Why translate pharmacovigilance reports? Because a safety report can begin as an ordinary patient’s description of feeling unwell, pass through a clinician or pharmacist, enter a structured safety database and later contribute to the evaluation of a medicine used across countries. Readers searching for pharmacovigilance translation, adverse event report translation, ICSR translation, MedDRA terminology or multilingual drug safety reporting need to understand how a story remains traceable as it moves between languages and systems. A seemingly minor change to timing, medicine identity or degree of uncertainty can change how a safety reviewer interprets the report.
Pharmacovigilance is the practice of detecting, assessing, understanding and preventing adverse effects or other medicine-related problems. Translation within this work is not merely a dictionary exercise. A patient’s own expression, the date a reaction began, the prescribed product, a suspected medicine, an alternative explanation and a professional’s later assessment are distinct pieces of evidence. The translation must neither turn a suspected link into a confirmed cause nor dismiss a patient’s account because it is informal. Regulators and industry exchange individual case safety reports through structured standards such as ICH E2B(R3), while controlled terminology such as MedDRA assists in describing reported medical concepts. Those systems are powerful only when the source meaning survives entry into them.
This guide explains why translation matters in pharmacovigilance, from everyday side-effect reports to multilingual case narratives, MedDRA coding, electronic data exchange, literature cases, safety signals and regulatory communication. It provides worked examples and review procedures for learners, translators and safety teams. It does not assess whether a particular medicine caused a symptom or advise a patient to change treatment. Those tasks belong to qualified healthcare professionals and relevant safety authorities. The key question for translation is whether every reader can recover the same person, medicine, event, chronology, evidentiary source and limits of what is known.
Why the original patient narrative must survive translation
Medicine-safety monitoring needs structured categories, yet it also needs the human story that gives those categories meaning. A symptom that seems medically familiar may still be a person’s own uncertain description, and an official form can have blank fields because the information is not known. Treating a blank as a negative answer or an unfamiliar phrase as a diagnosed condition can turn an incomplete but honest report into a misleadingly complete one.
A reported event is not automatically a proven drug reaction
A report may describe something that happened after a medicine was taken, but temporal sequence alone does not establish causality. In pharmacovigilance, adverse event, suspected adverse reaction, causality assessment and confirmed diagnosis can have different meanings depending on the data source and regulatory context. A translator should preserve the source’s degree of certainty and not silently turn ‘the patient believes the tablets caused the rash’ into ‘the tablets caused the rash’. Equally, ‘not yet assessed’ is not the same as ‘unrelated’. Keep patient statements, healthcare-professional comments and regulatory assessments distinguishable. The result should allow a trained safety reviewer to consider the evidence without having it filtered through the translator’s own clinical assumptions. This distinction supports signal detection because uncertain reports can still be informative when processed consistently and interpreted cautiously.
Preserve the reporter’s voice before assigning terminology
People describe symptoms in everyday language: a pounding heart, a stomach turning, a sudden hot feeling or an inability to focus. Their words may be culturally specific, approximate or uncertain. Translating immediately into a precise diagnosis can erase the meaning of the account and introduce a condition no clinician established. Start with a faithful rendition of what the reporter said, including the speaker’s degree of confidence, then let trained safety staff select an appropriate controlled term where the workflow requires it. The MedDRA Patient-Friendly Term List is designed to help reporters select understandable expressions while retaining opportunities for free-text descriptions. This shows why a patient narrative and the database classification serve different roles. A trustworthy report stores the patient’s account and the coding decision separately so one does not masquerade as the other.
Identify the exact product without inventing equivalence
A medicine can have a brand name, generic ingredient, pharmaceutical form, strength, manufacturer, batch identifier and country-specific label. Two products may contain the same ingredient but differ in formulation, concentration or other features relevant to the case. A translator should reproduce exact brand names, product identifiers and lot numbers as recorded, with separate verified standard names only where appropriate. A phrase like ‘one pill’ is not necessarily one milligram or one prescribed dose. If a reporter cannot remember a package name, the translator must not fill it in from a common product sold in that region. Good product identity work relies on a source field, not guesswork. A safety database can standardise information later, but any normalisation should be auditable and should not conceal what the original reporter actually knew.
Strength, dose and dosage form are separate concepts
Drug-safety reports often mix medicine strength, amount taken, number of units, frequency, route and duration. A label such as ’10 mg tablet’ describes a unit’s strength; a statement that the patient took two tablets describes a possible quantity consumed; neither tells us whether the dose was prescribed, repeated or missed unless the report says so. Translation must preserve these distinctions, including units, decimal points and time intervals. An abbreviation may differ across languages or jurisdictions, so an ambiguous shorthand should be queried rather than automatically expanded. Pharmacy and regulatory reviewers need to know whether an apparent discrepancy comes from the patient’s original description, a clinical instruction or a conversion introduced in translation. This is not a reason for an unqualified editor to calculate treatment advice; it is a reason to protect the data before clinical assessment.
Time relationships are the heart of a useful case narrative
A safety reviewer may need to know when treatment began, when a symptom first appeared, when the medicine was stopped, when the report was received and what happened afterward. Translation that collapses these into a generic date can falsely change the temporal relationship between treatment and event. Record exact source time expressions, including approximate phrases such as ‘about three days later’ or ‘sometime last week’. A date inferred from an imprecise account should remain a documented inference, not become the patient’s definite claim. Time zones may matter in international studies or digital reports, and source timestamps should be retained where possible. In review, place the reported events on a timeline and distinguish each person’s recollection from dated clinical records. Chronology must remain auditable even when the summary is concise.
Seriousness is not the same as symptom intensity
Everyday speakers use ‘serious’ to mean severe or worrying, but formal pharmacovigilance criteria for seriousness relate to defined outcomes such as death, life-threatening situations, hospitalisation or other medically important conditions under applicable rules. A very painful symptom is not necessarily reportable as serious under every criterion, and a clinically important event may be serious even if its immediate discomfort seems modest. Translators should preserve the reporter’s own words and avoid making a regulatory seriousness determination through word choice. If a structured field asks whether hospitalisation occurred, the data should come from the source record or authorised professional assessment. A narrative that says ‘I was terrified’ should not be silently encoded as ‘life-threatening’ simply because the emotions sound intense. Language accuracy prevents both overstatement and accidental underreporting.
Distinguish event outcome from causal certainty
A case can describe a reaction as recovered, recovering, not recovered or unknown, yet the cause may still be uncertain. Translating ‘the rash cleared’ into ‘the medicine was cleared as safe’ merges two unrelated ideas. Similarly, a death mentioned in a report may be an event outcome without demonstrating that a medicine caused it. Preserve whether the report talks about the clinical outcome, the medicine’s suspected role or an assessor’s conclusion. In structured data, avoid turning ‘not provided’ into a negative answer. The receiving reviewer should know when there is no follow-up information and when the follow-up positively establishes that something did or did not occur. This separation is essential when case narratives are condensed into database views used by people who never speak with the original reporter.
Dechallenge and rechallenge need exact chronology
A source may mention that a medicine was stopped and symptoms changed afterward. Another report may describe an event recurring after a medicine was started again. These observations can be relevant to causality assessment, but they should not be simplified into proof of causation, nor should a translation imply that anyone intentionally repeated an exposure to test a hypothesis. Translate exactly what happened, including the source of the information, dates, concurrent treatments and uncertainty. Keep discontinuation and dose reduction separate, and do not transform a patient decision into a clinician’s instruction. Safety writing must avoid encouraging people to stop or restart medication as an experiment. The translator’s role is to preserve historical evidence; a clinician or trained assessor determines what action and interpretation are appropriate.
Controlled medical terminology is a second layer, not a substitute
MedDRA provides a structured medical terminology for regulatory activities, with concepts arranged at different levels that support entry, analysis and aggregation. It should not be used as a licence to rewrite an unclear patient sentence into a diagnosis that the source does not contain. The correct workflow retains verbatim or accurately translated narrative and makes any selected coding term traceable. Version information matters because terminology evolves, and a study or regulator may specify a required version and coding conventions. When several English descriptions appear superficially similar, the broader medical concept and case context determine which code fits. Expert coding decisions should remain attributable to the coder. A translator may suggest terminology, but should not disguise a coding judgment as though it were the original reporter’s language.
E2B(R3) data fields need relational and technical integrity
The ICH E2B(R3) standard supports electronic exchange of individual case safety reports, including sender and case identification, patient characteristics, medicinal products, reported events, tests, outcomes and related information. Translation of accompanying narratives must not break the relationships among those fields. A patient may mention more than one medicine, several symptoms and a follow-up report. If the language version changes which medicine is suspected or which symptom occurred on a date, a technically valid electronic message can still be clinically misleading. Preserve stable identifiers and attachments; keep data values, coded terms and explanatory text in their intended locations. A database’s successful XML validation shows conformance to some technical rules, not clinical truth. Safety review therefore needs both structural validation and semantic comparison with the source record.
Reporter roles are evidence, not mere contact details
A case may originate from a patient, caregiver, pharmacist, physician, research participant, support programme or literature publication. Those sources may observe different aspects of an event and may report hearsay or direct experience. Preserve attribution: ‘the patient’s spouse reports’ should not become ‘the patient confirms’. A literature author may describe several cases in one paper, while the safety database needs to identify the relevant individual records without inventing duplicates. Translate qualifications accurately where relevant to professional source classification, but avoid assuming a foreign title maps directly onto a licensed role in another jurisdiction. Contact and personal data also require confidentiality protections and appropriate authorised access. The reader should be able to tell who supplied which fact and whether an opinion is that of the reporter, a treating clinician or an assessor.
Follow-up information may correct rather than merely extend a report
A later interview may clarify a dosage, update an outcome or identify a previously unknown product. A translation pipeline must support amendments with version history. Do not overwrite the original narrative in a way that makes it look as if every detail was known at the first contact. In a case chronology, distinguish information first reported, later corrected and later interpreted. If a patient initially used a colloquial symptom description and a clinician subsequently made a diagnosis, retain that progression rather than making the original report look diagnostic. Data linkage and duplicates need controlled processes because two languages or reporting channels can describe the same case under different identifiers. Transparent follow-up supports better signal review and protects against the accidental multiplication of reports into invented separate events.
Literature cases require source and patient separation
A safety article may describe several patients, report aggregate outcomes and cite earlier studies in different languages. Translating its title, abstract and clinical narrative requires care to avoid turning a group statistic into an individual patient’s result. If a case report quotes another source, preserve the chain of citation rather than making the present author appear to have observed everything firsthand. Publication date is not symptom onset, and a review article is not automatically a new original case. When using translated literature for safety screening, reviewers should know which statements are case-specific, which are author interpretations and which are background references. A well-indexed translation contains the original bibliographic identifier, article language and relevant extracts, with case-level data traced to the correct passage. This minimises errors in cross-border evidence collection.
Safety signals are hypotheses for evaluation, not instant verdicts
A signal can indicate a possible new or changed association warranting further review; it is not automatically proof that a medicine caused harm. Translation must preserve terms such as potential, under review, confirmed signal, emerging concern and regulatory action according to their defined context. A safety bulletin saying ‘we are investigating reports’ should not become ‘the medicine has been found unsafe’. Nor should a clear safety instruction be weakened into an optional suggestion by a translator attempting to avoid alarm. Readers need the current assessment, what remains uncertain, what action—if any—has been formally advised, and where updates will appear. A good multilingual communication separates observed reports from evaluated causality and separate both from instructions issued by responsible authorities.
Vulnerable and sensitive reports require careful privacy
Cases may contain pregnancy information, medical histories, ages, locations, contact details and unusually distinctive sequences of events. Even when names are removed, a highly specific narrative can identify a person. Translation should follow the data controller’s authority, confidentiality obligations, minimisation rules and secure handling procedures. Redactions in the source should not be filled in using external knowledge. A patient should not be made to sound more candid or less certain than they were. If a medical term is unclear and the required information is restricted, use an authorised clarification route rather than uploading the full report to an uncontrolled online service. Privacy and interpretive accuracy reinforce one another because the patient must be able to report sensitive concerns without their words being changed or exposed unnecessarily.
Thirty practical drug-safety translation errors and worked corrections
The patient says ‘my chest fluttered’ and the translation states an arrhythmia
Fluttering is an experience the patient described, not a confirmed electrocardiographic diagnosis. A precise target-language account should reproduce the sensation and any clinical finding separately. A medical coder may later assign a term under approved conventions, but the translation should not make that judgment invisible. Keep the verbatim expression and its context, including whether the patient described timing, duration or associated symptoms.
A report changes ‘after taking’ into ‘because of taking’
The first phrase establishes temporal sequence; the second asserts causality. Preserve the difference and identify who, if anyone, assessed the suspected relationship. A safety reviewer may use chronology as one component of evaluation, but translating a preposition must not decide the medical case. In a bilingual audit, the strength of causal language is as important as the medicine’s name.
A suspected medicine becomes a confirmed culprit
A case intake form may list a product as suspected because a reporter named it. That field is not a final conclusion. If the translated summary calls the product ‘the medicine responsible’, it overstates the evidence. Keep reporting categories and their defined purpose, then allow appropriate clinical assessment. A simple note about the source of suspicion can prevent a misleading narrative.
A brand name is replaced by a plausible generic product
A patient recalls a branded liquid whose formulation varies by country. A translator changes it to a familiar generic tablet with the same active ingredient. The product identity and form have changed. Preserve the original name and uncertainty; add verified ingredient details only in a clearly attributed field. Product equivalence belongs to a professional identification process rather than linguistic convenience.
Tablet strength is confused with the amount consumed
The package says 10 mg per tablet, and the patient reports taking two tablets. A target narrative says the patient took 10 mg in total. That statement can be false even though both components looked familiar. Keep unit strength, number of tablets and recorded dose separate. Any computed total should be explicitly derived and verified, not silently substituted for the patient’s report.
A daily instruction becomes a weekly instruction
An abbreviation or ambiguous time expression changes the apparent dosing schedule. Clinical consequences can be serious, and abbreviations differ across systems. Reproduce the source form, flag uncertainty and request qualified clarification rather than guessing frequency. Drug-safety translation is not a place to normalise instructions without verification. The reported regimen should remain traceable to a source document or person.
The event begins before the medicine, but the translation reverses them
Two dates are swapped or assigned to the wrong event labels. The resulting timeline falsely suggests that treatment preceded a symptom. Build a source-aligned timeline with treatment start, symptom onset, discontinuation and follow-up. Verify labels before normalising date formats. Keep approximate dates marked as approximate rather than using guessed exact days.
The report date is mistaken for the onset date
A safety form received on 14 May may describe a rash first noticed in April. Using the receipt date as onset changes the temporal association. Keep each event’s time source and label: first symptom, healthcare contact, reporter submission, company awareness and database entry are not synonymous. A timeline audit should always check the distinction.
‘Severe headache’ becomes a ‘serious adverse event’ by word matching
In ordinary language severe describes intensity; formal seriousness uses particular regulatory criteria. Preserve the patient’s intensity description and leave the seriousness field for authorised assessment using the applicable framework. A report can be severe but not meet a seriousness criterion, or may meet an important criterion for reasons not captured by simple intensity words.
‘I was admitted’ becomes ‘I visited the hospital’
Admission and attendance at an emergency department may have different implications. The translator should preserve the actual clinical event and not expand one into the other. If the source term is ambiguous, request more information through the proper route. A hospital contact is not evidence of an inpatient stay unless the record supports it.
A patient’s fear is translated as an objective threat to life
The patient was frightened by a symptom and described fearing death, but the report contains no clinical determination of a life-threatening episode. Both facts may matter, yet they are distinct. Translate the fear as the person’s account and avoid making an unsupported regulatory classification. Patient experience deserves accuracy without forced medical categorisation.
A symptom resolved, but the narrative says the case was closed
The symptom outcome and the administrative state of the case are not identical. ‘Recovered’ concerns the event, while ‘follow-up complete’ or ‘case closed’ describes workflow. A translation should not make recovery synonymous with the end of investigation, nor closure a clinical outcome. Keep case status and patient status in their intended fields.
A side effect is labelled ‘not recovered’ when no follow-up exists
Missing information is not proof that an adverse event persisted. If an outcome is unknown, preserve unknown rather than choosing the most pessimistic clinical category. If a later source confirms recovery, record the date and provenance of that update. Data completeness and event outcome are separate questions for a reviewer.
A medicine was stopped, but the translator says the dose was reduced
Discontinuation and dose reduction are different actions. Both may be relevant to later interpretation. Translate exactly what the source says and avoid assuming the person followed medical advice. If the patient reports independently stopping treatment, preserve that attribution rather than writing it as a doctor’s instruction.
A later reaction is described as proof of rechallenge
The patient mentioned taking the medicine again at an uncertain time. A summary declares a positive rechallenge without checking recurrence, chronology or context. Keep the observed sequence and leave technical interpretation to the assessor. Do not encourage deliberate re-exposure to test suspected harm. The translator describes what was reported, not a procedure the reader should attempt.
A patient reports ‘dizziness’ but the database uses a different term
Multiple terms may appear related in everyday English while mapping to different controlled concepts. The coder should review context using the appropriate MedDRA version and selection rules. Translation must preserve the original description and not retrofit it to a preferred code. A discrepancy should prompt a coding review rather than silent rewriting of what the patient said.
A symptom and a laboratory finding merge into one diagnosis
A patient reports weakness while a blood test shows an abnormal value. A narrative that joins them into a named disease can invent a diagnosis. Preserve each observation with its origin, unit, date and professional interpretation. A clinician may connect them, but a translator should not cross that boundary independently. Structured data should retain the source of each assertion.
The reporter is changed from caregiver to physician
A case submitted by a family member may later be reviewed by a clinician. The target report states that the physician originally witnessed the event. That alters the evidence chain. Label reporter, treating professional, reviewer and follow-up contact accurately. Names may be redacted for privacy while their roles remain traceable in authorised case data.
An older case reappears as a new foreign-language case
The same incident is reported through an original national authority and a translated literature summary, each with a different identifier. Treating them as two independent cases can skew counts. Translators should retain stable bibliographic and source identifiers to support formal duplicate-detection processes. They should not themselves merge records without authorisation, but they can preserve evidence that two reports may be linked.
A multi-patient paper is reduced to one invented patient
An article describes the outcomes of several people. A translated summary combines symptoms, ages and treatments into one fictional individual. Keep case-by-case boundaries and the original publication’s structure. When details are aggregated, label them as group findings rather than patient facts. A safety team should be able to trace every extracted event to the correct section.
A decimal separator changes the clinical measurement
In some languages, a comma serves as the decimal mark. Moving a value between systems can change a concentration or test result if formatting rules are not explicit. Keep the verified numeric value and units; do not allow a text editor to guess a conversion. Review data exports independently of prose and make any locale transformation testable.
A negative test result becomes a diagnosis of absence
The source reports that one test did not detect a specified signal at a stated limit. Translation turns it into a universal claim that the patient did not have a condition. Preserve test scope, timing and reporting threshold. A negative result may be informative, but it should not be broadened beyond what the method and source say.
A literature article’s date is used as the event date
Publication, acceptance and event onset can be years apart. A translated reference list must distinguish those dates and preserve the patient’s chronology if reported. Do not fill a missing onset date with the article date merely to complete a database field. The correct response is to retain unknown or clarify through the authorised process.
A follow-up report erases the original uncertainty
The patient first reported ‘some kind of skin change’; a doctor later documented a specific diagnosis. The summary rewrites the first report as though the diagnosis was known immediately. Keep the evolution of knowledge visible through source dates and version history. This is crucial when safety reviewers assess what each source knew at a given point.
A precautionary safety notice is translated as a confirmed recall
A manufacturer says it is reviewing a possible issue and will provide updates. The target headline announces a regulatory recall that no authority has ordered. Distinguish investigation, restriction, safety communication and actual recall decisions. Use the issuing regulator’s wording and verify the relevant jurisdiction before publishing consequential instructions.
A regulatory term is copied without its jurisdiction
A product’s safety classification or reporting deadline may differ by regulator and context. Translating a rule from one authority as though it applies globally is misleading. Identify the governing document, region and effective version. Explain a concept for international readers without promising one universal filing obligation.
An age band is turned into an exact age
A de-identified report says ‘adult in their forties’, but a translation assigns the patient age 44. That adds personal specificity and can distort safety analysis. Preserve the reported age band and do not infer an exact age from unrelated contextual clues. Privacy and data integrity both benefit from restraint.
A source redaction is filled in from another record
A report conceals the patient’s location, but a translator recognises it from the source study and writes the location into the target version. This defeats a disclosure decision. Translate the redaction as a redaction and route any need for additional information through authorised access. More detail is not always more accurate in a shared safety narrative.
A machine translator simplifies medical uncertainty
An automated draft substitutes ’caused’ for ‘possibly associated with’ and drops a negative laboratory result. The fluent target sentence looks authoritative. Compare clinically consequential propositions, polarity, evidence source, product identity and time fields directly with the original. Human review should use a risk checklist, not merely proofreading for grammar.
An ICSR XML file passes validation but carries the wrong medicine
Technical validation checks file structure and permitted values, yet cannot guarantee that a narrative associates the correct suspect product with the reported event. Quality assurance must inspect both structured fields and their clinical relationships. Review the source language, translated narrative, coded terms and product identifiers together. A successful technical upload is not a substitute for semantic accuracy.
How to read, translate and quality-check safety information
Quick guide for a patient or caregiver reading a safety report
A medicine-related safety report is not a verdict. Read first for the medicine actually named, what the person said happened, when it happened, and which facts remain uncertain. Distinguish a suspected association from an established medical conclusion, and look for information about diagnosis and follow-up if it is present. If you notice that a translated description is stronger or different from what you said, contact the reporting organisation through its official correction or follow-up channel. Do not use an article or database entry to decide independently to stop, resume or change a prescribed treatment. Medical questions should be discussed with an appropriate clinician, and urgent symptoms require prompt help through local emergency services.
A translation workflow for a safety operations team
Begin with intake provenance: source language, reporter category, original text, document date, case identifiers and access restrictions. Identify the medication, patient-related event, chronology, test results and quoted statements in separate fields. Translate the narrative without normalising away uncertainty or inventing terminology. Where controlled coding is required, have a competent reviewer verify that it fits the available source evidence and the designated version of the medical dictionary. Next compare numerical values, units, dates, medication roles, outcomes and negation against the source. Keep questions, clinical decisions and corrections in an audit log. Finally, validate the electronic format and then separately perform a semantic case review. Technical conformance does not relieve the duty to preserve accurate meaning.
An independent quality checklist
Before release, ask whether every person, product and event is still linked to the correct source and whether the temporal order of events is unchanged. Check that the patient’s own words remain distinguishable from a clinician’s diagnosis and a company’s causality assessment. Inspect event intensity against formal seriousness fields rather than assuming a shared meaning. Compare outcomes with missing or unknown follow-up. Verify exact units, decimal marks, route of administration and stable record IDs. Review any MedDRA selection with its relevant rules. Confirm that privacy controls and redactions have survived. The independent reviewer should be able to reconstruct what happened and what remains uncertain without seeing a more confident story than the source provides.
Learning exercise: construct a safe timeline
Give learners a fictional source report in which a person started a medicine on Monday, experienced an unspecified symptom midweek, called a pharmacist later, and submitted the form on Friday. Ask them to produce a dated or approximate timeline with separate fields for medicine start, event onset, healthcare contact and receipt. Provide a second note stating that the patient is unsure of the exact onset date. The correct target timeline should maintain that uncertainty, not use Friday as onset or write a particular Wednesday without evidence. This exercise illustrates why exact time relations matter more than a fluent summary with all the missing dates filled in.
Learning exercise: distinguish three voices
Create a small fictional packet with a patient’s sentence, a clinician’s later comment and a safety assessor’s note. The patient reports a subjective sensation; the clinician records a clinical finding; the assessor states that a causal link has not been established. Translate each statement separately, including its speaker and certainty. Then ask another learner to summarise the packet in two sentences without changing anyone’s role. The goal is to prevent a familiar error: turning a patient’s suspicion or an assessor’s hypothesis into a single authoritative diagnosis. Mastery is visible when the translation remains accurate even after the text is shortened.
Frequently asked questions
What is pharmacovigilance translation?
It is translation of medicine-safety information while preserving patient reports, product identity, clinical terms, timelines, uncertainties and the source of each claim.
Does an adverse event report prove a medicine caused harm?
No. Reports can describe suspected events or associations that require professional assessment; the meaning of the report must not be made stronger during translation.
What is an ICSR?
An individual case safety report: a structured record used in pharmacovigilance to describe a suspected medicine-related safety case and relevant information.
What is E2B(R3)?
An ICH specification for electronic exchange of individual case safety report data, including structured fields and associated implementation guidance.
What is MedDRA?
A standardised medical terminology used in regulatory activities, including the coding and analysis of reported adverse medical concepts.
Can machine translation replace medical safety review?
No. Automation may assist drafting or triage, but clinically consequential meaning, coding, numeric fields and source attribution need qualified review.
Is a ‘severe’ symptom automatically a serious adverse event?
Not necessarily. Symptom intensity and formal seriousness criteria are different concepts and should be evaluated under the applicable reporting framework.
How should suspected side effects be reported?
Use the relevant national regulator’s or healthcare provider’s reporting process; do not rely on a translated article to make treatment decisions.
Authoritative terminology and reporting resources
The ICH E2B(R3) individual case safety report documentation describes the structure and technical exchange of case data. The US FDA E2B(R3) safety reporting resources explain implementation for FDA submissions, while the European Medicines Agency EudraVigilance overview provides European context. The MedDRA Patient-Friendly Term List explains how patient language can coexist with controlled terminology. Always check the regulator and version applicable to a real submission.
The larger lesson
Pharmacovigilance translation is valuable because it lets human reports travel across languages without losing the evidence on which safety decisions depend. It should never turn a suspicion into a verdict, a number into another dose, a blank into a negative result or a patient’s phrase into an invented diagnosis. Sound language work makes the person, product, event and uncertainty visible so qualified professionals can evaluate them. That is how translation supports better medicine-safety information while remaining clear about its own limits.
Continue with Why Translate: meaning and communication, Why translation matters in pharmaceuticals and clinical trials, and Master Art of Translation: the medical translation system.