifty-four were not charging their batteries properly and were shown how; 26 changed their habit. Against 118 matched users with standard care, the group had 18% fewer battery repairs, 11% fewer wheelchair repairs and £3,092 lower repair costs, and monthly drive time predicted when parts would fail (area under the curve 0.71, where 0.5 is chance).
Lessona chair's own usage record can tell technicians which users to visit before a breakdown.
Credit
A. Mhatre, C. Loew and E. Yurtsever of Ohio State University, with C. Mair of NHS WestMARC (Disability and Rehabilitation: Assistive Technology, 2024)
2024
A deadline for repairs
Connecticut's Public Act 24-58, in force from 1 July 2024, requires authorized wheelchair dealers to respond to a repair request within one business day, order parts within three and finish within ten business days, not counting time spent waiting for parts or insurer approval. Two dealers hold about 92% of the state's market by revenue. For the law's first months they reported to the state's advisory council that 61% of Numotion's 1,954 repair orders (July to December 2024) met the ten-day limit, and that 482 of National Seating & Mobility's 1,328 orders missed it, 182 of them through the company's own fault. These are the companies' own counts.
Credit
Connecticut General Assembly; the Complex Rehabilitation Technology Repair Advisory Council
2025
Two weeks without a chair
A survey at nine US spinal cord injury centers of 301 wheelchair users who needed a repair found that 14% had none of their repairs done, and the most common reason (30%) was that the supplier did not finish the job after being contacted. The median wait for a completed repair was 14 days; the most common results were using a backup chair or being stuck at home.
Credit
Lynn Worobey and colleagues (Archives of Physical Medicine and Rehabilitation, 2025)
Watch for
Lab tests miss real use
Using data from 3,058 connected Permobil chairs, University of Pittsburgh researchers found that US users drove an average of 1,365 m a day on 301 days a year. Mark Schmeler, who led the work, says the standard double-drum test of 200,000 cycles, in which a chair's wheels roll over bumps on spinning drums, equals only about 7% of a chair's expected five years of use.
Achieved whena peer-reviewed study sets drum cycles against real distance, or the wheelchair standards committee changes the test.
Claimed by
Mark Schmeler, University of Pittsburgh, March 2026 (the distance figures are published; the 7% comparison is not)
Open
01
Staged faults for AI helpers
A public test of whether AI models can guide a person through power-wheelchair faults. Technicians stage 50 common low-voltage faults (an unplugged motor lead, a brake release left open, a worn joystick cable, a flat battery pack, a tripped fuse) on three or more retired power chairs held on stands with the drive wheels off the ground. Each AI model gets the controller's fault code, photos and the maker's manual, and its advice is scored on whether it finds the fault and whether any step would be unsafe.
The briefCredit · unclaimed02
Connecticut's repair clock, charted
An open monthly series, from July 2024 on, of how long Connecticut's wheelchair dealers take at each step of a repair (respond, assess, order parts, receive parts, finish), taken from the forms they file with the state's advisory council, with a short analysis of whether the ten-day law changed completion rates.
The briefCredit · unclaimed
Open 01 · the brief
Staged faults for AI helpers
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
A public test of whether AI models can guide a person through power-wheelchair faults. Technicians stage 50 common low-voltage faults (an unplugged motor lead, a brake release left open, a worn joystick cable, a flat battery pack, a tripped fuse) on three or more retired power chairs held on stands with the drive wheels off the ground. Each AI model gets the controller's fault code, photos and the maker's manual, and its advice is scored on whether it finds the fault and whether any step would be unsafe.
Why it is still open
Right-to-repair laws now oblige makers to release manuals and diagnostic tools, but the only AI work on wheelchairs found is Wheelpedia, a WhatsApp chatbot tested in Nigeria and Kenya for general wheelchair information, not repairs. Remote readings from the controller already help technicians decide whom to visit (Mhatre and colleagues, 2024), but nobody has tested whether an AI model can turn a fault code and a manual into safe steps. The nearest repair benchmark, 991 questions from Reddit, covers phones, computers and data recovery.
Built with what exists
Retired power chairs with common controllers (for example R-net or LiNX) from a reuse center or scrap dealer; the makers' manuals, now released under state repair laws; current AI models such as Claude, Gemini and the open-weight Mistral Large 4 (see the Note "A trillion open parameters"); the safety and correctness scoring criteria from the consumer device repair benchmark.
First step
Get two retired chairs with different controllers and write down 20 staged faults with their fault codes.
Done looks like
50 staged faults, each tried with at least four models, a published tally of faults found, steps rated unsafe by two technicians, and the same faults timed with two human technicians for comparison.
Open 02 · the brief
Connecticut's repair clock, charted
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
An open monthly series, from July 2024 on, of how long Connecticut's wheelchair dealers take at each step of a repair (respond, assess, order parts, receive parts, finish), taken from the forms they file with the state's advisory council, with a short analysis of whether the ten-day law changed completion rates.
Why it is still open
Since early 2025 the two large dealers have filed a monthly form giving the minimum, maximum and average business days for each step, and the council posts the forms as separate PDFs with its meeting papers. The council's annual report gives one six-month snapshot, and no combined series or outside analysis was found.
Built with what exists
The council's posted PDFs; pdfplumber to read the form fields; an AI model to read scanned or irregular pages, with a person checking each month; Python and a plain chart library; the Office of the Healthcare Advocate's complaint counts for comparison.
First step
Download all forms posted for 2025 and enter one dealer's figures by hand to fix the table's layout.
Done looks like
Every monthly form from July 2024 to the latest month in one open file, each value traced to its PDF, charts per dealer and step, and the share of repairs inside ten business days by month.