After a major repair, monitor alarm testing is not a pump template
When a biomedical equipment technician completes a major repair, upgrade, or configuration change on a bedside or transport patient monitor, the labeled alarm system still needs verification before the device returns to a clinical assignment. Returning a physiological monitor to an active patient bed is a different technical and regulatory job than returning a fluid-delivery or mechanical-therapy device to service.
In infusion therapy, return-to-service protocols centered on infusion pump return-to-service verification focus heavily on downstream occlusion pressure limits, volumetric delivery accuracy, and ultrasonic air-in-line detection under IEC 60601-2-24 and 21 CFR 880.5725. In respiratory care, validating critical-care ventilator essential performance requires verifying delivered tidal volume accuracy, airway pressure relief, positive end-expiratory pressure stability, and fraction of inspired oxygen regulation under ISO 80601-2-12. Similarly, post-service clearance for an external defibrillator, as outlined in our analysis of defibrillator energy verification after service, demands precise measurement of delivered energy across standard resistive loads and cardioversion synchronization delays under IEC 60601-2-4.
A bedside patient monitor, by contrast, does not deliver fluids, breaths, or electrical shocks to a patient. Its essential performance is surveillance and diagnosis: faithfully acquiring physiological signals, calculating vital parameters, and immediately alerting clinical staff when patient conditions deteriorate or technical faults invalidate monitoring. Consequently, after major hardware repair or software reconfiguration, return-to-service testing is primarily an alarm-system verification problem.
The regulatory obligation governing this inspection originates in Centers for Medicare & Medicaid Services (CMS) hospital maintenance requirements. In Survey and Certification Letter S&C 14-07-Hospital (issued December 20, 2013), CMS defined hospital medical equipment as devices intended for “diagnostic, therapeutic, or monitoring care provided to a patient by a hospital” and laid down the definitive rule: “all equipment must be inspected and tested for performance and safety before initial use and after major repairs or upgrades.”
Under current CMS survey guidance in State Operations Manual Appendix A, revised by QSO-25-24 (dated 5 September 2025), Tag A-0724 interprets 42 CFR 482.41(d)(2) by stating that equipment “should be inspected and tested for performance and safety before initial use and after major repairs or upgrades.” Meanwhile, the same survey tag maintains the statutory requirement that the hospital maintenance program “must” inspect, test, and maintain equipment to ensure safety, availability, and reliability. Healthcare technology management (HTM) leaders must recognize both citations: the 2013 memo established the foundational performance-and-safety gate for monitoring equipment using “must”, while the 2025 survey manual instructs surveyors using “should” for the specific post-repair inspection sentence. Neither CMS document defines “major repair” by a dollar figure, parts list, or firmware event. Whether the work is a major repair, upgrade, or configuration change is a hospital documentation decision; when that inspect-and-test gate applies, the labeled alarm-system functions are part of the performance-and-safety inspection.
Crucially, verifying a repaired patient monitor is not an exercise in compiling model-specific error codes scraped from third-party repair forums, nor does it involve guessing internal firmware registers. It requires an evidence-based functional testing protocol that verifies the alarm system's ability to communicate urgency across visual and audible channels in accordance with recognized collateral standards.
Use IEC 60601-1-8 as the collateral map, not a decibel cookbook
The structural architecture of medical device alarm systems is governed by the horizontal collateral standard IEC 60601-1-8 (Medical electrical equipment — Part 1-8: General requirements for basic safety and essential performance — Collateral Standard: General requirements, tests and guidance for alarm systems in medical electrical equipment and medical electrical systems). Understanding how this standard applies to hospital bench testing requires examining its recognition status and its public object.
On December 21, 2020, the U.S. Food and Drug Administration (FDA) published Federal Register Recognition List 055, formally recognizing IEC 60601-1-8 Edition 2.2 2020-07 CONSOLIDATED VERSION under Recognition Number 5-131 within the General I (Quality System/Risk Management) specialty. FDA recognized the complete standard, included it within the Accreditation Scheme for Conformity Assessment (ASCA), and designated the identical national adoption as ANSI/AAMI/IEC 60601-1-8:2006 and A1:2012 including AMD 2:2021. Internationally, the IEC webstore catalog identifies Edition 2.2 as published on July 23, 2020, with a formal stability date through 2028.
Consensus standard recognition under section 514(c) of the Federal Food, Drug, and Cosmetic Act is a premarket regulatory mechanism. As articulated in FDA's guidance document, Appropriate Use of Voluntary Consensus Standards in Premarket Submissions for Medical Devices, recognition allows medical device manufacturers to declare conformity during premarket reviews (such as 510(k) clearances). Premarket recognition does not convert a 254-page laboratory type-testing standard into a hospital biomedical routine maintenance procedure, nor does it establish in-service survey tolerances.
Instead of serving as a field testing protocol, the public object of IEC 60601-1-8 provides clinical engineering with an indispensable functional map. The standard establishes basic safety and essential performance for alarm systems by defining three distinct functional families:
Alarm Categories (Priorities) by Degree of Urgency: Classifying alarm conditions as high, medium, or low priority by degree of urgency, using the priority presentation the OEM labeling specifies—not an invented harm-delay matrix or numeric timeout.
Consistent Alarm Signals: Generating the labeled audible and visual alarm signals the device is supposed to produce, without treating paywalled flash-rate, pulse-pattern, or color-coding tables as in-service tolerances.
Consistent Control States and Marking: Honoring the marked control states the OEM IFU names for inactivation and inhibit—commonly pause, audio-off, alarm-off, and reset—without inventing timeout seconds.
A common pitfall in biomedical shops is treating IEC 60601-1-8 as an in-service acoustic cookbook. Audio-component vendors republish paywalled pulse-frequency and sound-pressure figures for speaker selection. Those numbers are not verified here and are not hospital pass/fail limits. Clinical engineering benches do not become type-test laboratories by quoting vendor notes. Verify that the monitor generates the labeled audible and visual signals and respects marked control states; leave quantitative acoustic limits to the manufacturer’s validated procedures and the OEM IFU.
Confirm the device is a multifunction patient monitor, not every alarmed ME device
Because IEC 60601-1-8 is a horizontal collateral standard, FDA Recognition Number 5-131 does not map to a single device product code. Premarket testing laboratories evaluate ventilators, dialysis consoles, infusion pumps, and monitors against 1-8 under the same umbrella. To define the specific testing boundary for a bedside monitor, clinical engineering must turn to the particular standard governing multifunction patient monitors.
The current FDA-recognized particular standard is IEC 80601-2-49 Edition 1.1 2024-09 CONSOLIDATED VERSION (Medical electrical equipment — Part 2-49: Particular requirements for the basic safety and essential performance of multifunction patient monitors). FDA recognized this standard on FR Recognition List 065 as Recognition Number 3-203 on December 22, 2025, included it in ASCA, and granted it partial recognition.
The partial nature of FDA's recognition is a specific technical point that clinical engineers must understand. The sole non-recognized part of IEC 80601-2-49 is Subclause 202.8.101, which defines electromagnetic compatibility (EMC) immunity test setup geometry—specifically the test table height in Figure 202.101—because it conflicts with IEC 61000-4-6. This non-recognition is purely an EMC type-testing laboratory setup issue. It is not an exemption from alarm testing, nor does it permit technicians to invent loose parameter accuracy windows.
The public scope of IEC 80601-2-49 defines the precise boundary of equipment covered by this return-to-service framework:
Multifunction Scope: Applies to multifunction patient monitors intended for use in professional healthcare facilities, emergency medical service (EMS) environments, or home healthcare settings, provided the equipment connects to a single patient and incorporates two or more physiological monitoring units (with a pregnant mother and her fetus or fetuses legally treated as a single patient).
Individual Parameter Exclusions: The standard explicitly excludes stand-alone physiological monitoring units—such as stand-alone electrocardiographs, stand-alone invasive blood pressure monitors, or stand-alone pulse oximeters—which remain governed by their own individual particular standards.
Implantable Device Exclusions: Implantable parts and devices are entirely excluded from the scope.
The evolutionary history of IEC 80601-2-49 reveals a crucial architectural transition. In public statements accompanying Edition 1.0 (2018), which replaced IEC 60601-2-49:2011, the drafting committee highlighted that the most substantial modifications occurred in Clause 208 (Alarm Systems), because many historical particular alarm requirements were harmonized into and superseded by the collateral standard IEC 60601-1-8. Consequently, verifying alarms on a modern multifunction bedside monitor is fundamentally a 1-8 collateral verification problem executed through the OEM instructions for use, rather than a replay of outdated 2011 particular standard requirements.
When clinical engineering services a dedicated, stand-alone hospital ECG monitor rather than a multifunction unit, the applicable particular standard is IEC 60601-2-27 Edition 3.0 2011-03 (FDA Recognition Number 3-126, recognized in part for an EMC patient-cable exemption). IEC 60601-2-27 explicitly excludes home-use ECG monitors, ambulatory ECG (Holter) systems, telemetry-only transmitters, fetal heart rate monitors, and pulse plethysmographs. Establishing this device boundary ensures that technicians do not apply bedside monitor criteria to telemetry transmitters or consumer-grade ambulatory trackers.
What the CFR already names as the alarm job
While international consensus standards define technical architectures, Title 21 of the Code of Federal Regulations establishes the legal identity and clinical purpose of physiological monitoring devices in the United States.
Under 21 CFR 870.1025, an Arrhythmia detector and alarm (including ST-segment measurement and alarm) is legally identified as a device that monitors an electrocardiogram and is designed to produce a “visible or audible signal or alarm when atrial or ventricular arrhythmia, such as premature contraction or ventricular fibrillation, occurs.” The regulation classifies these devices as Class II (special controls) and designates the FDA Class II Special Controls Guidance Document: Arrhythmia Detector and Alarm as the applicable special control.
The public scope of FDA's arrhythmia special controls guidance governs devices classified under product codes DSI (arrhythmia detector and alarm), MLD (ST-segment monitor with alarm), and MHX (patient physiological monitor with arrhythmia detection or alarms). The guidance explicitly notes that MHX devices frequently incorporate other monitoring components—such as non-invasive blood pressure (21 CFR 870.1130), pulse oximetry (21 CFR 870.2700), capnography, and temperature—which each require substantial equivalence demonstrations. Nevertheless, the legal identity of an MHX device centers on its ability to generate visible or audible alarms in response to cardiac arrhythmias.
Similarly, 21 CFR 870.2300 defines a Cardiac monitor (including cardiotachometer and rate alarm) as a Class II device used to measure heart rate from an analog signal produced by an electrocardiograph, vectorcardiograph, or blood pressure monitor. The regulation explicitly states that the device “may sound an alarm when the heart rate falls outside preset upper and lower limits.” In FDA's recognition database, Recognition Number 3-203 maps 870.2300 not only to stand-alone rate alarms (product code DRT) but also to product code MSX (system, network and communication, physiological monitors).
These statutory classifications define the public clinical functions in scope: 21 CFR 870.1025 names visible or audible arrhythmia-alarm generation, and 21 CFR 870.2300 says the device may sound an alarm when heart rate falls outside preset upper and lower limits. They do not publish in-service algorithm sensitivity percentages, nor do they authorize technicians to invent heart-rate tolerance bands. Furthermore, the inclusion of MSX networked communications does not authorize a biomedical technician to perform unauthorized network vulnerability scans, penetration tests, or security bypass routines during bench return-to-service. Network reconnection checks must be strictly confined to verifying that the monitor successfully negotiates communication with its authorized central station and faithfully delivers distributed alarm signals as specified by the manufacturer.
Priority, sound, and inhibit: verify the labeled states, not invented limits
A compliant after-service alarm testing procedure focuses on verifying the three core functional families labeled on the device and described in the OEM instructions for use:
Priority Presentation: Verify that the conditions the OEM IFU requires you to present produce the priority presentation that labeling names for those conditions. High, medium, and low are urgency families, not a license to invent red or yellow color rules, flash rates, or which arrhythmia must map to which priority on every model.
Audible and Visual Signal Generation: Confirm that the labeled audible and visual signals still generate. Watch and listen for the signals the IFU names at the indicator and volume states that labeling specifies. Do not convert that check into a sound-pressure measurement or an invented brightness or volume pass window.
Control-State Logic and Marked Inactivation: Exercise the inhibit, pause, off, and reset controls the OEM IFU names, and confirm that markings and restoration behavior match that labeling for the software version and alarm profile under test. Do not invent silence-timer seconds or require a particular icon unless the IFU specifies it.
The question then arises: where do the quantitative pass/fail acceptance criteria come from if technicians are prohibited from inventing tolerances? The answer lies in official manufacturer labeling. In FDA's final guidance issued on May 10, 2024, Remanufacturing of Medical Devices: Guidance for Industry, Entities That Perform Servicing or Remanufacturing, and FDA Staff, the agency recommends that OEM labeling for reusable medical devices provide:
Key performance and safety specifications.
Device-specific performance specifications.
Recommended routine testing, troubleshooting, and acceptance criteria to confirm the device remains within specifications.
A comprehensive description of error codes, alerts, and alarm features on the device.
Software version numbers and release dates required to return the device to performance specifications.
In accompanying announcements, FDA emphasized that providing these servicing instructions is a best practice to ensure devices are safely returned to OEM specifications without disclosing proprietary trade secrets. Consequently, the specific acceptance criteria for which alarm conditions must trigger, how visual indicators flash, and how control states behave must be obtained directly from the OEM service manual and instructions for use (IFU) for the specific model and software build under test. This article borrows only that reusable-device labeling recommendation; it does not retell the six-principle servicing-versus-remanufacturing decision tree.
To present physiologic or technical conditions, use the patient simulator or equivalent the OEM IFU names. Test-equipment vendors publicly describe simulators that can present ECG, arrhythmia, apnea, NIBP, and SpO2 conditions so a monitor can be checked against those presented conditions. That is useful context for typical function families, not a CMS or IEC 60601-1-8 checklist and not a substitute for the IFU.
While patient simulators are indispensable tools for functional verification, technicians must maintain a strict methodological boundary: a simulator check is not a clinical accuracy study. Connecting an SpO2 finger sensor to an optical simulator or connecting an NIBP cuff port to a pressure simulator verifies that the monitor's acquisition hardware detects signals and that alarm thresholds trigger at preset limits. It does not validate clinical SpO2 accuracy as a pulse-oximeter clinical study, nor does it validate clinical blood pressure determination under ISO 81060-2. Documenting return-to-service requires recording that the monitor detected the presented conditions and triggered alarms in accordance with OEM criteria, rather than claiming clinical calibration.
Detectability is an accreditation check, not an alarm-fatigue essay
In addition to CMS regulations and consensus standards, hospital clinical engineering departments must satisfy healthcare accreditation standards governing clinical alarms.
The Joint Commission's National Patient Safety Goal NPSG.06.01.01, the 2025 hospital-program goal to improve the safety of clinical alarm systems, requires hospitals to establish written policies addressing:
Clinically appropriate alarm settings for specific care units.
Clear criteria determining when alarm signals can be disabled.
Specific clinical authority regarding who is permitted to change alarm parameters or set them to “off.”
Protocols for monitoring and responding to alarm signals.
Checking individual alarm signals for “accurate settings, proper operation, and detectability,” with an explicit cross-reference to Environment of Care Standard EC.02.04.03.
Under The Joint Commission's Accreditation 360 program, effective January 2026, this requirement transitioned to National Performance Goal NPG.01.05.01. The relocated standard retains the identical mandate: hospitals must identify the most important alarm signals to manage and ensure that policies govern alarm disabling, authority to turn alarms off, and individual signal detectability.
Clinical alarm literature frequently focuses on alarm fatigue—the phenomenon where clinical staff become desensitized to frequent, non-actionable alarms. While alarm fatigue is a vital clinical workflow issue, clinical engineers must distinguish between hospital operational policy and after-service equipment verification. HTM's role during return-to-service is not to rewrite nursing alarm escalation protocols, but to provide objective technical evidence that the monitor's alarm hardware and software deliver audible and visual signals that are fully operational, accurately configured, and physically detectable.
Under Joint Commission Standard EC.02.04.03 (and the 2025 Hospital Life Safety & Environment of Care Document List and Review Tool), hospitals must maintain a written inventory of medical equipment and achieve a 100% completion rate for required inspection, testing, and maintenance of high-risk equipment. High-risk equipment is defined as medical equipment for which failure poses a risk of serious injury or death to patients or staff, including life-support equipment.
Here, clinical engineers must avoid a critical classification error: physiologic patient monitors are not automatically classified as life-support equipment. FDA's regulatory classification for product code MHX establishes that the generic device type is not life-sustaining or life-supporting. Whether a bedside monitor is categorized as high-risk depends upon the hospital's written Medical Equipment Management Plan (MEMP) and its formal risk-scoring taxonomy (evaluating equipment function, physical risk, and maintenance history). If the hospital classifies a monitor as high-risk, it must achieve the 100% completion rate, but technicians must not mislabel general monitoring devices as life-support assets on service work orders.
Electrical safety is required and not sufficient
After repair, electrical-safety testing remains a necessary technical checkpoint under the hospital’s already-named in-service method. In-service electrical safety is typically governed by recognized standards such as IEC 62353:2014 (Medical electrical equipment — Recurrent test and test after repair of medical electrical equipment) or NFPA 99 (Health Care Facilities Code, incorporated by CMS into 42 CFR 482.41(c)). For a detailed examination of standard selection, refer to our analysis of electrical safety testing after repair.
The public scope of IEC 62353 explicitly applies to testing medical electrical equipment before initial use, during routine preventive maintenance, and after repair to assess safety. Crucially, the standard's public abstract notes that it “does not define requirements for repair, exchange of components and modification of ME equipment or ME systems” and presumes that repairs executed in accordance with manufacturer instructions maintain conformity to original design standards.
A dangerous misconception in biomedical servicing is the belief that passing an electrical safety test—such as verifying protective earth continuity and chassis leakage current—constitutes complete return-to-service clearance. Electrical safety testing verifies only that the equipment does not present an electric shock hazard to patients or operators. It provides zero evidence regarding essential monitoring performance or alarm functionality.
A repaired patient monitor can display flawless ground resistance and virtually non-existent leakage current while suffering from fatal functional defects: a disconnected internal speaker wire, a dead visual alarm beacon LED, an uncalibrated analog front-end that fails to detect ventricular fibrillation, or a corrupted software profile that sets alarm volumes to zero. Electrical safety testing is a necessary prerequisite, but it is never sufficient on its own. It must be paired with active alarm-system verification.
A scope-and-evidence matrix, not an error-code library
To establish a rigorous, audit-ready return-to-service protocol, healthcare technology management departments should implement a structured four-layer verification workflow. The following diagram illustrates how standards, regulations, and manufacturer criteria integrate to release a repaired monitor to clinical service:
flowchart TD
subgraph Scope["Layer 1: Scope"]
A["Bedside or transport monitor after service"] --> B{"Which particular-standard identity?"}
B -->|"IEC 80601-2-49 / Rec# 3-203: multifunction, 2+ units, single patient"| C["Multifunction monitor track"]
B -->|"IEC 60601-2-27 / Rec# 3-126: hospital stand-alone ECG"| D["Stand-alone hospital ECG track"]
B -->|"Home, Holter, fetal, telemetry-only, implantable"| E["Out of this article's scope"]
end
subgraph Trigger["Layer 2: CMS trigger"]
C --> F{"Inspect-and-test after major repair or upgrade?"}
D --> F
F -->|"S&C 14-07 used must; QSO-25-24 A-0724 uses should"| G["Performance and safety inspection before return to service"]
end
subgraph Execution["Layer 3: Parallel gates"]
G --> H["Electrical safety under the hospital named in-service method"]
G --> I["Alarm-system verification against the OEM IFU"]
I --> I1["Priority presentation as labeled"]
I --> I2["Audible and visual signals the IFU requires"]
I --> I3["Labeled inhibit, pause, off, and reset states"]
end
subgraph Documentation["Layer 4: Work-order evidence"]
H --> J["Compile work-order evidence"]
I1 --> J
I2 --> J
I3 --> J
J --> K["Identity, software version, alarm profile, conditions presented, results versus OEM criteria, instruments, authorization"]
K --> L["Authorized return to the assigned unit"]
endThe technical relationships governing this verification workflow are summarized in the following evidence matrix, delineating the specific role and boundaries of each standard and regulatory authority:
| Framework / Authority | Role in Return-to-Service | Verified Alarm Functions | Governing Standard / Limits | Prohibited In-Service Assumptions |
|---|---|---|---|---|
| IEC 60601-1-8 Edition 2.2 (FDA Rec# 5-131) | Horizontal Collateral Map | Priorities by urgency (high, med, low); consistent audible and visual signals; marked control states (pause, off, reset). | FDA-reprinted public scope and IEC webstore abstract; complete recognition on FR List 055; included in ASCA. | Do not invent decibel levels, pulse frequency tolerances, or delay limits from paywalled type tests. |
| IEC 80601-2-49 Edition 1.1 (FDA Rec# 3-203) | Multifunction Particular Standard | Identifies multifunction bedside monitor device class; delegates alarm requirements to IEC 60601-1-8 under Clause 208. | Public scope: 2+ physiological units, single patient. Recognized in part (EMC table height excluded). | Do not treat EMC table height exclusion as an alarm waiver; do not invent parameter measurement accuracy windows. |
| 21 CFR 870.1025 & 870.2300 (MHX, MLD, DRT, MSX) | Statutory Device Identity | Names visible or audible arrhythmia-alarm generation (870.1025) and optional rate-limit alarms (870.2300) as legal identity, not in-service algorithm or bpm tolerances. | Class II special controls; MHX encompasses multi-parameter units; MSX covers networked central communications. | Do not invent algorithmic sensitivity tables; do not execute network penetration attacks under the guise of MSX testing. |
| CMS S&C 14-07 & QSO-25-24 (Tag A-0724) | Hospital Regulatory Trigger | S&C 14-07 used “must” for inspect-and-test after major repair or upgrade; current A-0724 uses “should” for that sentence. | 42 CFR 482.41(d)(2); S&C 14-07 uses “must”; QSO-25-24 survey guidance uses “should”. | Do not define major repair by dollar or parts thresholds; do not treat a power-on self-test or electrical-safety result as alarm-system evidence. |
| Joint Commission NPSG.06.01.01 / NPG.01.05.01 | Accreditation Compliance | Requires checking individual alarm signals for accurate settings, proper operation, and physical detectability. | Accreditation 360 NPG.01.05.01 (2026); EC.02.04.03 100% completion for high-risk equipment. | Do not automatically classify MHX monitors as life-support equipment; follow hospital MEMP risk inventory. |
| FDA May 2024 Remanufacturing Guidance | Service Labeling Authority | Identifies OEM labeling as the source of routine testing, troubleshooting, acceptance criteria, and error-code descriptions. | FDA final guidance on medical device remanufacturing and reusable device servicing instructions. | Do not invent third-party acceptance criteria when OEM documentation specifies validated test procedures. |
| IEC 62353:2014 / NFPA 99 (2012 Edition) | Electrical Safety Gate | After-repair electrical-safety testing under the hospital’s named in-service method; not a substitute for alarm-system verification. | In-service electrical safety standards; incorporated into CMS CoP physical environment requirements. | Never treat passing an electrical safety test as evidence of functional alarm integrity or sensor operation. |
To ensure complete regulatory audit readiness under CMS Tag A-0724, Joint Commission EC.02.04.01, and ISO 13485 servicing record controls (as detailed in our guide on medical equipment service record requirements and our analysis of recording configuration after a software restore), the technician must capture specific data fields on the CMMS work order before releasing the device:
Asset Unique Identification: Document the internal asset control number, manufacturer serial number, equipment make, model, and assigned clinical department.
Software and Firmware Baseline: Record the primary operating system version, installed module firmware versions, and configuration profile build date.
Active Clinical Profile and Alarm Presets: Identify the clinical profile or alarm preset selected during verification, as named in the hospital configuration or OEM IFU, and record that it was the profile actually tested.
Simulated Conditions and Test Methods: Record the physiologic and technical conditions the OEM IFU required the technician to present, and the simulator or equivalent actually used, without converting those presentations into invented bpm or saturation windows.
Observed Functional Alarm Responses: Document pass/fail against the OEM acceptance criteria for priority presentation, audible and visual signal generation, and labeled inhibit, pause, off, and reset states.
Test Instrument Traceability: Record the model, serial number, and current calibration expiration date for the patient simulator and the electrical safety analyzer used during testing.
Electrical Safety Verification Data: Record the electrical-safety method the hospital already uses after repair and the results against that method’s criteria. Those results are not alarm-system evidence.
Authorizing Service Signature: Capture the name, signature, and date of the qualified biomedical equipment technician authorizing return to clinical operation.
A repaired physiologic monitor is ready for a bedside or transport assignment when the work order can show that the labeled alarm system still presents priority, generates the audible and visual signals the IFU requires, and honors marked inhibit and inactivation states, with electrical safety handled as a parallel gate. That is an evidence job, not a compliance guarantee and not a model-specific error-code library.
