What FDA and EMA inspectors actually flag most often

Comparing inspection findings across regulators is harder than it sounds, since different agencies categorise deficiencies differently and inspect under different legal frameworks. A descriptive analysis addressing exactly this, examining 26 clinical investigator sites and 23 sponsors or contract research organisations inspected by both the FDA and EMA across 31 marketing applications between 2009 and 2015, found something worth taking seriously: despite the structural differences, the two agencies converged on remarkably similar conclusions about where the real problems sit.

Different labels, same underlying concern

At clinical investigator sites, the FDA's inspections centred primarily on protocol compliance deficiencies, essentially whether the site actually followed the approved study protocol. The EMA's inspections of the same sites centred primarily on documentation deficiencies, including trial master file completeness and quality. On the surface, these look like different concerns: one about execution, one about paperwork.

At the sponsor and CRO level, a similar pattern appeared with the labels reversed in emphasis. The FDA's inspections focused on trial management issues. The EMA's focused again on documentation.

It would be easy to read this as the two agencies simply prioritising different things, protocol adherence for one, paperwork for the other, and conclude their findings aren't really comparable. The concordance data tells a more interesting story.

The agreement rate is the real finding

When the researchers examined how often the two agencies' findings actually agreed, meaning both flagged the same underlying issue at the same site or sponsor, regardless of which category label each agency used, concordance was around 90% for protocol compliance findings at investigator sites and trial management findings at sponsors and CROs. Documentation-related concordance across both inspection types sat somewhat lower, around 70%, but still substantial.

That level of agreement, from two regulators operating under different legal frameworks, with different inspection protocols and different categorisation systems, is a strong signal that the underlying problems being detected are real and consistent, not an artefact of one agency's particular inspection style or documentation preferences. If the FDA and EMA were essentially finding different things and just describing them with overlapping language, concordance this high wouldn't appear. Two independent measurement systems converging on the same answer is one of the stronger forms of evidence available in a field where randomised comparison isn't really possible.

Why protocol compliance and documentation keep recurring

The specific categories that recur, protocol compliance at sites, documentation across both levels, trial management at the sponsor and CRO level, point toward a consistent underlying picture of where trials actually go wrong. It's rarely a single catastrophic failure. It's closer to a mismatch between what a study was designed to do and what actually happened in practice, combined with an inability to demonstrate, after the fact, exactly what did happen and why.

Protocol compliance deficiencies at the site level suggest execution drifting from design, whether through misunderstanding, workload pressure, or ambiguity in how the protocol was actually written. Documentation deficiencies suggest that even where execution was reasonable, the record of it wasn't good enough to demonstrate that clearly to an inspector after the fact. Trial management deficiencies at the sponsor and CRO level point at oversight gaps, the sponsor's ability to actually know what's happening across its sites in close to real time, rather than discovering problems only when an inspector does.

What this means for where to focus, regardless of which regulator is inspecting

Given the strength of the concordance finding, a study team doesn't need to treat FDA-style and EMA-style compliance as two separate exercises requiring two separate strategies. The overlap suggests a shared foundation would address most of what either agency is likely to flag:

• Protocol compliance needs active, ongoing verification, not just training at study start. If drift from protocol is the leading site-level finding for one major regulator and closely correlated with the leading finding for the other, it's the single highest-value thing to monitor continuously rather than assume was covered by initial site training.
• Documentation quality is a shared vulnerability across both regulators and both organisational levels. A trial master file that's complete, current, and genuinely usable by someone other than the person who built it addresses a finding category that both agencies independently prioritise.
• Sponsor-level oversight, the ability to know what's happening at sites without waiting for a scheduled visit, addresses the trial management category directly. This is precisely the case for continuous data visibility over purely periodic, scheduled monitoring, a theme regulatory findings return to repeatedly across different specific studies and inspection types.

The broader takeaway is genuinely useful for prioritisation: with two major regulators independently converging on the same handful of problem areas at a concordance rate this high, a study team doesn't need to guess where to focus limited quality assurance effort. Protocol compliance, documentation, and active oversight aren't just three items on a long compliance checklist. They're the specific areas where the evidence, from two independent inspection systems, says problems are most likely to actually be found.