Mobility Atlas research: home-hospital-bed-guide
Data and method for the article at mobilitatlas.com/blog/home-hospital-bed-guide/
License: CC BY 4.0 (credit "Mobility Atlas")
Prepared: 16 October 2026

WHAT WAS READ
FDA MAUDE medical device reports filed under product code LLI (Bed, Therapeutic, AC-Powered, Adjustable Home-Use; 21 CFR 880.5100), FNJ (Bed, Manual; 21 CFR 880.5120), FNK (Bed, Hydraulic, Adjustable Hospital; 21 CFR 880.5110) or QTC (Adult Portable Bed Rails For Medical Purposes; 21 CFR 890.5050), read in openFDA on 16 October 2026, that carry text: (a) every such report received 1 January 2019 to 31 December 2025 (423); (b) every such report received in 2026 whose manufacturer field does not contain "Medline" (20; six of them are about Medline beds, with the field blank or misspelled); (c) a systematic sample of 250 of the 2,078 reports with text that Medline filed in 2026, taken as 50 consecutive reports at five evenly spaced positions of the openFDA result list (skips 0, 415, 830, 1245 and 1660). Pool: 693 reports, all coded; 3 describe no bed event (coded N), leaving 690. The stratum of each report is in strata.csv.
All 34 FDA recall records for product codes LLI, FNJ, FNK and QTC, downloaded through openFDA on 16 October 2026 (recall_facts.py).
Makers: grouped from the manufacturer and brand names in each report (Hillrom/Baxter, Invacare, Medline, Craftmatic, Drive, other).

HOW IT WAS CODED
1. A codebook (codebook.txt) was written before any report was coded. Its SHA-256 fingerprint, recorded at 2026-10-16T04:17:56Z, is 69510f6805b5560708bcd7c44266df438f9efefec1e1f2cc3f1156cf22b7a335. 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 91.3% (kappa 0.766); category 88.5% (kappa 0.859); injury 99.7% (kappa 0.989); fall 99.6% (kappa 0.976); setting 97.5% (kappa 0.952); what was found 96.7% (kappa 0.927).
4. Every report where the coders differed was decided by the lead analyst under the written rules below (164 reports marked adjudicated). A consistency pass then re-applied the rules across the whole table and changed codes in 0 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
B1: DEVICE: O for a consumer adjustable bed base (every Craftmatic base), a mattress alone, or a report that names no bed, rail or bed part; B for a hospital, homecare, manual or enclosed medical bed (SleepSafe, Beds by George) or a bed used in a care facility, and a pendant, hand control or bed remote counts as naming the bed, so pendant or remote reports are B.
B2: DEVICE (rails): B when the text ties the rail to the bed itself (a side rail that raises, lowers, latches or releases with the bed, or rails delivered with and attached to a hospital bed as one set); R when the rail is an assist bar, handle or separately bought rail, or when the rail is itself the reported product and no bed make is named (including a companion report that covers the bed separately); the bare words side rail or bed rail do not decide.
B3: CATEGORY C vs U: C when the pendant, hand control, motor, actuator, cable, power cord, power supply or electronics failed, as the codebook lists them, whether or not the bed stopped moving; this includes a damaged, frayed, cut or exposed cord or cable, a shock from the bed or its cord with no smoke, sparks or heat, and a failed bed-exit, nurse-call or brake-not-set alarm (bed electronics); chosen so decided rows match the rows on which both coders agreed C for cords and alarms; U only when a device problem is alleged and no part or function is named.
B4: CATEGORY H: smoke, sparks, flames, burn marks, melting, a burning smell, or a pendant, cable or motor that is hot or warm to the touch is H; a fire started by smoking, oxygen or another outside source with no bed fault stated is X, and its finding is N.
B5: CATEGORY M vs C vs F: M needs a described movement (moved, dropped, sprang up, ran up, overshot or did not stop); a bed or section stuck in a position, including stuck raised or resting tilted, is C; a section that dropped with nothing broken is M, while a section or headboard that fell or collapsed because a part broke, bent or a bolt came out is F.
B6: CATEGORY F vs C vs K: F when a frame part, weld, deck, section, leg or caster broke or bent, or a bolt came out of the frame or a frame bracket, and by precedence F beats C and K (a broken caster is F even when brakes or locks fail); unspecified broken pieces, a plastic piece, motor housing, gear or gearbox of the drive are C.
B7: CATEGORY (arrival): a part found broken on arrival or at unboxing is coded by the broken part exactly as a break in use (frame F, drive or controls C); X is kept for labeling, cleaning, delivery or assembly issues with no broken or failing part and for allegations about assembly practice with no use event.
B8: CATEGORY (residual): the category codes what the report alleges even when the maker later finds no fault or a user cause (that goes in FINDING); a fall with no failure from 1-7 named is L even when an unspecified bed problem or a shifting mattress is blamed, because L precedes U; N when no event is described at all (a general design concern).
B9: INJURY: y when the text states bodily harm caused by the event (scratch, cut, bruise, burn, fracture) even without medical care; n when only a near miss, an impact with no harm, an almost-needed hospital visit, or a shock with no harm or care is stated.
B10: FALL: y when the person went to the floor from the bed while lying, sitting, getting in or out, transferring or leaning on it, including sliding down or being found on the floor beside the bed with the head still at the bed; n when the person stayed on the bed.
B11: SETTING: C when a facility is the reporter or location, or residents or facility staff are named; H when a home or a room of a house, a family member, or the user's own aide or carer is named; otherwise U; a nurse or case manager, the words patient, customer or homecare bed, and the brand Medline at Home do not decide by themselves.
B12: FINDING: D when the maker, its technician or its review of photos or video identifies the fault, or the maker or technician replaced, swapped or reset a part to fix it; N when the maker says no fault, could not replicate, works as designed, or names an outside cause (wrong mattress or plug, missing parts, misuse); X when only the reporter, family or provider describes the fault, or the maker gives advice or goodwill parts with no conclusion.
B13: PAIRS: reports that repeat the same event or the same text (end-plug pair, plastic-piece pair, cord series, alarm series) get identical codes.

KEYWORD COUNTS (Python regular expressions, matched against the upper-case report text without its header line)
KW = {
 'neck_or_head_trapped': r'NECK|HEAD',
 'between_rail_and_mattress': r'BETWEEN THE (BED )?(SIDE )?RAIL AND (THE )?MATTRESS|BETWEEN (THE )?MATTRESS AND (THE )?(BED )?(SIDE )?RAIL',
 'asphyx_strang': r'ASPHYX|STRANGUL|SUFFOCAT',
 'dementia_alz': r'DEMENTIA|ALZHEIMER|CONFUS',
 'air_mattress': r'AIR MATTRESS|ALTERNATING|LOW AIR',
 'pendant_or_hand_control': r'PENDANT|HAND ?CONTROL|REMOTE|HANDSET',
 'fire_smoke_melt': r'FIRE|SMOK|MELT|SPARK|BURN',
 'fracture': r'FRACTUR',
 'hospital_or_er': r'HOSPITAL|EMERGENCY ROOM|\bER\b',
 'skin_tear_bruise': r'SKIN TEAR|BRUIS|LACERAT|ABRASION',
 'assist_bar_handle': r'ASSIST BAR|BED HANDLE|BED ASSIST|CANE|SWIVASSIST|GRAB BAR',
 'strap': r'STRAP',
 'weight_limit': r'WEIGHT (LIMIT|CAPACITY)|OVER ?WEIGHT',
 'motor_actuator': r'MOTOR|ACTUATOR',
 'cord_cable': r'\bCORD|CABLE',
 'alarm': r'ALARM',
 'weld': r'WELD',
 'caster': r'CASTER|WHEEL',
}
E_KW = {'neck': r'NECK', 'arm': r'\bARM', 'leg_foot': r'\bLEG|FOOT|KNEE|CALF', 'asphyx_strang': r'ASPHYX|STRANGUL|SUFFOC'}

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 home-hospital-bed-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."
- The Medline 2026 reports are a sample of one maker's retrospective batch; they are counted separately and not added to the other strata.
- 160 of the 440 reports outside that sample are short technicians' service notes from one maker, and 74 are one series about end caps on one maker's assist rails; both inflate their categories.
- Rule B3 counts damaged power cords and failed bed alarms under controls (C), so decided rows match rows both coders had agreed on.
- The same event is sometimes filed twice; we count reports, not events.

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
strata.csv          which part of the pool (stratum) each report came from
The full report texts are public: search MAUDE or openFDA by report key.
home-hospital-bed-guide-documents.csv  the documents and pages read, with links