ICH-GCP E6 Overview

It is Friday, 4:50pm. The pharmacy fridge alarm went off overnight, and the log shows the trial's entire vial stock spent six hours at 11°C against a label that says 2–8°C. The protocol points to the pharmacy manual; the manual's excursion table covers neither this temperature nor this duration. A participant is booked for dosing on Monday morning, these are the only vials in the building, and everyone who could tell you what to do has gone home.

Every trial eventually produces a moment like this: a question the documents you were trained on do not answer. This module is about the rulebook built for that moment — ICH E6(R3) itself: who wrote it, how it is organised, and how to use it when it seems, at first, to be silent. ICH is the author; Good Clinical Practice is what it wrote. Module 1 told you what GCP is, and Module 6 showed you the Swiss law that makes it binding — what you have not yet seen is the rulebook as a document: its architecture, its principles, its borders.

That matters now because you are crossing the course's seam. Part 1 gave you the why — ethics, law, consent. Part 2 is the how — documents, product, data, safety, monitoring — and it starts here, with the map every one of those modules will cite. E6(R3) opens not with rules but with principles, and it expects you to reason from them when the specifics run out.

Learning Objectives

After this module, you can:

• Explain the problem ICH was created to solve — and why GCP is one worldwide standard instead of many
• Navigate E6(R3): locate where your obligations, the sponsor's, and the definitions live
• Apply the GCP principles when a situation the rules don't explicitly cover arises
• Distinguish a trial's critical-to-quality factors from everything else, and decide what a risk-proportionate approach looks like at your site
• Decide which standard governs a given trial — E6(R3), Swiss law where it is stricter, or ISO 14155 for devices — and place the other ICH guidelines you will meet

The module makes three moves. First, the story and the map: why a single worldwide standard exists, and how E6(R3) is put together. Then the machinery: the eleven principles, and the quality-by-design thinking that drives the current revision. Last, the borders: where Swiss law takes over, where medical devices go, and which of E6's sibling guidelines you will meet later in the course.

One Standard Instead of Three

Imagine running the same trial three times — same molecule, same design, three continents — because no regulator will accept another regulator's data. Until the mid-1990s that was not a thought experiment; it was how drug development worked.

The cost was not only money and delay. Every duplicated trial exposes a fresh group of participants to an experimental treatment in order to generate knowledge that already exists. A worldwide standard for trial conduct is not regulatory tidiness — it is the thing that makes one trial enough.

That is what the International Council for Harmonisation (ICH) was created to deliver. It began — under its original name, the International Conference on Harmonisation — as a joint project of the drug regulators of three regions: the European Union, Japan, and the USA. ICH E2A Its GCP guideline, E6, states its objective in one sentence: to provide a unified standard so that clinical trial data are mutually accepted by the regulatory authorities of ICH members. ICH E6(R3)

E6 has been revised twice, and the difference between the revisions matters at your site today:

• E6, adopted in 1996, was the original: a single numbered rulebook that made "GCP" a global term.
• E6(R2), adopted in 2016, was an integrated addendum — new expectations, quality management among them, grafted onto the 1996 text rather than a rethink of it.
• E6(R3), adopted on 6 January 2025, is the rethink: a full rewrite around principles, with the operational detail moved into annexes. ICH E6(R3)

In Switzerland, E6(R3) has been the legally binding GCP text since August 2025 — the Swiss-law section near the end of this module shows the exact mechanism.

A Reader's Map of E6(R3)

Open a sponsor audit report or a monitor's follow-up letter and every finding arrives with an address — "Annex 1 §2.12" over a source-data correction that obscured the original entry, say — and the person who can read that address without looking it up has the advantage in every conversation that follows. E6 addresses are the working currency of sponsors, monitors, auditors, and your own SOPs (standard operating procedures). (For a Swissmedic inspector the anchor is often a Swiss article instead — the two-layer rule of the Swiss-law section later in this module.) Three things are worth holding onto: the architecture, the four sections of Annex 1, and what the new Annex 2 adds.

The architecture is deliberately two-level. The guideline opens with the Principles (Section II), which are written to apply across trial types and settings and to stay relevant as technology and methods change. Annex 1 and its appendices then show how the principles apply in interventional clinical trials — and the guideline states openly that further annexes may be added as trial design evolves. ICH E6(R3) This is the structural break with E6(R2), which was one long numbered rulebook. R3 separates what should never change (the principles) from what will (the operational detail).

Annex 1 has four sections, and knowing which is whose tells you where any finding, question, or obligation lives:

• §1 — IRB/IEC: the duties of the ethics committee itself (in Switzerland: the cantonal ethics committees, Module 5's territory)
• §2 — Investigator: your section — qualifications, protocol compliance, consent, participant safety, investigational product at the site, records
• §3 — Sponsor: quality management, monitoring, safety reporting; your obligations and the sponsor's interlock, and when the sponsor fails, the site usually absorbs the consequences
• §4 — Data Governance: addressed to investigator and sponsor together, covering the full life of trial data — Module 12 takes this section in depth

Three appendices follow: A (the Investigator's Brochure — Module 13's home), B (the protocol — Module 15's), and C (essential records — Module 11's). The Glossary deserves more respect than most people give it: definitions like "source records" (the original records of what happened to a participant) decide real questions, such as what a monitor may verify against what — Module 12 again.

Annex 2 is the newest part of the map. Adopted on 3 June 2026, it adds GCP considerations for trials that incorporate decentralised elements (trial activities away from the investigator's location — home visits, video calls), pragmatic elements, and/or real-world data. It is explicitly not an endorsement of any methodology, and it is read in conjunction with the Principles and Annex 1 — it extends the map, it does not replace it. ICH E6(R3) Annex 2 Where those elements touch your daily work, the operational detail lives in the module that owns it: remote and electronic consent in Module 7, product shipped to participants in Module 10, remote monitoring in Module 14.

The Eleven Principles

Go back to Friday, 4:50pm. Nothing in your protocol or your pharmacy manual answers the Monday question — but the guideline still does, and this is the section of it that does.

Principle 1 is the hierarchy that governs everything else: trials follow the ethical principles that have their origin in the Declaration of Helsinki, and the rights, safety, and well-being of participants come first. Its first sub-principle is the single most load-bearing sentence in GCP.

Now watch the principles answer the Monday question. Principle 11 requires investigational products to be manufactured to GMP (Good Manufacturing Practice, the manufacturing-quality standard) and managed in accordance with the product specifications and the trial protocol — with measures in place to ensure the product given to participants retains its quality. ICH E6(R3)

Six hours at 11°C is outside the product's specification, and no table in your manual makes it acceptable — so as of Friday evening, nobody at your site can vouch for those vials. Principle 1 settles what you do with a product nobody can vouch for: the participants' safety prevails over the interests of science and society — and over Monday's schedule, which is neither.

So the vials go into quarantine, the excursion report goes to the sponsor now, and Monday's dosing waits for the sponsor's written assessment of the product — rescheduled if that is what it takes. The accountability and quarantine mechanics are Module 10's. But notice what just happened: the rules were silent, and the answer was still there.

The full set, in one pass — each with what it looks like when it reaches your site:

• 1 — Participant primacy. The tiebreaker you have already met — and wider than it first looks: it also demands that participant selection be representative of the people the product is meant to help, that a qualified physician (or dentist) holds overall responsibility for trial-related medical care, and that participants' confidentiality is protected. A trial can violate Principle 1 in its enrolment strategy before a single participant is at risk.
• 2 — Informed consent. Participation is voluntary and rests on a consent process that ensures participants are genuinely well-informed — designed so a person can actually weigh benefit, risk, and burden, not merely receive a document. Everything Module 7 taught you hangs off this one sentence; even its emergency-deferral rule appears here in miniature.
• 3 — Independent review. An IRB/IEC reviews the trial before it starts and keeps reviewing it — and the trial must then be run in compliance with the protocol that approval covered. The approval you wait for on day one and the amendments, updates, and periodic reviews that follow (Modules 5 and 9) are the same principle at two moments in time.
• 4 — Scientific soundness. The design must rest on adequate, current knowledge — and because knowledge moves, the principle requires periodic re-checking of the science mid-trial. A trial that was sound at approval can become unsound by year two; Principle 4 is why nobody gets to say "but it was approved."
• 5 — Qualified individuals. Every trial task is designed and performed by someone qualified for that task by education, training, and experience. At a site this principle wears two familiar faces: the training record and the delegation log — proof, not assumption, of who was qualified to do what.
• 6 — Quality by design. Quality is built into the trial's scientific and operational design, not inspected in afterwards.
• 7 — Proportionality. Effort follows risk: processes are proportionate to the risks to participants and the importance of the data, avoiding unnecessary burden on participants and investigators. Principle 6's partner.
• 8 — The protocol. Clear, concise, scientifically sound, and operationally feasible — a protocol a site cannot actually execute is a defect in the protocol, which is why your feasibility review is quality work, not paperwork. Module 15 takes the document apart in full.
• 9 — Reliable results. The information a trial generates must be fit for purpose, on systems that protect data integrity, in records that stay traceable and retrievable for as long as the law requires — and the trial itself must be publicly registered, results posted. If a result cannot be trusted or found, the participants' contribution is spent for nothing (Modules 11 and 12).
• 10 — Clear roles. Responsibilities are documented; sponsors may transfer activities and investigators may delegate them, but overall responsibility stays put. The machinery — agreements, oversight, the transfer/delegation distinction — is Module 3's.
• 11 — The product. Investigational products are GMP-manufactured and managed according to their specifications and the protocol, with measures ensuring the product participants receive retains its quality — you just watched this principle decide the Monday question. The site-level mechanics are Module 10's. ICH E6(R3)

Two of the eleven turn on their exact words, not their gist — the first two checks below hinge on reading them closely. The third puts the principles to work on a question the protocol won't answer for you.

All eleven of the principles you just met bind. But two of them — 6 and 7 — are the reason GCP practice looks different under R3 than it did for twenty years before, and they get the next section to themselves.

Quality by Design: Deciding What Matters Before Anything Goes Wrong

Here is the sentence that retired a whole style of clinical research. ICH E8(R1) — the guideline E6(R3) is explicitly built on — states that checking documents and data after the fact, even when combined with audits, is not sufficient to ensure the quality of a trial. ICH E8(R1) If inspection at the end cannot produce quality, quality has to be put in at the beginning. That idea has a name, a tool, and a consequence — quality by design, critical-to-quality factors, and proportionality — and this section takes them in that order.

Quality by design (QbD) means quality is a property of the trial's design: the protocol, the procedures, the operational plans, the training — all built, prospectively, to prevent the errors that matter. ICH E8(R1) E6(R3) makes this Principle 6: quality should be built into the scientific and operational design and conduct of the trial, where quality means fitness for purpose — not maximal paperwork. ICH E6(R3)

But you cannot build everything to the same standard, so QbD forces a question most trials historically never asked out loud: what, in this trial, actually matters? The answer is the trial's critical-to-quality (CtQ) factors: the attributes of a trial that are fundamental to the protection of participants, the reliability and interpretability of the results, and the decisions made based on those results. ICH E6(R3)

The working test: if this went wrong, would the participants' protection or the answer itself be undermined? If yes — eligibility criteria that exclude an unsafe population, the integrity of randomisation and blinding, the primary endpoint measurement, safety reporting — it is CtQ. If no, it is still worth doing right, but it does not deserve the same share of the trial's attention.

Identifying CtQ factors is not a form-filling exercise. E8(R1) starts it from first questions — are the objectives clear, is the question meaningful to patients, is the design operationally feasible? — and explicitly recommends consulting patients early in the design, avoiding unnecessary complexity and unnecessary data collection, and building a culture of open dialogue about what is critical, beyond sole reliance on tools and checklists. For each factor, the risks that threaten it are weighed by probability, detectability, and impact, then accepted or mitigated. ICH E8(R1)

Proportionality is the consequence. Once you know what is critical, spending equal effort on everything stops being diligence and becomes a design flaw — attention spent on trivia is attention taken from the factors that protect people.

For your site, this changes what "being compliant" looks like. Expect the sponsor's quality system to reach you as documented procedures focused on your trial's actual risk points rather than a uniform template; expect monitoring that concentrates on critical data instead of verifying everything equally — the monitoring mechanics are Module 14's. And expect to be asked: feasibility input from sites is part of how CtQ factors get identified, which makes pushing back on an unworkable procedure part of the quality system, not friction with it.

How E6(R3) Lands in Swiss Law

E6(R3) binds no one — not by itself. A guideline issued by a harmonisation body has no legal force in any country until that country's law gives it some. Switzerland gives it force in one sentence, and that sentence repays careful reading.

KlinV Art. 5 requires clinical trials to be conducted in accordance with the rules of Good Clinical Practice "as specified in Annex 1 number 2" — and KlinV's Annex 1 names the text: for clinical trials of medicinal products and transplant products, the ICH-GCP E6(R3) Guideline in the version dated 6 January 2025. KlinV Note what that means: which version of GCP applies is not a training preference or a sponsor convention — it is a fact of Swiss law, pinned to a date. When the guideline changes, the ordinance is amended to adopt it, which is how E6(R3) became binding here in August 2025.

The same article carries two refinements, and both change what you can expect to see in real protocols.

The first: not every clinical trial in Switzerland runs on the E6 text. KlinV's Chapter 4 covers trials whose intervention is neither a therapeutic product nor a transplantation — a surgical technique, a physiotherapy regimen, a psychological intervention — and these may follow other rules recognised in the specialty, provided participant protection and data quality and security are guaranteed. KlinV GCP is the default rulebook, not the only conceivable one, and the "recognised standards" architecture behind that choice is Module 6's.

The second: even where E6 does govern, Swiss law says its measures must be adapted to the extent of the risks participants face — a low-risk trial genuinely may deviate from the rules of GCP. But this is a designed flexibility, not a loophole: any deviation must be recorded in the protocol, up front and reviewable, and participant protection and data quality and security must be guaranteed in all cases. Swiss law arrived at proportionality in its own words; Principle 7 and Art. 5(3) point the same direction, which is why a Category A trial's lighter paperwork is compliance, not corner-cutting. KlinV

Where Swiss law says more, Swiss law governs. E6(R3) is written for every jurisdiction at once, so wherever Swiss law is more specific or more restrictive, the Swiss provision is the one that binds you:

• Consent — including incapacitated adults, minors, and emergency research — is regulated in detail by the HRA and KlinV: Module 7's ground
• Risk categorisation and who authorises what (ethics committee, Swissmedic) are pure Swiss-law constructs: Modules 6 and 5, with the ongoing duties in Module 9
• Safety-reporting timelines to the ethics committee and Swissmedic are set by KlinV, not left to GCP defaults: Module 13
• Data protection adds the nDSG on top of GCP's data-integrity expectations: framework in Module 6, practice in Module 12

And this two-layer pattern — E6 made binding by local law, local law layering its own requirements on top — is not a Swiss peculiarity. It is how a guideline written for mutual acceptance is designed to be used: one standard underneath, one legal system on top, in every ICH jurisdiction.

Devices Are Different: ISO 14155

Change one word in a protocol — "tablet" to "stent" — and E6(R3) quietly lets go. Its scope is interventional clinical trials of investigational products, a term the guideline defines as synonymous with drugs, medicines, medicinal products, vaccines, and biological products. ICH E6(R3) A trial of a medical device answers to a different standard — same ethical spine, different instrument.

Internationally, that instrument is ISO 14155, the standard for good clinical practice in clinical investigations of medical devices — the device world's counterpart to E6. It exists for the same reason E6 does: a common conduct-and-data standard so that a device investigation done well once is credible everywhere.

In Switzerland the routing is explicit. Device trials fall under their own ordinance, ClinO-MD (SR 810.306) — the clinical-trials ordinance you already know, KlinV, excludes them and hands them over. ClinO-MD then requires sponsors and investigators of clinical investigations to fulfil the requirements of the EU Medical Device Regulation, Article 72 and Annex XV Chapters I and III — and, for performance studies of in-vitro diagnostics, Article 68 and Annex XIII Part A of the EU-IVDR. ClinO-MD The investigator-qualification rule makes the standards knowledge personal: device-trial investigators must demonstrate adequate knowledge of the internationally recognised requirements for the conduct of clinical trials, alongside the specialist knowledge the investigation demands. ClinO-MD

What this means at the level of this module is a single reflex: the first question about any trial is what is being tested, because that decides which rulebook you are in.

• Medicinal product — E6(R3), via KlinV
• Device — ISO 14155 and the EU-MDR requirements, via ClinO-MD

The operational consequences for device studies at your site — accountability, vigilance, deficiency reporting — are Module 10's.

The Wider ICH Family

E6 keeps its own borders deliberately tight — and says so: it is to be read in conjunction with the other ICH guidelines relevant to the design and conduct of clinical trials. ICH E6(R3) Four siblings matter for this course, and each already has a home module:

• E8(R1) — General Considerations for Clinical Studies (2021): the source of quality by design and critical-to-quality factors, which you met in the quality section of this module
• E2A — Clinical Safety Data Management (1994): the definitions of AE, SAE, and SUSAR and the expedited-reporting framework — three decades old and still operative alongside E6(R3); Module 13 is built on it
• E9 — Statistical Principles for Clinical Trials (1998) and its addendum E9(R1) on estimands: the statistical backbone of protocol design, taken up in Module 15
• E10 — Choice of Control Group (2000): placebo, active control, and the design questions between them — also Module 15's ground

The point is not to know these guidelines now. It is to stop expecting E6 to contain everything: when E6 seems thin on statistics, safety definitions, or study design, that is not a gap — it is a pointer to a sibling.

Module Summary

E6(R3) is one standard so that one good trial is enough — that was the founding bargain among three regions' regulators, and it is why the guideline's structure puts durable principles above changeable detail. You now know the map: Principles, Annex 1's four sections and three appendices, the Glossary, and the 2026 addition of Annex 2 for decentralised, pragmatic, and real-world-data elements. You know the machinery: eleven principles with participant primacy at the top, and the quality-by-design logic — critical-to-quality factors, proportionate effort — that R3 placed at GCP's centre. And you know the borders: KlinV pins the exact GCP version and Swiss law prevails where it says more; devices route to ISO 14155 and ClinO-MD; statistics, safety definitions, and design questions live with E6's siblings.

You can now:

• Explain the problem ICH was created to solve — and why GCP is one worldwide standard instead of many
• Navigate E6(R3): locate where your obligations, the sponsor's, and the definitions live
• Apply the GCP principles when a situation the rules don't explicitly cover arises
• Distinguish a trial's critical-to-quality factors from everything else, and decide what a risk-proportionate approach looks like at your site
• Decide which standard governs a given trial — E6(R3), Swiss law where it is stricter, or ISO 14155 for devices — and place the other ICH guidelines you will meet

Which returns us, one last time, to Friday at 4:50pm and a fridge full of vials nobody can vouch for. The excursion table was silent; the principles were not. That is the deepest change R3 made, and the real subject of this module: GCP no longer runs out where the rules do — and neither, now, do you.