25.14 — The Rest of the World: EMA, ICH and Global Submissions

A company has a drug that works. It wants to sell it in the United States, the European Union, Japan, China, India, Brazil and Canada. Does it run seven sets of trials?

Before about 1990, in effect, yes. Each country wanted studies done its own way, in its own population, reported in its own format. The same drug was tested repeatedly, patients were exposed to trials that answered no new question, and the delay between approval in one country and another ran to years.

The fix was not a world regulator. It was an agreement between regulators and industry on the technical content — one set of scientific guidelines, one document format, one data standard. That agreement is ICH, and it is why a global submission is possible at all.

What ICH did not do is make approval a single decision. Every country still decides for itself, and the differences that remain are exactly the ones that will cause you trouble.

ICH: what was actually harmonised

The International Council for Harmonisation began in 1990 as a project between the regulators and industry associations of Europe, Japan and the United States. In 2015 it became a formal international association, and it now includes regulators from many more countries, including China, Canada, Switzerland, Brazil, South Korea and Singapore.

Its output is a set of guidelines, and their naming is a code you should be able to read on sight.

Q is quality — manufacturing, specifications, stability, analytical methods. S is safety — the nonclinical and toxicology studies. E is efficacy — clinical trial design, conduct and reporting. M is multidisciplinary — things that fit nowhere else, including the document format itself.

A handful appear so often they are worth carrying in your head.

Guideline	Subject
Q1	Stability testing
Q2	Validation of analytical methods
Q7	GMP for active ingredients
Q8–Q12	Quality by design, risk, quality systems, lifecycle
E2 series	Safety reporting and pharmacovigilance
E3	Clinical study report structure
E6	Good Clinical Practice
E9	Statistical principles for trials
M4	The Common Technical Document
M8	The electronic version, eCTD

Two things about the status of these documents matter, and people get them wrong constantly.

An ICH guideline is not law until a regulator adopts it. Each region implements it in its own way and on its own date. Chapter 25.9 gave the live example: the revised Good Clinical Practice guideline E6(R3) was adopted by ICH on 6 January 2025, became effective in Europe on 23 July 2025, and was published by the FDA as final guidance on 9 September 2025 — where "guidance" states the agency's expectations rather than binding law.

And harmonisation covers the technical content, not the administrative content. Which is exactly what the Common Technical Document expresses.

The Common Technical Document: one dossier, five modules

The CTD is the agreed structure of a marketing application, and its shape encodes the limits of harmonisation.

Module 1 is regional. Application forms, proposed labelling, prescribing information in the local format, local requirements. This module is different for every region, deliberately.

Module 2 is the summaries — the quality overall summary, the nonclinical overview and summaries, the clinical overview and summaries. These are the documents reviewers actually read first, and they are written by senior people because they carry the argument.

Module 3 is quality — everything about the substance and the product: manufacture, characterisation, controls, stability.

Module 4 is the nonclinical study reports.

Module 5 is the clinical study reports and the datasets.

Modules 2 to 5 are common across regions. Only Module 1 changes. That single design decision is what makes a global submission programme feasible.

The electronic form, eCTD, is a defined folder structure with an XML backbone describing every file and, critically, how each new submission relates to what came before — a replacement, an addition, or a deletion. A product's eCTD is therefore not a document but a growing, versioned history, sometimes running for decades, and building or migrating one is a genuine engineering problem that regulatory publishing teams own (Chapter 25.24).

Europe: one agency, several routes

The European Medicines Agency, established in 1995 and based in Amsterdam since the United Kingdom left the European Union, is not structured like the FDA and the difference matters.

The EMA does not approve anything. It coordinates a scientific assessment carried out largely by experts from the national regulators of member states, and its committee issues an opinion. The legal approval decision is then taken by the European Commission and applies across the whole Union.

The main committee for medicines is the CHMP, the Committee for Medicinal Products for Human Use. Two national agencies are appointed as rapporteur and co-rapporteur, and they lead the assessment while the whole committee votes.

Four routes to market exist and clients choose between them deliberately.

The centralised procedure gives one approval valid in all member states. It is compulsory for biotechnology products, cancer drugs, HIV, diabetes, neurodegenerative disease, orphan drugs and a few other categories, and optional for many others. The assessment runs to a defined timetable — a maximum of 210 active review days, with the clock stopped while the company answers questions, which is why elapsed time is always longer than the stated period.

The national procedure approves in one country only.

The mutual recognition procedure takes an existing national approval and extends it to other member states.

And the decentralised procedure applies to several member states at once when the product is not yet approved anywhere.

Clinical trials in Europe changed recently and completely. The Clinical Trials Regulation replaced the older directive: since 31 January 2023 all new trials are submitted through a single portal, the Clinical Trials Information System, and by 30 January 2025 all ongoing trials had to be transitioned to it. One application covers all the member states involved, with a coordinated assessment and a single decision per country. This was one of the largest regulatory system migrations the industry has been through, and it is a good example of a change that generated enormous service work with a hard deadline.

Two other European specifics worth knowing. The Qualified Person for Pharmacovigilance, the QPPV, is a named individual legally responsible for a company's entire safety system in Europe — a personal accountability with no American equivalent. And EudraVigilance is the European safety database that all cases flow into (Chapter 25.22).

The other major regulators, briefly and usefully

The United Kingdom, since leaving the European Union, is regulated by the MHRA as a standalone authority. It has built routes to lean on other regulators' decisions rather than duplicating them, which is the pattern smaller markets increasingly follow.

Japan's PMDA carries out the scientific review, with approval granted by the health ministry. Its distinguishing feature is a strong culture of formal consultation meetings during development, and a historical expectation of data in Japanese patients, which has softened as global trials routinely include Japanese sites.

China's NMPA changed dramatically after China joined ICH in 2017. It now accepts foreign clinical data far more readily than before, and approval timelines that once ran many years behind the West have narrowed sharply.

India's CDSCO, headed by the Drugs Controller General of India, operates under the New Drugs and Clinical Trials Rules of 2019, which set defined timelines for trial and approval decisions. India matters twice over: as a market, and as the source of a large share of the world's generic medicines and active ingredients, which means Indian manufacturing sites are inspected by the FDA and the EMA regularly.

Health Canada, Australia's TGA and Brazil's ANVISA follow broadly similar models and participate in various reliance arrangements.

And the World Health Organization runs prequalification, an assessment used by United Nations agencies and many lower-income countries to decide what they will purchase — which for vaccines and essential medicines is a market-defining decision in its own right.

Reliance, and where global regulation is going

Most of the world's regulators do not have the resources to do a full independent review of every product, and the honest response has been reliance: formally taking another regulator's assessment into account rather than repeating it.

Two mechanisms make this practical. PIC/S, the Pharmaceutical Inspection Co-operation Scheme, is an arrangement among more than fifty inspection authorities to use common GMP standards and share inspection outcomes, so a plant inspected by one member is trusted by others. And mutual recognition agreements do the same bilaterally, most significantly between the United States and the European Union for routine manufacturing inspections.

The practical effect for a client is real money: fewer duplicate inspections, and a faster path to smaller markets on the back of a major approval.

What actually still differs, and will bite you

Harmonisation is real for science and weak for administration, and the gaps below are where global programmes lose time.

Labelling and product information. Content, format and language differ by region, and the European product information must be produced in every official language of the Union.

Local clinical requirements. Some countries still expect local patients in the programme, either by regulation or by review practice.

Pricing and reimbursement. In most of Europe and much of Asia, approval is only the first gate: a separate national body then decides whether the health system will pay and at what price, and that assessment can take longer than the approval did. A drug can be legally available and effectively unavailable, and Chapter 25.27 shows how differently the United States handles this.

Data privacy. European rules on personal data apply to clinical and safety data, with consequences for where systems may be hosted and how transfers are handled — a constraint that lands directly on engineering decisions.

And safety reporting timetables and formats vary, even after the E2B standard harmonised the message structure. Chapter 25.22 gives the clocks.

The practical shape of a global submission programme

A large product launches into dozens of markets over several years, and the work is a pipeline rather than an event.

One core dossier is built. Regional modules are prepared. Submissions are sequenced — usually the largest markets first, because their approvals support reliance elsewhere. Questions arrive from several regulators at once, often about the same data, and answers must stay consistent across all of them, since regulators talk to each other. Commitments made in one region are tracked because they affect others. And every subsequent change — a new manufacturing site, a new indication, a label update — must be propagated to every market that has already approved the product.

That last item is the one people underestimate. A product approved in fifty countries with a change to its manufacturing process needs fifty regulatory transactions, each with its own format, language, timing and fee. The systems that manage this — regulatory information management, tracking every product, every market, every submission, every commitment and every deadline — are among the most valuable and most neglected systems in the industry, and improving them is a straightforward, provable value case for a services company.

Next: Chapter 25.15, what happens when the patents run out — generics, biosimilars, exclusivity, and why medicines get cheap all at once.