Annual Safety Reporting and the Development Safety Update Report

Part of Regulatory Affairs and the Jurisdiction Map: Safety Reporting Obligations.

About 5 minutes. Free, in your language, no account needed.

Why this matters

Expedited reports tell an authority about a single event. Periodic reporting tells it what the accumulating evidence means, and it is the mechanism by which a regulator forms a view of a product's developing safety profile rather than a series of disconnected incidents.

It is also the reporting obligation most likely to be missed, because it has no triggering event. Nothing happens to prompt it. A date arrives, and if nobody is watching the date, nothing happens at all.

What the periodic report does

The Development Safety Update Report, described in ICH E2F, is the common format for annual safety reporting during clinical development. Its purpose is to present a comprehensive annual review and evaluation of safety information relevant to the product across its development program.

Two features distinguish it from expedited reporting.

It is cumulative and program-wide. It covers the product across studies, not one trial, and it presents the evolving picture rather than the individual events. Signals that are invisible in any single trial become visible in aggregate, which is the entire point.

It requires a conclusion. The report evaluates whether the information obtained during the period is consistent with the previous understanding of the product's safety, and whether changes are needed to the reference safety information, the protocol, or the information given to participants. A report that assembles data without reaching a conclusion has not done what it exists to do.

The date it runs from

The reporting period is anchored to the development international birth date, which is the date of first authorization to conduct a clinical trial in any country. The report is prepared annually from that anchor and submitted inside a defined window after the data lock point.

Two errors follow from getting the anchor wrong. Using the date of the first trial in your region rather than the first anywhere produces a period that is out of step with the product's actual program. Using a study-specific date produces a different report period for each study, which defeats the program-wide purpose.

For a sponsor in-licensing a product, establishing the correct anchor is genuine work, because the originating company holds the date and it will not be in the transferred dossier unless someone asks.

Recipients and the local layer

The report goes to the authorities in the jurisdictions where trials are ongoing, and frequently also to ethics committees, whose annual reporting requirements may be separate and may run on the study's own approval anniversary rather than the product's development anniversary.

That mismatch is worth stating plainly: a sponsor can owe a product-anchored annual report to the authorities and a study-anchored annual report to each ethics committee, in the same year, with different dates and different content. Treating them as one obligation causes one of them to be late.

Several jurisdictions also require a local annual report with national content requirements that the international format does not cover.

In practice

A regulatory lead takes over a product in development with trials in nine countries.

Her first action is not to read the last report. It is to build a single calendar with three classes of entry: the product-anchored annual safety report with its data lock point and submission deadline, each authority's own periodic requirements, and each ethics committee's continuing review and annual reporting anniversary.

The calendar has forty-one entries in the coming year. Eleven of them she inherited with no owner named, and two are overdue.

The deliverable is not the reports. It is the calendar, because the reports were never going to be missed for lack of ability. They were going to be missed for lack of a date with a name against it.

What this looks like when it goes wrong

A sponsor submits its annual safety report on time, and the report is a well assembled compilation of line listings and summary tables covering the period.

It contains no evaluation. It does not say whether the accumulated information remains consistent with the reference safety information, and it does not state whether any change to that document, the protocol or the participant information is warranted.

The authority's question, when it comes, is not about the data. It is about what the sponsor concluded, and the honest answer is that nobody concluded anything, because the report was produced as a compliance deliverable by a team assembling data rather than as a safety evaluation by people qualified to reach a view.

That is a more serious finding than a late submission, because it indicates that the mechanism by which the sponsor is supposed to notice a developing safety problem is not functioning.

Key takeaways