How should a medical-device clinical investigation plan, informed consent document, adverse event (AE), serious adverse event (SAE) report or device deficiency record be translated under ISO 14155:2026? People searching for ISO 14155 translation, medical device clinical trial document translation, CIP terminology, AE versus SAE meaning and multilingual informed consent need a target-language record that preserves participant protections, study methods and the precise status of safety evidence.
This educational guide examines the March 2026 fourth edition of ISO 14155, which replaced ISO 14155:2020, and shows how to preserve the language of clinical investigation plans (CIPs), sponsor and investigator roles, subject communication, adverse events, protocol deviations, device deficiencies, data handling, ethical review and study reporting. It is designed for translators, sponsors, study personnel, ethics-review teams, technical writers and learners.
Translation is not medical judgment or regulatory authorization. A linguist must never reclassify a reported event, change eligibility criteria, weaken a warning, invent a reporting deadline or make an unapproved claim of device safety. Applicable laws, ethics approval, approved study records and qualified clinical oversight control real activities. The examples below are fictional teaching scenarios and never replace professional review.
Why the 2026 edition matters
ISO 14155:2026 is the current fourth edition of the good clinical practice standard for clinical investigations of medical devices involving human subjects. It replaced the 2020 edition in March 2026. Edition control is essential because training documents, contracts and templates may still cite the older edition.
A sponsor may use a 2020-era translated consent template while updating its study plan for the new edition. The translator should not merely change 2020 to 2026 in the header; any required technical revision belongs to an authorized study team.
Show both the original source edition and the actual target-document revision. If a governing regulatory approval still refers to a legacy version, preserve that fact while escalating the need for alignment.
The safety-review question is: Can a reviewer tell which standard edition the original and translated study records invoke? Check the answer against the exact approved source, not a generic assumption about clinical practice. Record the accountable clinical, ethics or regulatory reviewer and any question still pending. A translation can be clear without becoming an unsupported medical judgment, and it should preserve that boundary in every language.
Good clinical practice is about evidence and participant protection
Good clinical practice in this context concerns the ethical and scientific conduct, recording and reporting of clinical investigations intended to assess medical-device safety and clinical performance or effectiveness. It is not a marketing badge guaranteeing that a product is safe or approved for every use.
A supplier brochure saying clinically proven should not be translated as ISO 14155 certified unless the source substantiates such a specific claim. Completion of a clinical investigation and authorization to market a device are separate questions.
Keep approved intended purpose, investigation scope, study status and any safety limitation distinct from promotional statements. Avoid making stronger claims than the source makes.
At handover, ask the responsible reviewer: Does the translation state what the study investigated rather than implying a universal guarantee? Check the answer against the exact approved source, not a generic assumption about clinical practice. Record the accountable clinical, ethics or regulatory reviewer and any question still pending. A translation can be clear without becoming an unsupported medical judgment, and it should preserve that boundary in every language.
The scope is medical-device clinical investigation
The standard addresses investigations of medical devices involving human subjects. It does not generally apply to in vitro diagnostic medical devices, while particular situations and national rules can affect what principles are considered relevant. Translators should not invent automatic applicability to unrelated research.
A document concerning an investigational implant, an imaging device and a laboratory-only assay may require different regulatory context. The title clinical investigation alone does not settle the classification.
Retain the product classification, intended use, investigational status, sponsor’s regulatory basis and applicable jurisdictional references. Ask the authorized regulatory team when scope is unclear.
Before the document is authorized, verify: Would the target reader know which device and investigation are actually being governed? Check the answer against the exact approved source, not a generic assumption about clinical practice. Record the accountable clinical, ethics or regulatory reviewer and any question still pending. A translation can be clear without becoming an unsupported medical judgment, and it should preserve that boundary in every language.
Clinical investigation plan is not a generic business plan
The clinical investigation plan, often abbreviated CIP, describes the design, conduct and analysis intentions of a particular investigation. It may define objectives, population, endpoints, procedures, monitoring and data handling. Its details establish the study that is actually approved or proposed.
If an investigator receives CIP revision 4 while a translated site manual still cites revision 2, the procedures may differ. A fluent translation cannot resolve that mismatch by choosing the newest-looking paragraph.
Preserve the CIP title, study identifier, protocol revision, date and approval status. Use consistent terminology between the CIP and every controlled supporting document.
The safety-review question is: Does every translated instruction point to the correct currently approved investigation plan? Check the answer against the exact approved source, not a generic assumption about clinical practice. Record the accountable clinical, ethics or regulatory reviewer and any question still pending. A translation can be clear without becoming an unsupported medical judgment, and it should preserve that boundary in every language.
Investigational device versus comparator
An investigational medical device is studied within the investigation, while a comparator may supply a reference for assessment. A product under investigation is not automatically approved for commercial use. The distinction affects labels, procedures and how claims are reported.
A trial packet might contain two devices with similar appearances and different study roles. Translating comparator as test device can misidentify which device generated a reported performance result.
Protect product identifiers, serial or lot references, intended purpose, study arm and authorization wording. Use a controlled termbase rather than substituting device synonyms for style.
At handover, ask the responsible reviewer: Could a study-team member distinguish the investigational device from the comparator in the target language? Check the answer against the exact approved source, not a generic assumption about clinical practice. Record the accountable clinical, ethics or regulatory reviewer and any question still pending. A translation can be clear without becoming an unsupported medical judgment, and it should preserve that boundary in every language.
Sponsor and principal investigator have separate duties
The sponsor and principal investigator have different responsibilities in a clinical investigation. Translating both as organizer or lead researcher can obscure who initiates, manages, conducts, oversees or reports particular activities.
Suppose the sponsor is responsible for a regulatory submission while the principal investigator is responsible for oversight at a site. If the translated instruction assigns both to the same unnamed leader, an essential handoff disappears.
Retain the formal role names, site affiliations, delegation records and applicable responsibility matrix. A local job title must be mapped to the defined study role before being used as an equivalent.
Before the document is authorized, verify: Does the translated workflow identify the accountable party for each decision? Check the answer against the exact approved source, not a generic assumption about clinical practice. Record the accountable clinical, ethics or regulatory reviewer and any question still pending. A translation can be clear without becoming an unsupported medical judgment, and it should preserve that boundary in every language.
Ethics committee is not merely an administrative office
An ethics committee has a defined role in reviewing participant-related ethical aspects and study documentation. A translated reference should not imply that routine administrative registration is equivalent to ethical approval.
A document may state submitted to the ethics committee, another may state approved, and a third may state approval pending. All three statuses mean different things to the receiving study team.
Keep the committee identity, reference number, submission date, approval conditions, correspondence and expiry or continuing-review requirements exactly as documented.
The safety-review question is: Can a reader tell whether the relevant document is under review, conditionally approved or finally approved? Check the answer against the exact approved source, not a generic assumption about clinical practice. Record the accountable clinical, ethics or regulatory reviewer and any question still pending. A translation can be clear without becoming an unsupported medical judgment, and it should preserve that boundary in every language.
Informed consent is a process, not only a signed page
Informed consent involves communicating relevant information and supporting a participant’s informed decision under applicable requirements. A signature page records part of the process but does not, by itself, make a difficult or misleading translation acceptable.
A translated form may use a familiar everyday word for treatment that falsely suggests therapeutic benefit from an investigational device. Such a change could undermine a participant’s understanding of the research nature of participation.
Use culturally clear language without weakening or adding study promises. Where a validated or approved target-language version is required, follow that process before anyone presents it to prospective participants.
At handover, ask the responsible reviewer: Can a prospective participant distinguish research participation from guaranteed treatment benefit? Check the answer against the exact approved source, not a generic assumption about clinical practice. Record the accountable clinical, ethics or regulatory reviewer and any question still pending. A translation can be clear without becoming an unsupported medical judgment, and it should preserve that boundary in every language.
Voluntary participation and withdrawal wording
Consent materials commonly explain voluntary participation and how withdrawal works within the approved investigation and local law. Words like may, can, without penalty and subject to particular data-handling conditions require exact interpretation rather than blanket reassurance.
A source might explain that a participant can stop procedures but that previously collected data may be handled under specific lawful conditions. Translating this simply as all data will be deleted immediately could create a promise the study cannot fulfil.
Preserve each applicable condition, consequence and contact channel. Have qualified ethics and privacy reviewers evaluate the final target-language wording.
Before the document is authorized, verify: Does the translation maintain the precise rights and limits described in the approved consent material? Check the answer against the exact approved source, not a generic assumption about clinical practice. Record the accountable clinical, ethics or regulatory reviewer and any question still pending. A translation can be clear without becoming an unsupported medical judgment, and it should preserve that boundary in every language.
Risks, benefits and uncertainty must stay balanced
Study communications need to describe known or reasonably anticipated risks and potential benefits without overstating certainty. Translators sometimes smooth cautious language into promises or soften serious warnings to sound comforting. Both approaches can misinform participants.
A statement that benefit is not guaranteed cannot become the device will improve your condition. Similarly, a risk labelled possible should not be translated as impossible merely because the translator thinks it unlikely.
Check each qualifier, conditional sentence and comparative claim. Preserve both the severity and the uncertainty of the source without adding clinical advice.
The safety-review question is: Would a reader in the target language perceive the same level of uncertainty and seriousness? Check the answer against the exact approved source, not a generic assumption about clinical practice. Record the accountable clinical, ethics or regulatory reviewer and any question still pending. A translation can be clear without becoming an unsupported medical judgment, and it should preserve that boundary in every language.
Eligibility criteria are decision rules
Inclusion and exclusion criteria define who may enter an investigation under its approved design. Small words such as at least, more than, prior, within and not can decide whether a person meets a condition. Translation must protect these boundaries.
If a protocol includes an age threshold or a specified time window for a prior procedure, changing ≥ to > or twelve weeks to twelve months alters the defined population.
Verify numerical criteria, units, conditions, exclusion logic and source amendments. A qualified investigator decides eligibility; a translator does not infer exceptions.
At handover, ask the responsible reviewer: Do the target-language criteria select exactly the same eligible and ineligible groups? Check the answer against the exact approved source, not a generic assumption about clinical practice. Record the accountable clinical, ethics or regulatory reviewer and any question still pending. A translation can be clear without becoming an unsupported medical judgment, and it should preserve that boundary in every language.
Endpoints and outcomes must not be merged
An endpoint is a study-defined quantity or event used for evaluation. Primary and secondary endpoints can have distinct statistical and interpretive roles. A result reported for a secondary endpoint must not be translated as the principal demonstrated outcome.
A study may record an improvement on an exploratory measure while missing its primary objective. An overconfident summary saying trial success erases an important distinction.
Retain endpoint name, time point, analysis population, statistical qualifier and actual status. Leave claims of clinical significance to qualified assessment.
Before the document is authorized, verify: Can the reader distinguish what was planned to be measured from what was actually observed? Check the answer against the exact approved source, not a generic assumption about clinical practice. Record the accountable clinical, ethics or regulatory reviewer and any question still pending. A translation can be clear without becoming an unsupported medical judgment, and it should preserve that boundary in every language.
Clinical performance and effectiveness language
Clinical performance and effectiveness can have defined contextual meanings in medical-device investigations. A device’s technical accuracy, user experience and clinical result are related but not identical questions. Generic synonyms such as works well can hide the specific claim.
A study report about diagnostic measurements should not be translated into a claim that patient outcomes improved unless those outcomes were actually investigated and supported.
Preserve the precise evaluation objective, result metric and scope. Separate analytical measurement, clinical performance, usability and broader therapeutic claims.
The safety-review question is: Has the translation kept the type of evidence attached to each performance statement? Check the answer against the exact approved source, not a generic assumption about clinical practice. Record the accountable clinical, ethics or regulatory reviewer and any question still pending. A translation can be clear without becoming an unsupported medical judgment, and it should preserve that boundary in every language.
Adverse event and adverse device effect
An adverse event is a clinical event recorded in connection with a subject or investigation under the applicable definitions. It is not automatically proven to have been caused by the investigational device. An adverse device effect involves an assessed relationship under the governing classification.
A subject experiences a symptom during follow-up. The initial event report may record it while causal assessment remains pending. Translating every event as device-caused harm inserts a conclusion that may not exist.
Keep the exact event term, onset and resolution times, attribution assessment, seriousness and outcome separately. Qualified clinical personnel make causal classifications.
At handover, ask the responsible reviewer: Does the target maintain the distinction between occurrence and assessed device relationship? Check the answer against the exact approved source, not a generic assumption about clinical practice. Record the accountable clinical, ethics or regulatory reviewer and any question still pending. A translation can be clear without becoming an unsupported medical judgment, and it should preserve that boundary in every language.
Serious adverse events require exact qualifiers
SAE is not interchangeable with every adverse event or with the everyday adjective severe. Seriousness is determined under applicable definitions and circumstances; severity can describe intensity. Translation must not replace one classification with another.
A severe headache may not automatically have the same regulatory classification as a serious adverse event, and an event can be serious without being described by the strongest pain adjective.
Protect the original classification, supporting criteria, medical outcome, investigator judgment and reporting status. Never invent a reportability conclusion or statutory deadline.
Before the document is authorized, verify: Could the reviewer recover both the severity description and the regulatory seriousness category? Check the answer against the exact approved source, not a generic assumption about clinical practice. Record the accountable clinical, ethics or regulatory reviewer and any question still pending. A translation can be clear without becoming an unsupported medical judgment, and it should preserve that boundary in every language.
Device deficiency is not always an injury
A device deficiency may involve inadequacy of identity, quality, durability, reliability, usability, safety or performance, including issues involving use and information supplied. It can arise without a patient injury. Misreading deficiency as adverse event can distort the safety record.
A monitor displays an incorrect status message but no subject is harmed. The incident might still need assessment as a device deficiency, yet the translation should not invent a clinical injury.
Preserve device ID, malfunction description, use context, user action, observed consequences and classification status. Refer reportability and investigation decisions to the responsible professionals.
The safety-review question is: Can the target reader distinguish product inadequacy from an event affecting a subject? Check the answer against the exact approved source, not a generic assumption about clinical practice. Record the accountable clinical, ethics or regulatory reviewer and any question still pending. A translation can be clear without becoming an unsupported medical judgment, and it should preserve that boundary in every language.
Protocol deviation is not automatically misconduct
A protocol deviation is a departure from requirements of the clinical investigation plan under the applicable framework. The description may range from a minor scheduling issue to a potentially important effect on participant protection or data reliability. Translation should not turn every deviation into deliberate wrongdoing.
A visit occurred two days outside a study window. If the investigation team is assessing its significance, a translation that calls it fraud or a critical violation invents a conclusion.
Record the actual departure, date, cause as known, affected subject or data, corrective action and classification status. Preserve the responsible investigator’s assessment.
At handover, ask the responsible reviewer: Does the wording describe the observed departure without overstating intent or consequence? Check the answer against the exact approved source, not a generic assumption about clinical practice. Record the accountable clinical, ethics or regulatory reviewer and any question still pending. A translation can be clear without becoming an unsupported medical judgment, and it should preserve that boundary in every language.
Reporting deadlines are jurisdiction-dependent
Standards, local law, regulatory approvals, ethics requirements and sponsor procedures can interact in defining reporting responsibilities. A universal deadline guessed by a translator would be particularly dangerous in a safety workflow.
A sponsor may require internal notification immediately while a regulator has a different legally defined timeline for a classified event. These are distinct obligations and may start from different triggering knowledge.
Translate exact deadlines, recipients, start conditions and jurisdiction names from the approved source. If one is missing, flag the omission for regulatory review instead of filling it from memory.
Before the document is authorized, verify: Can the study team tell exactly which deadline applies to which recipient and event? Check the answer against the exact approved source, not a generic assumption about clinical practice. Record the accountable clinical, ethics or regulatory reviewer and any question still pending. A translation can be clear without becoming an unsupported medical judgment, and it should preserve that boundary in every language.
Data monitoring committee and clinical events committee
The 2026 edition includes updated attention to oversight structures, including data monitoring committees and clinical events committees. These bodies have particular study-related purposes and should not be collapsed into one generic independent review team.
A document may explain why no data monitoring committee was established or describe criteria considered for a clinical events committee. The translation must preserve the nature of the decision and the responsible sponsor’s justification.
Keep the committee designation, charter references, independence arrangements, scope, decisions and pending questions. Do not assert that every study is required to establish an identical committee without reading the approved obligations.
The safety-review question is: Can the reader distinguish the roles, presence and documented rationale of each oversight body? Check the answer against the exact approved source, not a generic assumption about clinical practice. Record the accountable clinical, ethics or regulatory reviewer and any question still pending. A translation can be clear without becoming an unsupported medical judgment, and it should preserve that boundary in every language.
The 2026 risk-based approach
Clinical investigation oversight and risk management should be interpreted with the current edition’s risk-based and proportionate context. Device-related residual risks and risks from clinical procedures require careful attention, without assuming they are governed identically.
If a protocol involves an additional invasive procedure performed solely for the study, the related participant risks should not disappear into the product’s technical risk summary in translation.
Preserve how risks are linked to device use, study conduct and non-routine procedures. A translator must not judge that an intervention is safe or waive a control.
At handover, ask the responsible reviewer: Does the target-language risk description identify the source of risk and the relevant protection? Check the answer against the exact approved source, not a generic assumption about clinical practice. Record the accountable clinical, ethics or regulatory reviewer and any question still pending. A translation can be clear without becoming an unsupported medical judgment, and it should preserve that boundary in every language.
Investigator’s brochure and device instructions
An investigator’s brochure, instructions for use, labelling and training records may all contain information needed to conduct an investigation correctly. They are not interchangeable documents, and an approved device instruction cannot be rewritten merely because the brochure is more detailed.
A site may be asked to follow device preparation steps in one controlled instruction while a later document supplies background performance data. Mixing them into one newly invented sequence can change handling.
Maintain document identifiers, revision dates, warnings, preparation steps, storage conditions and training references. Technical reviewers should approve any localized operational instruction.
Before the document is authorized, verify: Could a site team find the correct approved instruction rather than treating background explanation as an instruction? Check the answer against the exact approved source, not a generic assumption about clinical practice. Record the accountable clinical, ethics or regulatory reviewer and any question still pending. A translation can be clear without becoming an unsupported medical judgment, and it should preserve that boundary in every language.
Case report forms and data fields
Case report forms and electronic data capture systems depend on stable field names, values, coding rules and study identifiers. Changing a category label can break comparability or cause data to be entered under the wrong meaning.
A source form distinguishes event onset, awareness date and reporting date. Translating all three as date of occurrence destroys chronology. A checkbox labelled unknown must not be interpreted as no.
Retain variable IDs, permitted responses, units, skip logic, protected codes and validation rules. Test the localized interface with representative users before live entry.
The safety-review question is: Does each target-language field still capture the same observation or decision as the source field? Check the answer against the exact approved source, not a generic assumption about clinical practice. Record the accountable clinical, ethics or regulatory reviewer and any question still pending. A translation can be clear without becoming an unsupported medical judgment, and it should preserve that boundary in every language.
Subject identifiers and confidentiality
Participant identity, coded study identifiers and confidential records require different handling. A translator should not infer or expose real identity when a source is intentionally coded, and should not substitute names for subject numbers in a translated case narrative.
A safety report may use a pseudonymized identifier to connect events without revealing a person’s name. Translating the label as patient name could encourage improper disclosure.
Keep the established identifier system, access restrictions, secure transfer instructions and privacy statements. Follow applicable data-protection arrangements and sponsor-approved workflows.
At handover, ask the responsible reviewer: Does the translation protect the same confidentiality boundary as the original? Check the answer against the exact approved source, not a generic assumption about clinical practice. Record the accountable clinical, ethics or regulatory reviewer and any question still pending. A translation can be clear without becoming an unsupported medical judgment, and it should preserve that boundary in every language.
Dates, dosage-like units and measurement precision
Clinical investigation documents may contain device settings, measurement units, visit windows and timestamps. A numeral can govern an inclusion criterion, a device operating parameter or a reported clinical observation. Changing it can distort safety or scientific meaning.
A 0.5 value with a defined unit cannot be rounded to 5, and a follow-up at week six cannot become six months. Decimal comma conventions, superscripts and timezone differences need technical checking.
Preserve values and units exactly, and keep instrument settings distinct from subject outcomes. Use a separate numeric verification process in addition to linguistic review.
Before the document is authorized, verify: Are all study-related quantities tied to the same unit, endpoint and timepoint? Check the answer against the exact approved source, not a generic assumption about clinical practice. Record the accountable clinical, ethics or regulatory reviewer and any question still pending. A translation can be clear without becoming an unsupported medical judgment, and it should preserve that boundary in every language.
Amendments, versions and reconsent
Protocol amendments, consent changes and site training updates have different effective dates. A multilingual study may require controlled target-language updates before revised materials are deployed. Reconsent obligations depend on the approved requirements and nature of changes.
If a participant received consent version three, a translated record must not imply version four was already provided. The source chronology should show what was available and what was actually completed.
Keep version numbers, amendment reasons, approval dates, affected sites, distribution and completion status. Obtain appropriate review of any new participant-facing translation.
The safety-review question is: Can the audit trail establish exactly which language and version each participant encountered? Check the answer against the exact approved source, not a generic assumption about clinical practice. Record the accountable clinical, ethics or regulatory reviewer and any question still pending. A translation can be clear without becoming an unsupported medical judgment, and it should preserve that boundary in every language.
Monitoring findings and corrective action
Monitoring observations can identify missing data, deviations, process weaknesses or unresolved source verification. A corrective-action plan differs from a closed finding. Translating planned into completed gives the false impression of resolved deficiencies.
A monitor may ask a site to reconcile two dates. Until the discrepancy has been addressed and reviewed, the file should retain its open or pending status.
Preserve finding classification, owner, deadline, source evidence, correction and closure decision. Do not infer quality-system approval from the presence of an action plan.
At handover, ask the responsible reviewer: Can the target-language audit reader distinguish observed problems from verified remediation? Check the answer against the exact approved source, not a generic assumption about clinical practice. Record the accountable clinical, ethics or regulatory reviewer and any question still pending. A translation can be clear without becoming an unsupported medical judgment, and it should preserve that boundary in every language.
Statistical conclusions need restrained language
A confidence interval, p value, subgroup analysis or descriptive estimate is not a plain-language proof of clinical benefit. Translators may be tempted to turn qualified statistical statements into unqualified patient-facing conclusions.
A report may state that results are exploratory, that a subgroup was not powered for a particular comparison or that evidence remains uncertain. These limitations must survive translation.
Keep numerical results, analysis population, timepoints, statistical qualifiers and the investigators’ actual interpretation. Seek statistical review for ambiguous terminology rather than inventing certainty.
Before the document is authorized, verify: Does the translation reflect the stated strength and limitations of the evidence? Check the answer against the exact approved source, not a generic assumption about clinical practice. Record the accountable clinical, ethics or regulatory reviewer and any question still pending. A translation can be clear without becoming an unsupported medical judgment, and it should preserve that boundary in every language.
Human comprehension testing for target audiences
A technically exact consent translation may still be difficult for prospective participants to understand. Readability matters, but simplification must not remove risks, conditions, rights or important procedures.
A target-language reader may interpret a literal term for randomization as a personal choice by the investigator. Comprehension testing can identify such misunderstandings before approved use.
Work with qualified clinical and language reviewers to use understandable approved explanations. Document changes and confirm that the revised wording still expresses the same study process.
The safety-review question is: Can the intended audience accurately explain what will happen and what choices remain theirs? Check the answer against the exact approved source, not a generic assumption about clinical practice. Record the accountable clinical, ethics or regulatory reviewer and any question still pending. A translation can be clear without becoming an unsupported medical judgment, and it should preserve that boundary in every language.
Automated translation and safety-critical proofreading
Machine translation can accelerate preliminary drafting but can change negation, conditional language, seriousness labels, device identifiers or visit timing. In clinical records, polished sentences can be dangerous if they conceal these errors.
An instruction stating do not use after a specified condition must not become use after that condition. An AE still under assessment cannot become confirmed device-caused harm because the tool preferred a shorter phrase.
Protect coded terms, numbers, warnings and participant rights through a bilingual terminology system, independent checks and qualified approval. Do not publish a machine-only output as a validated clinical document.
At handover, ask the responsible reviewer: Has a human domain reviewer checked every safety-relevant statement against the original? Check the answer against the exact approved source, not a generic assumption about clinical practice. Record the accountable clinical, ethics or regulatory reviewer and any question still pending. A translation can be clear without becoming an unsupported medical judgment, and it should preserve that boundary in every language.
Scenario: an AE is declared device-caused during translation
A site note records an adverse event with causality not yet assessed. The translator changes it to adverse device effect because the symptom occurred after device use. This creates a clinical conclusion. The correct target text preserves the recorded AE and pending assessment. The authorized investigator or sponsor process determines the eventual relationship, and the translation must follow the final documented classification.
Scenario: severe is confused with serious
A source describes a severe but non-hospitalized symptom while a different subject’s event meets a serious-event criterion. Replacing both with one generic word loses an important distinction. A termbase should separately capture intensity and the applicable regulatory classification so that reviewers can interpret the study record as written.
Scenario: a device deficiency causes no injury
A device screen displays an incorrect alert during a controlled procedure, but no injury is reported. The translator should preserve the deficiency and its actual consequences without inventing a harmed subject. The sponsor may still need to assess the event under applicable procedures; the language record must not prejudge that assessment.
Scenario: consent wording becomes a guarantee
The original form states that participation may or may not benefit a subject. A translation promises improvement. The translated text must be corrected and reviewed because it changes the basis of informed choice. Clarity does not justify adding certainty. The participant-facing version should convey the same limits and allow real comprehension.
Scenario: an old protocol is mistaken for the approved revision
A translated site packet cites CIP revision two while the sponsor’s released packet cites revision four. The correct action is to identify the mismatch, retain the audit trail and obtain an authorized updated translation. Quietly copying the revision-four label onto the old body is an uncontrolled and misleading document change.
Scenario: a reporting clock is invented
A safety email asks the site to follow the applicable urgent reporting procedure. The translator adds a fixed twenty-four-hour deadline from memory. That new line may conflict with local obligations or source policy. Remove the invented deadline and request an approved jurisdiction-specific instruction when needed.
Scenario: an electronic form changes answer meaning
A dropdown with yes, no and unknown is localized so that unknown appears to mean not applicable. Data users may then misclassify observations. The correction requires review of the field’s coded value, screen label and reporting logic, not merely a more attractive translation.
Scenario: exploratory evidence becomes a confirmed result
A clinical investigation report notes an exploratory observation that needs additional study. The translated executive summary claims the device has proved clinical superiority. Restore the source’s exploratory status and statistical qualification. A linguistic summary cannot strengthen the underlying study outcome.
Frequently asked questions
What is ISO 14155:2026?
The fourth edition of the medical-device clinical investigation good clinical practice standard, published in March 2026.
Did it replace ISO 14155:2020?
Yes. The 2020 edition is withdrawn and the 2026 edition is published.
Does ISO 14155 generally cover in vitro diagnostic devices?
Its stated scope does not generally apply to IVD devices; specific legal or study contexts need separate review.
Is an adverse event always caused by the device?
No. Occurrence and causality must not be merged without the appropriate assessment.
Are serious and severe equivalent?
No. Intensity and seriousness classification are different concepts.
Can a device deficiency exist without injury?
Yes. Deficiencies can concern device reliability, usability, information or other characteristics without a documented injury.
Can a translator decide whether an event is reportable?
No. Appropriate clinical and regulatory personnel make that decision using applicable requirements.
May patient-facing consent materials be simplified?
Only through an approved process that preserves rights, risks, conditions and meaning.
Should a protocol deviation be translated as misconduct?
Not automatically. Preserve the recorded departure and any documented assessment of seriousness or intent.
What must an electronic case report form retain?
Variable identifiers, allowed values, units, skip logic, timepoints and source meaning.
Do all studies need the same monitoring committees?
Do not assume so; use the current standard, sponsor’s documented justification and applicable requirements.
What is the safest bilingual handover?
A controlled target file with source version, reviewed terminology, numerical verification, approvals and unresolved questions.
Practical quality gate for a multilingual clinical investigation
Identify current document version and source authority; distinguish sponsor, investigator and ethics roles; preserve intended purpose and device identity; verify all CIP criteria and visit windows; compare AE, SAE, device-effect and deficiency terms with approved definitions; retain participant rights and uncertainty in informed consent; validate units, dates, codes and electronic-form fields; identify unresolved safety classifications and reporting obligations; review all amendments and committee language; and obtain qualified clinical, regulatory and linguistic approval before using the target document in an actual study. This article is educational and not a clinical instruction.
Sources and protected learning routes
Verify the current fourth edition at ISO 14155:2026 and read the dated April 2026 overview of changes from NSF. The actual licensed standard and applicable regulatory or ethics requirements govern real studies. This narrow application connects to the existing clinical-trials translation architecture, the Vocabulary Learning Hub and How English Works. It is not a replacement master hub.