Safe, Cheaper, Essential: What the Data Really Says About Independent Medical-Device Servicing — and What It Means for Ultrasound

Act I — Servicing is a strategic issue, not a back-office one

Takeaway: A large, expanding, aging imaging install base means the equipment bought in the 2010s is now in its peak service-and-repair window — and for ultrasound specifically, the economics of board-level repair versus replacement are decisive.

The fleet is big, and it is getting older

Neither the WHO nor the OECD tracks diagnostic ultrasound as a separate line item; both track the heavier imaging modalities. That makes the imaging fleet a useful — and conservative — floor for the install base ultrasound sits within, because ultrasound units are more numerous and more widely distributed than any of the modalities below.

Across OECD-area countries, 184,809 imaging and radiotherapy units span six modalities, latest year per country: CT 61,422, mammography 47,511, MRI 40,546, gamma cameras 22,026, radiotherapy 9,116, and PET 4,188 ¹⁵. The United States alone fields 74,638 units — the single largest fleet — followed by Japan (29,269), Brazil (9,069), Korea (9,045), Italy (7,611), and Germany (6,028) ¹⁵.

An expanding fleet is, by definition, an aging fleet. On OECD's balanced panel, per-million density rose across every active modality from 2010 to 2023: PET +97% (1.15 → 2.27), MRI +78% (10.9 → 19.4), CT +41% (20.7 → 29.0), radiotherapy +17%, mammography +17% — with only gamma cameras declining (−15%, the one legacy modality) ¹⁵. Units bought during that 2010s buildout are now mid-to-late life: precisely the window where spare parts, board-level repair, and timely service determine uptime.

The access stakes are not evenly distributed. MRI density spans a greater-than-2,900× range between the best- and worst-served countries; 24 countries report zero MRI and 8 report zero CT ¹⁴. Where a single scanner serves a region, an OEM service monopoly does not just raise a budget line — it directly gates patient access. And the servicing-footprint split matters: CT runs roughly 70% hospital / 30% ambulatory ¹⁵, and it is the smaller ambulatory clinics, with the thinnest OEM-contract budgets, that benefit most from competitive independent servicing.

For ultrasound, the repair-versus-replace math is decisive

Ultrasound's wear part is the probe (transducer), and probes fail. The most concrete pricing on record — dated, and therefore understated in 2026 dollars — comes from an ITN/MedPro Imaging table: a standard probe repair averages ~$1,500 (range a few hundred to a few thousand) versus $5,000 for a refurbished replacement or $14,000 new; a transesophageal (TEE) probe repairs for $5,000–$8,000 versus $14,000–$20,000 to replace ¹⁹, with new TEE purchases running up to ~$70,000 ¹⁸. The headline industry figure — 60–70% cost savings repairing versus replacing — flows from this gap ¹⁷.

A word of caution on these numbers. The widely repeated trio — ~25% of probes in clinical use are defective, ~75% of those are repairable, and 60–70% repair-versus-replace savings — is single-lineage and vendor-adjacent, tracing largely to a 2013 trade article resting on a 2002 study (Weigang et al.) ¹⁷. Present it as an industry rule of thumb, not settled fact.

It does, however, survive independent corroboration. A peer-reviewed multicentre survey of 219 probes across 12 sites found 37% had at least one fault, 25% needed action ("amber"), 13% were unfit for use ("red"), and only 63% were fault-free — with more than 90% of faults detectable by simple, low-cost tests ²⁰. A separate multicentre survey put the average annual probe failure rate at 13.9% ²⁰. Even an OEM concedes the point: Siemens cites a "high incidence of defective ultrasound transducers in use in routine clinical practice" ²¹. The "roughly one in four probes is defective" rule of thumb is, if anything, conservative against the peer-reviewed 37% fault rate.

Value is migrating from the box to the service-and-parts economy

The broader trade data shows hardware commoditizing while the service-and-parts economy expands. Finished ultrasound-system exports (HS 901812) grew from $12.32B (2010) to $18.87B (2024), a CAGR of about 3.1%; the broad medical-instrument parts/accessory basket (HS 901890) grew from $148.27B to $366.50B over the same period, a CAGR of about 6.7% — more than twice as fast ¹². Tellingly, the implied price per kilogram of finished systems fell from ~$296 to ~$251 ¹²: the box is getting cheaper while keeping it running gets more valuable.

One important caveat: HS 901890 is a very broad parts basket spanning all instruments of heading 9018, not ultrasound-specific, and our Comtrade pull carries no country dimension. Treat the parts figures as a directional proxy for where value is shifting — never as "ultrasound parts" in absolute dollars, and not as evidence of any single country's share ¹².

The probe-repair market itself is estimated at roughly $0.94B (2024), projected toward $2.5B by 2035 at a 9.3% CAGR ²², against a new-transducer market of about $3.7B (2024) growing at 6.8% ²³. (A second report's "$2.98M" probe-repair figure is almost certainly a units error and is excluded.) The signal is consistent: an expanding, aging fleet feeding a fast-growing repair economy.

Act II — The debate from the primary record

Takeaway: When you go to the source documents — FDA's 2018 study, MAUDE, the recall and enforcement databases, and a second regulator in the UK — the "third-party servicing is a safety crisis" narrative does not survive contact with the evidence.

Who actually showed up to the FDA docket

First, a correction that matters for credibility. The genuinely on-topic servicing/remanufacturing docket — FDA-2018-N-3741 — contains 87 comments and 12 documents from 69 unique commenters ¹². Larger figures sometimes cited ("1,087 comments," "6,795 comments") conflate this with an unrelated Laboratory Developed Tests docket (FDA-2023-N-2177) that must be excluded; the servicing-specific count is 87 ¹³. We lead with FDA's quantified findings below rather than comment volume precisely because volume is the weakest, most easily mischaracterized part of this record.

Of those 87 comments, the right-to-repair camp outnumbered the OEM camp. By stakeholder: individuals (mostly HTM/biomed) 32 (36.8%); OEM trade associations 12 (13.8%); hospitals/health systems/HTM 10 (11.5%); OEMs/manufacturers 10 (11.5%); independent service organizations (ISOs) 9 (10.3%); ISO/HTM trade associations 9 (10.3%); patient-safety NGOs 3 (3.4%); and clinician/accreditation bodies 2 (2.3%) ¹. Grouped by side: OEM-side 22 (25.3%) versus independent-side 28 (32.2%), plus 32 largely pro-independent individuals ¹.

FDA's 2018 Section 710 study: the load-bearing exoneration

FDA studied the safety question directly and reported its conclusion in plain language. The headline finding, verbatim:

"The currently available objective evidence is not sufficient to conclude whether or not there is a widespread public health concern related to servicing, including by third party servicers, of medical devices that would justify imposing additional/different, burdensome regulatory requirements at this time." ⁴

And, just as directly:

"Rather, the objective evidence indicates that many OEMs and third party entities provide high quality, safe, and effective servicing of medical devices." ⁴

The single quote that guts the OEM safety narrative addresses where the cited harms actually come from:

"A majority of comments, complaints, and adverse event reports alleging that inadequate 'servicing' caused or contributed to clinical adverse events and deaths actually pertain to 'remanufacturing' and not 'servicing'." ⁴

FDA also affirmed the category's essentiality: "The continued availability of third party entities to service and repair medical devices is critical to the functioning of the U.S. healthcare system" ⁴ — a position the agency restates on its public remanufacturing-and-servicing page ⁵. It declined to impose new regulation, committing instead to four non-regulatory actions, the first of which was to promote voluntary adoption of quality-management principles — the framework an ISO-certified servicer already operates inside ⁴.

The data funnel behind these conclusions is the most persuasive single artifact in the entire record:

Of 2,114,303 MDRs analyzed (2006–2015), only 86 (0.004%) were plausibly servicing-related ⁴. A separate pull found 4,301 MDRs since 1992 explicitly naming a third-party servicer — 40 deaths, 294 serious injuries, 3,791 malfunctions, 176 other ⁴. Of the 40 death reports, only 3 carried enough information to conclude servicing caused or contributed — and 2 of those 3 were deaths of the service engineer, not a patient (a field service engineer died servicing a CT scanner; another died servicing an MRI). The single conclusive patient death involved a patient-lift rail reinstalled incorrectly; a fourth death traced to a remanufactured, uncleared imaging system ⁴.

The OEM trade association AdvaMed cites these same numbers — "more than 4,300 negative incidents... including 40 deaths and 294 serious patient or device user injuries" from "up to 21,000 companies" ²⁷. But every one of those figures traces back to this FDA report — the report that itself attributed the majority of them to remanufacturing, not servicing ⁴. In fairness, FDA also noted that third-party servicers have no mandatory MDR-reporting obligation, so servicing issues may be under-reported ⁴ — a caveat that cuts both ways, and one FDA weighed before declining to extend mandatory reporting.

MAUDE: the scale and nature of failures, read honestly

MAUDE shows what kind of failures ultrasound generates at scale — and the answer is malfunctions, not patient harm. On the cleanest cut, the core diagnostic codes (IYN + IYO + ITX), 12,919 reports break down as 91.8% malfunction (11,864), 6.3% injury (816), and 0.6% death (75) ⁹. The broader diagnostic-imaging subset (15,374 reports) tells the same story, and 93.8% of those reports are manufacturer-filed — they read as device-reliability signals, not clinical-harm complaints ⁹.

Crucially, the failure families that dominate are the most-serviced parts: diagnostic ultrasound transducers (912 reports), ultrasonic probes (722), and the imaging systems they plug into ⁹. That is the texture of service demand, not a safety crisis. And where the data lets us isolate refurbished units, the pattern holds: of 426 core reports tagged remanufactured/refurbished, 425 were malfunctions, 1 an injury, and 0 deaths ⁹ — even refurbished gear fails almost entirely as benign malfunction.

The volume is rising with the fleet. Core diagnostic ultrasound MAUDE reports climbed from 568 in 2015 to a peak of 1,895 in 2024 (1,688 in 2025) — roughly a 3× rise that tracks an expanding installed base and growing servicing demand ⁹.

The honest limit: MAUDE has no device-problem code in our export, no narrative text, no servicer-identity field, and no installed-base denominator ⁹. It cannot prove independent servicing is safe, and it cannot attribute any failure to any servicer. What it can establish — that ultrasound failures are overwhelmingly benign malfunctions and that servicing is never recorded as the cause — is exactly the claim we make from it, and no more.

Recalls and enforcement: the cause is the factory, not the field

The recall record points the finger squarely at the manufacturer. Of 1,509 FDA ultrasound recalls (2000–2026), a 59.6% majority (900 records) trace to OEM device-design or software-engineering root causes — the two largest sub-codes alone being Device Design (379) and Software design (376) — and 21.9% to manufacturing/process/material, while only 0.3% (5 records) carry a servicing/repair/maintenance/installation root cause ¹⁰. Independent servicing is nearly invisible as a recall cause.

Severity undercuts the catastrophe framing too. Only 4.3% of the 680 classified ultrasound enforcement recalls are Class I (most serious); 95.7% are Class II or III ¹¹. And of the 29 Class I events, 18 are contaminated ultrasound gel — a consumable bacterial-contamination issue (Eco-Med) — leaving just 11 non-gel Class I events, none implicating independent repair ¹¹.

When you search the entire enforcement record for an event that actually blames non-OEM servicing, you find essentially one. Four records describe a single distinct event — a fetal/maternal monitor "serviced or remanufactured using non-OEM equivalent components that have not been appropriately verified or validated" ¹¹. That is the sole primary-data point supporting the OEM thesis across 680 enforcement records — and it is a fetal monitor, not an imaging system, and it involves remanufacturing-grade component substitution. Meanwhile, 42.8% of recalls and 45.1% of enforcement actions were initiated by the major imaging OEMs themselves ¹⁰¹¹ — the OEMs are the largest single recall source.

A second regulator corroborates the picture. The UK MHRA logged only 3 ultrasound-related alerts across its entire device-alert corpus over ~13 years (2013–2026) — 0.21% of 1,407 total alerts, with zero recalls and zero field-safety notices for ultrasound ¹⁶. All three are infection-control/decontamination matters (probe-cover contamination, reusable-probe decontamination, prion cross-transmission); none relate to servicing or third-party maintenance ¹⁶. Across two independent regulators, servicing-attributable ultrasound events are effectively nil.

(Two caveats. On the enforcement data: ultrasound is identified by keyword, not a product-code field, so a small number of false positives are possible, and no field links a recall to who serviced the device — the near-absence of servicing as a cause is strong but inferential ¹¹. And the MHRA corpus is thin at n=3 — best read as corroboration of a 13-year window in which almost nothing happened, not as a standalone proof ¹⁶.)

Act III — The regulatory reality in 2026

Takeaway: The 2026 regulatory framework draws a clean line between servicing (generally unregulated) and remanufacturing (fully regulated). Right-to-repair is advancing for everything except medical devices — and the genuine barrier is access to manuals, parts, and passwords, not safety.

Servicing vs remanufacturing — FDA's 2024 final guidance

FDA's May 2024 final guidance, "Remanufacturing of Medical Devices," defines the two activities precisely. Service is "Repair and/or preventive or routine maintenance of one or more parts in a finished device, after distribution, for purposes of returning it to the safety and performance specifications established by the OEM and to meet its original intended use" ⁷. Remanufacture is any act that "significantly changes the finished device's performance or safety specifications, or intended use" ⁷. The test for a "significant change" is one that "exceeds the OEM's specifications, introduces new risks, or significantly modifies existing risks" ⁸.

The regulatory consequence is a clean split. FDA "generally does not require independent third parties... that only perform servicing activities to comply with its medical device regulatory requirements," but imposes the full load — registration and listing, MDR reporting, QMSR/GMP, marketing submissions, labeling/UDI, corrections and removals — on any entity that remanufactures ⁷.

QMSR / ISO 13485:2016 — the quality framework, in force

As of February 2, 2026, FDA's Quality Management System Regulation (QMSR) is in force, amending 21 CFR Part 820 to incorporate ISO 13485:2016 by reference ⁶ (final rule FR 2024-01709, published February 2, 2024). On that date FDA retired the QSIT inspection approach in favor of Compliance Program 7382.850 ⁶.

Here is the nuance that matters for how a servicer should position itself. QMSR applies to manufacturers and remanufacturers — not to pure third-party servicers ⁶. So an independent servicer holding ISO 13485:2016 is not "complying with a mandate"; it is voluntarily operating the very quality framework FDA's 2018 report urged servicers to adopt, and the same standard QMSR now bakes into U.S. manufacturing law ⁴⁶. That is a credible mark of quality precisely because it is elective and externally audited — and it directly undercuts the "only OEMs deliver quality servicing" claim.

Right-to-repair: advancing everywhere, exempting medical devices everywhere

The legislative momentum is real but pointedly carved out. All 50 states have introduced right-to-repair legislation (Wisconsin last, February 2025), and 7 have enacted comprehensive laws — California, Colorado, Maine, Massachusetts, Minnesota, New York, and Oregon — but every one exempts medical devices ²⁶.

The medical-device-specific efforts have stalled. California's SB 605 (the Medical Device Right to Repair Act) would have required OEMs to provide documentation, software, and parts on "fair and reasonable terms"; it passed Senate Judiciary 9–0 in April 2021 — backed by the California Hospital Association, Dignity Health, Sutter Health, Cedars-Sinai, and others — but died in Senate Appropriations ²⁴²⁵. Parallel 2021 bills in Texas, Arkansas, and Hawaii also failed ²⁹. At the federal level, the FY25 NDAA's Section 828 (Sen. Warren) sought "fair and reasonable" repair access for military systems, but an AdvaMed-led coalition successfully lobbied to strip the medical-device provision ²⁶³⁰. Warren's own framing: maintenance and repair "can make up as much as 70 percent of the overall cost of a program" ³⁰.

The real barrier is access, not safety

If servicing isn't a safety problem, what is the actual constraint? Access to the OEM's manuals, parts, passwords, and specifications. In biomed surveys, 91.8% reported being denied service information for critical equipment (2020), and the share frequently blocked from critical repair documentation rose from 64% (2020) to ~79% (2026) ²⁸. The irony of FDA's 2024 guidance, as Morgan Lewis notes, is that it asks third-party servicers to self-assess whether they have crossed into remanufacturing — yet they "do not have access to the OEMs' confidential device design and specification information" needed to make that call ⁷. The barrier is informational and commercial, not a safety record.

Independent-side institutions reinforce the point. The Association of Medical Device Reprocessors notes that FDA-cleared single-use-device reprocessing is used at 9,400+ hospitals, saving $468M and diverting 20.3M lbs of waste in 2021 — evidence that regulated reuse is safe and effective ³². AMDSO stated it is "supportive of and agrees with" FDA's 2018 conclusions ³¹, and the Alliance for Quality Medical Device Servicing found "no credible evidence of systemic shortages" ³³.

Act IV — Strategic implications: choosing a credible independent servicer

Takeaway: Public buyers in Europe have already voted with their budgets — independents win the majority of maintenance work by both count and value. For HTM and procurement teams, the decision is no longer "OEM or risk"; it is "which independent meets a defensible quality bar."

The market has already chosen independents

EU public procurement is the cleanest revealed-preference dataset available, and it is decisive. Of EU TED maintenance awards with a named winner, 74% by count (2,086 of 2,818) and ~82% by value (€24.2B independent vs €5.5B OEM) went to non-OEM/independent servicers ¹³.

Maintenance is also the larger procurement category, not an afterthought: 65.1% of awarded contract value went to maintenance/repair versus 34.9% to new equipment (€94.7B vs €50.8B across valued awards), and maintenance award volume grew roughly 6.7× from 2016 to 2025, running about 2:1 over equipment every year ¹³. Maintenance contracts are also about 1.9× cheaper at the median (€620,783 vs €1,194,394) ¹³. Notably, only 2.0% of maintenance notices are no-competition direct awards, and exclusivity language ("authorized service," "original parts") is near-absent from tender titles ¹³ — little evidence that buyers treat servicing as OEM-only.

Some caveats on the TED data: `total_value` is assumed to be EUR (no currency field), only ~11,271 awards report a value, and the OEM/independent split is a winner-name heuristic that counts OEM country subsidiaries as OEM and may classify some OEM-authorized distributors as independent ¹³. The direction — independents as the default, by a wide margin — is robust to these caveats.

What a credible independent servicer looks like

If the choice is now among independents rather than against them, the buyer's job is to verify quality. The defensible bar combines an externally audited quality system with operational proof points:

• Audited quality system — ISO 13485:2016 (medical-device QMS) plus ISO 9001:2015, the same ISO 13485:2016 that QMSR now incorporates into U.S. manufacturing law ⁶.
• Real, witnessed testing before shipment — not a paper attestation.
• Parts traceability across a deep SKU base, so a repair is built from known-good, tested components.
• A warranty that puts the servicer's own balance sheet behind the repair.
• A predictable turnaround an HTM team can plan downtime around.

Guangzhou Rongtao Medical Technology, an independent ultrasound service provider founded in 2013, maps onto that bar: ISO 13485:2016 and ISO 9001:2015 certification; a 48-hour real-machine test before shipment (with photos or video where applicable); a 5–8 business-day standard turnaround; a typical 90-day warranty; 3,000+ parts SKUs; 35+ senior engineers; and service partners in 140+ countries ³⁴. The point is not that one provider is uniquely virtuous — it is that this configuration of certification, testing, traceability, warranty, and turnaround is what turns "independent servicing" from a category claim into an auditable one. (Rongtao provides service and compatibility coverage for major OEM platforms; it does not claim OEM authorization or affiliation.)

Conclusion

The safety case against independent servicing collapses against the primary record. FDA's own 2018 study found no widespread public-health concern and concluded that the harms attributed to "servicing" were largely remanufacturing; recalls trace overwhelmingly to factory design, and two regulators on two continents log essentially zero servicing-attributable ultrasound events ⁴¹⁰¹⁶.

The economic case for repair is strong and independently corroborated. Probe repair runs 60–70% cheaper than replacement ¹⁷, and a peer-reviewed survey confirms roughly a quarter to over a third of in-service probes carry actionable faults ²⁰ — real demand, profitably and safely met.

The 2026 regulatory framework legitimizes — it does not threaten — quality independents. Servicing remains outside mandatory QMSR while remanufacturing carries the full load; an ISO 13485:2016-certified servicer voluntarily operates the exact quality framework FDA endorsed ⁶⁷.

The real contest is over access, and the market has already chosen. With biomeds widely denied service information ²⁸ and right-to-repair carving out medical devices, the binding constraint is OEM control of manuals and parts — yet EU public buyers already send ~74% of maintenance work to independents ¹³. The decision facing HTM and procurement leaders is no longer whether to trust an independent, but how to verify one.

Let's talk

If your team is weighing OEM versus independent servicing — or needs board-level ultrasound repair, tested replacement parts, or added repair capacity without building every line in-house — we are happy to talk through the specifics: certifications, testing protocol, parts traceability, warranty, and turnaround. Bring a failing probe model or a downtime problem, and we'll show you the data behind the answer.