What Regulatory Affairs Actually Owns

Part of Regulatory Affairs and the Jurisdiction Map: What Regulatory Affairs Owns.

About 6 minutes. Free, in your language, no account needed.

Why this matters

Ask ten people in a clinical trial what regulatory affairs does and you will get ten versions of "they handle the submissions." That is like saying a pilot handles the takeoff. It is true, it is the visible part, and it is not the part where the judgment lives.

Regulatory affairs owns a single question, asked continuously: is this trial lawful, right now, in every country where it is running? Not was it lawful when it started. Not will it be lawful when the amendment lands. Right now, in all of them, simultaneously.

That question has a different answer in Germany than in Nigeria, and a different answer this month than last month, and the person who owns it is expected to know which answers changed.

The three things you own

One: the authorization to conduct the trial in each jurisdiction. Not a general sense that approval exists. The specific instrument, its date, its scope, its conditions, and the date it expires or must be renewed. A trial running in fourteen countries has at least fourteen of these, usually more, because ethics approval and regulatory authorization are separate in most places and sites are often approved individually.

Two: the currency of that authorization. Protocols change. A substantial modification in the European Union must be authorized before implementation; the same change in another country may require notification only, or may require nothing. Knowing which is which, per country, per change, is the work. Getting it wrong means sites conduct a trial under a version their authority never saw, which is not a paperwork problem.

Three: the record that proves both. Regulatory affairs is the function an inspector goes to when they want to know why the sponsor believed it was allowed to do what it did. The answer has to be a document, with a date, from the authority, not a recollection.

What you do not own

You do not decide whether the science is sound. That is the protocol team and, in the end, the investigator and the ethics committee.

You do not decide whether an adverse event is related to the product. That is a clinical judgment, and regulatory affairs receives the assessment rather than making it. What you own is the clock that starts once the assessment exists.

You do not have authority to promise an authority anything the sponsor cannot deliver. A commitment made in a cover letter to buy a faster review becomes a condition of the authorization, and it binds the whole study.

The shape of the world you work in

Most of the countries a trial runs in apply the same technical baseline: ICH E6(R3), Good Clinical Practice. That baseline is genuinely global, and it is why a monitor trained in one country is not useless in another.

What differs is everything around it. Who authorizes. On what legal instrument. In what language. Inside what statutory clock. With what appeal if refused. Whether ethics review is centralized or site by site. Whether the import of the investigational product needs its own license from a different body entirely.

CRLN's own jurisdiction register makes this concrete. It holds 200 jurisdictions. Of those, 45 are verified against a named primary instrument, and the remaining 155 record an inference from the global baseline rather than a checked national law. That distinction is the honest center of this course: knowing what you have verified, and refusing to speak with more confidence than your sources support, is the professional skill. Everything else is filing.

In practice

A sponsor runs a Phase III trial in eleven countries. A protocol amendment adds a laboratory assessment at week twelve and removes an exclusion criterion.

The regulatory lead does not ask "can we implement the amendment?" That question has no single answer. She asks it eleven times.

In the seven European Union member states, this is a substantial modification under Regulation 536/2014: one submission through the shared portal, one coordinated assessment, and no site may implement before authorization. In the United States, the change goes to the IND as a protocol amendment and sites may proceed once the IRB has approved it, a different sequence entirely. In Japan, Canada, and the two remaining countries, three more answers.

She produces a single table: country, instrument, submission type, expected clock, earliest lawful implementation date. Eleven rows. The earliest date differs across them by nine weeks.

That table is the deliverable. Without it, the study team assumes one date, the fastest one, and five countries implement early. With it, the study team plans a staggered rollout and the inspection three years later finds eleven implementation dates that each sit after their own authorization.

What this looks like when it goes wrong

The failure is almost never a missing approval. Missing approvals are loud; somebody notices that the site cannot start.

The failure is a stale approval. The trial was authorized, correctly, two years ago. Since then the protocol has moved through four amendments, two of which were substantial in three countries. Somewhere in there, one country's submission was prepared, reviewed internally, and never sent, because the person who owned it changed roles and the handover listed it as complete.

Nobody notices, because nothing visibly breaks. Sites keep enrolling. Data keeps arriving. The discrepancy surfaces during an inspection, when an inspector asks to see the authorization for the version the site is actually running, and the most recent one on file is two versions behind.

Key takeaways