Adverse Events

On a Thursday evening, a participant in your hypertension trial faints in a supermarket queue. The emergency department keeps her overnight; you learn of it at 08:10 on Friday. Nothing in her history explains it, the trial drug plausibly could, and an overnight admission is an overnight admission. By Friday afternoon your report is with the sponsor: serious, possibly related.

In trial vocabulary, her faint is an adverse event: any unfavourable medical occurrence in a participant, drug-related or not. The two labels you attached — serious, possibly related — are exactly what this module teaches.

At your site, that is the whole story — one participant, one faint, no pattern. Whether the same thing happened in Lyon or Rotterdam, your site cannot even ask. Somebody can. This module is about the system that turns scattered reports into the one thing no single investigator ever sees: a signal.

The route runs in the reporting direction. First the vocabulary: the ladder from adverse event up to the SUSAR — the suspected unexpected serious adverse reaction — climbed by two judgements, causality and expectedness. Then the machinery: your clocks, the sponsor's clocks, and the committees that read the aggregate. Last, the device track, and what happens when harm becomes a claim.

Learning Objectives

After this module, you can:

• Classify any event on the safety ladder — adverse event, adverse reaction, serious, SUSAR — and keep severity distinct from seriousness
• Judge the two gates that decide expedited reporting: your causality assessment, and expectedness against the Investigator's Brochure
• Deliver the investigator's duties: serious events to the sponsor within 24 hours of awareness, screening-phase events per protocol, follow-up through to outcome
• Trace a SUSAR through the sponsor's machinery: the 7- and 15-day clocks to Swissmedic, the two unblindings, and the committees that read the aggregate
• Run the device track — classify device events and deficiencies, route them under the right ordinance, and locate liability when harm becomes a claim

The Ladder: Four Labels, Four Consequences

Labels route events. The name you attach to Thursday's faint decides who hears about it and when: the sponsor within a day, Swissmedic within a week, or nobody at all.

The ladder's bottom rung is deliberately wide. An adverse event (AE) is any unfavourable medical occurrence in a trial participant administered the investigational product — whether or not it has anything to do with the treatment. A headache, a fall on ice, an abnormal lab value: all adverse events. Causality is not part of the definition — the net catches everything first and sorts afterwards. ICH E6(R3)

One rung up: causality enters. An adverse drug reaction (ADR) is an adverse event where a causal relationship to the medicinal product is a reasonable possibility. Every reaction is an event; only suspected ones are reactions. ICH E6(R3) That judgement is the next section's subject.

One distinction decides reporting obligations and is misjudged constantly.

Severity is intensity: mild, moderate, severe. Seriousness is consequence — what the event did or threatened to the participant. A severe headache is intense but usually harmless. A mild stroke is a stroke.

Notice what the criteria measure: outcomes and consequences, never causes. The fall on the icy pavement becomes an SAE the moment it puts the participant in a hospital bed. Whether the trial drug had anything to do with the ice is a separate question — one the definition deliberately does not ask.

The top rung stacks three judgements. A SUSAR — suspected unexpected serious adverse reaction — is an adverse reaction that is all three at once:

• Suspected — there is a reasonable possibility the drug caused it
• Unexpected — its nature or severity is not consistent with the product information; the expectedness section below owns the test
• Serious — the criteria above

Drop any one of the three and it is not a SUSAR. This is the label that starts the module's fastest clocks, because an unexpected serious reaction is precisely the news the current safety picture failed to predict. ICH E6(R3)

Swiss law adopts the vocabulary wholesale: the KlinV takes its definitions of adverse events, serious adverse events and SUSARs from the rules of Good Clinical Practice it declares binding. KlinV

What Swiss law scales to risk is the documentation burden. In a Category C trial, the highest-risk category as Module 6 taught, the investigator documents every adverse event in a standardised manner; the protocol may waive documentation only for events identified as not critical to the safety evaluation, in justified exceptional cases. In Category B, standardised documentation is owed for events the protocol flags as safety-critical, or where an authority requests it; in Category A there is no AE-documentation obligation at all. KlinV

The First Gate: Your Causality Call

No instrument measures relatedness. Between an event and its route through the system stands one clinical judgement, and it is yours: is there a reasonable possibility the investigational product caused this?

That phrase — reasonable possibility — is the working threshold. It is how the guideline defines "suspected" in the SUSAR definition, and it is the line that decides whether an event can travel the expedited path at all. It does not demand certainty, or even probability. It asks whether a causal relationship can honestly be entertained on the evidence in front of you. ICH E6(R3)

E6(R3) makes the assessment part of the report itself: when you report a serious adverse event to the sponsor, you include your assessment of causality. Not a placeholder to be filled in when the picture clears — your judgement, at reporting time, on what you know. ICH E6(R3)

The reasoning is ordinary clinical craft, and it belongs in the source record, not just in your head:

• Timing — did onset follow exposure at a plausible interval?
• Alternatives — does the underlying disease, a co-medication, or the participant's history explain it at least as well?
• Course — what happened when the product was paused or stopped?
• Prior knowledge — is there a known mechanism, or experience with the class?

Class experience can feed your suspicion — but, as the next section shows, it plays no role in whether the reaction counts as expected.

Many protocols ask you to grade the relationship on a scale — not related, unlikely, possibly, probably, definitely. Use whatever scale your protocol specifies. But do not let the grades obscure the regulatory question underneath, which is binary: reasonable possibility, or not. An honest "possibly" and a confident "definitely" travel the same road.

The asymmetry is the reason to lean toward reporting when genuinely torn. An over-cautious assessment is recoverable: the sponsor's pharmacovigilance physicians can downgrade it the moment better information arrives. An event never reported is not recoverable — it simply does not exist in the safety database, and no amount of review can weigh evidence that was never filed.

The Second Gate: The Brochure and Its List

Somewhere in your investigator site file sits the document that decides, in advance, which reports will race and which will walk: the Investigator's Brochure (IB). This module is its home — every earlier mention finally gets its full introduction here.

The IB is the sponsor's compilation of the clinical and nonclinical data on the investigational product relevant to studying it in humans. Its job is working knowledge: giving you the rationale behind the protocol's dose, dosing interval, administration and safety-monitoring procedures, and the basis for managing participants clinically. The sponsor develops it, keeps it current as significant new information arrives, and reviews it at least once a year. For an authorised product studied within its approved use, current scientific information — a summary of product characteristics, for instance — may serve instead. ICH E6(R3)

Buried in the IB is the list this module cares about most: the Reference Safety Information (RSI). It is a cumulative list of the adverse reactions that are expected for the product, with information on their frequency and nature. Expected does not mean common or acceptable. It means: already described, at this character and this severity, in the product's current safety picture. ICH E6(R3)

Whose call is it? For regulatory reporting, the formal expectedness assessment belongs to the sponsor, made against the product information — the RSI in the IB, or its authorised-product equivalent. ICH E6(R3)

Your part is upstream and just as load-bearing: work from the current IB version, and know what the RSI does and does not contain for the product you are dosing. Above all, never talk yourself out of a report because "everyone knows the class does this". Class knowledge is not in the RSI; expectedness is assessed against this product's documented picture, not its relatives'.

New safety information important enough to matter may reach you before it appears in a revised IB — the sponsor may communicate it to investigators, ethics committees and authorities directly. File it, read it, and treat it as part of the current picture. ICH E6(R3)

Your Clock: What the Investigator Owes, and When

Awareness starts the clock — not certainty, not paperwork, not the next team meeting. Every deadline in this section is measured from the moment you or your delegated team reasonably become aware of the event.

The core duty is fast and simple. Serious adverse events are reported to the sponsor immediately; Swiss law fixes the outer edge at 24 hours from the event becoming known. KlinV The only exempt events are those the protocol says need not be reported — a carve-out the sponsor typically reserves, with regulatory agreement, for endpoint events the trial exists to count. ICH E6(R3)

A suspected unexpected serious adverse reaction runs on the same 24-hour clock to the sponsor — with no protocol carve-out attached. KlinV Your causality assessment travels with either report. ICH E6(R3)

Report what you have, then keep reporting. An initial report on partial information is the system working as designed; subsequent information follows as follow-up reports. For deaths, you supply the sponsor, the ethics committee and, where applicable, the authority with any additional information they request — autopsy findings, terminal medical reports — when it becomes available. A case stays open until its outcome is documented: recovered, recovered with sequelae, ongoing at trial end, or death. ICH E6(R3)

The window before first dose is covered too. Unfavourable medical events occurring during screening — after consent, before the product — are considered and reported to the sponsor if the protocol requires it. A screening-phase admission cannot be a drug reaction; nobody has been dosed. But the reporting machinery for serious events does not run on causality, and the sponsor's safety review, as the next section shows, reads the pre-dose window as part of the trial's picture. ICH E6(R3)

Non-serious adverse events run on the protocol's schedule — typically collected at each visit and reported with the data flow, at whatever standard of documentation the trial's category demands. And everything in this section is delegable in execution, never in ownership: qualified site staff may run safety reporting, while responsibility for participant safety and reporting compliance stays with the investigator. ICH E6(R3) Your SUSAR duty to the ethics committee — 7 or 15 days, under KlinV Art. 41 — belongs to Module 9's reporting calendar. KlinV

What does a defensible AE record hold?

• Onset and course
• A description in medical terms, not just the participant's words
• Severity
• Your causality assessment, with its reasoning
• Action taken, and outcome
• Where the event was serious: the date you became aware, and the date you reported

The last pair is not bureaucracy: it is the documented proof that the 24-hour duty was met, KlinV written where Module 12's standard keeps it credible. ICH E6(R3)

The Sponsor's Machinery: From One Report to Swissmedic

Your Friday report lands in a database you will never log into. What happens there is the sponsor's half of the bargain — and since sponsor duties fall to the sponsor-investigator in an investigator-initiated trial, it may be closer to home than you think.

The sponsor owes the trial an ongoing safety evaluation of the investigational product, with the IB as its reference point. Incoming reports are aggregated and reviewed in a timely manner — including events from the screening window — and the review has consequences: updates to the protocol, the IB, the consent materials; communication to participants, investigators, ethics committees and authorities when emerging information warrants it. Urgent safety issues move without undue delay; the safety-measure routes themselves are Module 9's ground. ICH E6(R3)

For the reports you send, the sponsor runs the second gate — expectedness against the RSI — and what passes both gates is a SUSAR, owed to regulators on the expedited standard of ICH E2A. ICH E6(R3) ICH E2A

In Switzerland the duty is concrete. For Category B and C trials, the sponsor makes the SUSAR reports to Swissmedic: within 7 days where the reaction was fatal or life-threatening, with a complete follow-up report within a further 8 days; within 15 days for everything else. For Category A trials, the ordinary pharmacovigilance duties of the Therapeutic Products Act apply instead. The duty survives the trial: a SUSAR surfacing after completion is reported all the same. KlinV Swissmedic SAE Guidance

The machinery also runs back toward you. The sponsor reports SUSARs onward to the investigators and ethics committees of its trials, with urgency matching the action required. When those notifications arrive from other sites and other countries, they are not correspondence — they are your trial's safety picture changing between IB revisions. Read them, act where they demand it, file them. ICH E6(R3)

In a blinded trial, one more thing must happen before a SUSAR report can mean anything: someone has to know what the participant was taking. This is where trials keep two unblindings, built for different emergencies — and the distinction is worth a table. ICH E6(R3) ICH E2A Swissmedic SAE Guidance

Emergency readiness is a standing duty, not an aspiration. From the first participant onward, your site must be able to unblind at any hour without hunting for permission — the code-break mechanics sit with Module 10's IMP procedures. Around both unblindings, who may see unblinded information is defined and documented in advance. Any unplanned unblinding is documented and assessed for its impact on the trial. ICH E6(R3)

The Committee That Sees What You Cannot

Twelve weeks after your Friday report, four people you will never meet sit down with a listing your site cannot generate: every serious event in the trial, worldwide, by treatment arm. Your participant's faint is on it — the ninth syncope in the trial, the seventh in the same arm. At your site it was an anecdote. Here it is a curve.

That is the Independent Data Monitoring Committee (IDMC) — you will also hear data safety monitoring board (DSMB). It is a small group of independent experts outside the trial; the sponsor decides whether a trial gets one. It meets at intervals, and it sees what nobody inside the trial sees: the accumulating safety data — and the relevant efficacy endpoints — unblinded, by treatment group. What it produces is a recommendation to the sponsor: continue the trial, modify it, or stop it. Who may see the unblinded data is defined and documented in advance. ICH E6(R3)

A recommendation is all it is. The decision — and the accountability for overriding a recommendation — stays with the sponsor. Behind it stands the standing ethical duty: when risks are found to outweigh potential benefits, assess whether to continue, modify or immediately stop the research. Declaration of Helsinki An IDMC does not replace that duty; it informs it with the one view nobody else has. ICH E6(R3)

A sibling committee reads your endpoints rather than your safety data, and it exists for a plain reason: subjective endpoints need one consistent, bias-free yes or no. An endpoint adjudication committee reviews investigator-reported endpoints against the protocol's criteria — did this admission meet the trial's definition of a heart-failure hospitalisation? — so a hundred sites' judgement calls become one standard. ICH E6(R3)

Note the deliberate inversion in who sees the code. The IDMC is unblinded on purpose: it cannot find a pattern without seeing which arm the pattern sits in. The adjudication committee is blinded on purpose, typically even in an open trial: it must judge each event without knowing the arm. ICH E6(R3)

Committees of either kind carry the same governance: relevant expertise, managed conflicts of interest, a written charter, documented decisions. ICH E6(R3)

Your interface with both is humble and load-bearing: the listings they read are built from your reports, your dates, your causality reasoning. A signal is only as sharp as the reporting underneath it.

The Device Track: Same Instinct, Different Rulebook

A drug can harm only through the body. A device can do that too — and it can also simply break. Device-trial safety law therefore watches two things: what happened to the participant, and what the hardware did or failed to do.

The participant-side ladder will look familiar: adverse events and serious adverse events carry the same logic as the drug track. The device world then adds its own causality rungs — an adverse device effect (ADE) for an event causally related to the investigational device, and a serious adverse device effect (SADE) for a serious one. The terms come from the device-GCP standard ISO 14155, KlinV which Swiss law names for trials of device-like products; for device trials proper, the ordinance takes its definitions from the EU regulations it mirrors. ClinO-MD

The hardware-side concept has no drug equivalent. A device deficiency — a shortfall in the device itself or its performance, a malfunction included — is reportable even when nobody was harmed, provided it carried the potential for serious harm.

The routing inverts the drug track, and it comes in two regimes — split by what is being tested. Hold on to one question through both: who reports, to whom, on what clock.

Device trials (ClinO-MD). The sponsor documents the trial's safety record: every serious adverse event, the adverse events the protocol flags as critical, near-miss deficiencies, and new findings on any of them. The record is produced to Swissmedic or the ethics committee on request. ClinO-MD

In Category C, the sponsor reports without delay to the ethics committee and to Swissmedic: serious adverse events whose causal link to the device, the comparator or the investigation procedure is ascertained or reasonably possible; near-miss deficiencies; new findings on either. Conformity-related trials (sub-categories C1 and C2) that also run abroad report their foreign events too. An incomplete report first is expressly allowed — "without delay" outranks completeness — and Category A keeps narrower routes through the ethics committee and ordinary device vigilance. ClinO-MD

Device-like products (KlinV). Some products test like devices but sit under the Therapeutic Products Act; their trials run under the KlinV, and the investigator stands closer to the authority-facing duty. In Category C, the investigator reports to the ethics committee, within 7 days, serious adverse events whose link to product or procedure cannot be excluded. Near-miss deficiencies go the same 7-day route whatever the category — harm or no harm. The sponsor carries the Category C reports onward to Swissmedic, foreign events included. KlinV

Your site-level duties when a device misbehaves — recognise, document, quarantine, report to the sponsor — are Module 10's ground and unchanged by any of this.

When Harm Becomes a Claim: Liability

Every record in this module has a second audience nobody enjoys thinking about: the lawyers who arrive if a participant is harmed. Swiss law makes the stakes plain: whoever carries out the research is liable for the damage participants suffer in connection with it, and that liability must be covered HRA — the architecture is Module 6's ground.

Your contribution to that architecture is the file this module taught you to build. Whether an injury was connected to the trial, what was known when, how fast the site acted — a claim runs on exactly the record your safety reporting leaves behind. A complete, dated, honestly reasoned adverse-event file protects the participant's claim and the site's answer alike; a gappy one serves nobody, including you.

Module Summary

You filed one report on a Friday, and this module followed it. Up the ladder: event, reaction, serious — and, where your suspicion meets an RSI that never predicted it, SUSAR. Through your 24-hour clock, into the sponsor's aggregation, out to Swissmedic on the 7- and 15-day clocks, unblinded one participant at a time.

And onto the listing where nine faints became a signal no single site could see. Around the drug track ran the device track, sponsor-driven and timed in two words; behind everything, the liability file.

You can now:

• Classify any event on the safety ladder — adverse event, adverse reaction, serious, SUSAR — and keep severity distinct from seriousness
• Judge the two gates that decide expedited reporting: your causality assessment, and expectedness against the Investigator's Brochure
• Deliver the investigator's duties: serious events to the sponsor within 24 hours of awareness, screening-phase events per protocol, follow-up through to outcome
• Trace a SUSAR through the sponsor's machinery: the 7- and 15-day clocks to Swissmedic, the two unblindings, and the committees that read the aggregate
• Run the device track — classify device events and deficiencies, route them under the right ordinance, and locate liability when harm becomes a claim

You will never know whether your report was the one that completed a pattern — that is the system's design, not its flaw. The trial's safety net is woven from reports filed on time by people who could not see what they were contributing to. File as if the signal depends on your report, because somewhere, eventually, it does.