Mobility Atlas research: lift-chair-guide
Data and method for the article at mobilitatlas.com/blog/lift-chair-guide/
License: CC BY 4.0 (credit "Mobility Atlas")
Prepared: 19 October 2026

WHAT WAS READ
Every FDA MAUDE medical device report with text filed under product code INO (Chair, Positioning, Electric; 21 CFR 890.3110), received at any date up to the latest openFDA update read on 19 October 2026: 550 reports, received from 1993 to 15 July 2026, all coded. The code covers home lift chairs and clinical positioning chairs: 391 reports are about home lift chairs or power lift recliners, 87 about clinical positioning chairs, 27 about an electric positioning chair whose setting cannot be told, and 45 about another device.
All 5 FDA recall records for product code INO, downloaded through openFDA on 19 October 2026 (2 of them about home lift chairs), and the CPSC recall database entries for lift chairs and power recliners.
Makers: grouped from the manufacturer and brand names in each report (Pride, Stryker, Invacare, Altimate, Golden Technologies, other).

HOW IT WAS CODED
1. A codebook (codebook.txt) was written before any report was coded. Its SHA-256 fingerprint, recorded at 2026-10-19T13:42:03Z, is 461fcfd68fdf0b43a45431450481c55fc295e82342a6705e992148f61d4d7410. Any later change to the codebook would change this fingerprint.
2. The reports were split into chunks. Two coders (AI language models working separately, each in a private scratch folder, each without access to the other's answers, told to read every report and not to use keyword scripts) coded every report in every chunk against the codebook.
3. Agreement between the two coders: device 99.3% (kappa 0.984); problem 94.5% (kappa 0.931); injury 95.3% (kappa 0.904); fall 99.3% (kappa 0.982); fire 100.0% (kappa 1.0); user 98.9% (kappa 0.958); what was found 96.2% (kappa 0.941).
4. Every report where the coders differed was decided by the lead analyst under the written rules below (84 reports marked adjudicated). A consistency pass then re-applied the rules across the whole table and changed codes in 85 more reports (marked consistency).
5. Counts in the article were produced by script from the final table, not by hand. Some sub-counts (parts, places or words named in reports) are keyword counts over the report text; the exact word patterns are listed under KEYWORD COUNTS. Different wording would give somewhat different counts.

ADJUDICATION RULES
C1: DEVICE: L is decided by function: a recliner or chair whose powered seat rises to help the user stand is L wherever it stands (a nursing-home resident included); P needs a clinical product type or brand (Trurize, Winco, Barton, Invacare medical chairs, geri, podiatry, dialysis, transfer chair) or a clinical setting named for the event; an unbranded recliner whose only cue is the word patient, pt or resident is U; EasyStand standing frames, scooters, power chairs and patient lifts are O.
C2: PROBLEM (M vs F vs S vs Q for movement, tips and falls): M when the reporter says the chair or one of its parts moved abnormally (threw, dumped, ejected, catapulted or jolted the user, dropped, a footrest or ottoman fell, a back or recline would not lock, would not stop, stuck); F when the chair as a whole tipped, tilted over, leaned forward or flipped with nothing said to break, or the user fell or slid and no part or behaviour is named (a bare "many issues" or "malfunction" beside the fall does not make it Q, because F comes first in the list); S when a break, crack, pop or snap is alleged, or a named part is said to sit out of place (footrest protruding or not flush, seat sunk with a bar poking, a dip in the back); a bare "the chair broke" or "the chair is leaning" with nothing named is Q.
C3: PROBLEM (hand control vs motor): E when the reporter says the hand control itself misbehaved (button or switch stuck, control froze, stopped working, sparked or got hot); M when the report only describes the chair moving wrongly (kept rising after the button was released) even if the hand control was later found faulty or replaced, which belongs in FINDING.
C4: PROBLEM (accidental activation): when a working control was pressed by the user's body, a wedged position, an object or a third party and no device fault is alleged, the event is F if a fall followed and U if not.
C5: PROBLEM (clinical-chair summary terms): brakes not holding or reduced brake force are Q, since the codebook has no brake category and the maker's broken or dirty components belong in FINDING, unless the allegation itself names a damaged part (damaged casters is S); a chair-exit alarm, nurse call or "clinician not alerted/aware of possible patient fall" is E (electronic alert function); accessible AC current or exposed wires is E; a hazard with nothing broken or moving (tripping hazard, accessible sharp edges, fails to support occupant) is Q.
C6: PROBLEM and FIRE (house fires): a fire in a home or property where the report says only that the chair was present or located, or where the maker or fire report puts the cause outside the chair (external source, cigarette, power strip), is X; E only when the report alleges that the chair, its parts or electrics caught fire, smoked, sparked or may have caused the fire. FIRE is y only when the chair or its parts are said to have burned, smoked, sparked, melted or charred (burn marks or fire damage on the chair count), or when the report alleges that the chair or its electrics may have caused the fire; a house fire in which the chair was merely present is n.
C7: PROBLEM (other recurring choices): seat, back and hand control hot to the touch is E (electrical heat, not a heat-pad burn); when an electrical complaint (massage unit stopped and buzzes on the buttons) and a tip are both alleged, E wins by list order; a labelling or no-warning complaint with no event described is X; a power cut that leaves the chair stuck with nothing failed is X; an object under the chair damaging the lift mechanism is M; a skin scratch on a device part during transfer with no fault alleged is U.
C8: INJURY: y only when harm to the body, pain or medical care is stated (cut, bruise, burn, fracture, wound, rash, sores, an electric shock felt, a pinched limb, passing out, ER, hospital, death); a fall, being thrown or dumped, hitting the head, floor or a table, being stuck, or calling 911, EMTs, paramedics or the fire department to get the user up or out is n without stated harm, and event type Injury alone does not make it y.
C9: FALL: y when the user was thrown, ejected, dumped or catapulted out of the chair, the chair tipped or flipped over with the user in it, the chair fell on top of the user, or the user was released and dragged to the floor; n when the user was caught before reaching the ground or stayed in the chair.
C10: USER: y when the report or the maker names an act of the user or a third party as the cause, likely cause or possible cause (misuse, abuse, operating against an obstruction, moisture or liquid ingress, a cigarette or careless smoking, wrong installation or assembly, a field retrofit, a non-original part, a driving error such as turning too sharply), hedged forms like "cannot be ruled out" included; n for ordinary use such as leaning or reaching, or when an act is mentioned but not linked to the event.
C11: FINDING: D when the maker, a service technician or a clinic confirmed a fault, or a part was repaired or replaced to fix the complaint (provider or FST repair, hand control replaced, stuffing corrected, maintenance required, a burned part replaced); N when the evaluation found no fault, could not duplicate, found only cosmetic damage, or concluded the cause was outside the device (misuse, abuse, obstruction, moisture, improper installation or assembly, a field retrofit) even when the damage itself was confirmed and even when the conclusion is hedged, and N when any completed evaluation found no issue; X when nothing was evaluated, the device was not made available, evaluation is pending, the returned parts could not be correlated, the only remark concerns a different allegation, or no inspection lies behind "no malfunction reported"; a whole-unit or goodwill replacement is not a finding.

KEYWORD COUNTS (Python regular expressions, matched against the upper-case report text without its header line)
KW = {'fracture': 'FRACTUR', 'head': '\\bHEAD\\b', 'trapped': 'TRAP|PINN|PINCH|CRUSH|CAUGHT|STUCK UNDER|UNDER THE CHAIR', 'child': '\\bCHILD|\\bKID|\\bTODDLER|GRANDCHILD', 'pet': '\\bDOG\\b|\\bCAT\\b|\\bPET\\b', 'fire': 'FIRE|SMOKE|MELT|BURN', 'hand_control': 'HAND CONTROL|HAND-CONTROL|REMOTE|PENDANT', 'battery_backup': 'BATTERY BACK|BACKUP BATTER|BATTERY'}

QUALITY NOTES (disclosed in full)
- The article went through fact-check passes by checkers who had not written it (a full pass with a reader review, then a control pass over the changes), each recomputing numbers from the files here and reading quotes against their sources; every error found was corrected before publication.

AMOUNTS
Every report text was searched for money figures; none gives one. Prices in the article are prices shown on makers', dealers' and retailers' pages and in the Mobility Atlas catalog, listed in lift-chair-guide-documents.csv and the article sources.

LIMITS
- Reports are not rates. Nobody publishes how many devices of each model are in use.
- A report is an allegation or a description; the FDA says the submission of a report "itself does not necessarily demonstrate that the device caused or contributed to the adverse outcome or event."
- One maker (Pride) is named as the maker in 348 of the 550 reports, which reflects its sales and filing, not its safety; it also files reports from insurers' and attorneys' letters about house fires in which one of its chairs was present.
- Setup error disclosed: in the first coding pass, a fault in the coder prompts made the second coder write its answers for chunks 2 to 8 into the first coder's files. Those files were discarded unread and both coders recoded chunks 2 to 8 from scratch with corrected prompts before agreement was measured; chunk 1 was not affected.
- 8 report texts are exact copies of another report (none in the home lift chair group) and are counted as filed.

FILES
codebook.txt        the codebook as fingerprinted
coded_reports.csv   one row per coded report: both coders' answers, the final code, and how a disagreement was resolved (resolution, rule_applied)
report_pool.csv     the reports in the pool with product code, brand, date received and event type
The full report texts are public: search MAUDE or openFDA by report key.
lift-chair-guide-documents.csv  the documents and pages read, with links