What an Inspector Looks At, and What It Reveals

Part of Consent and Vulnerable Populations: Running a Consent Process That Holds.

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

Why this matters

Consent is among the first things examined in any inspection, because it is where participant protection is most directly evidenced and because consent findings are common.

Knowing what is examined, and what conclusions are drawn from what, tells you what to check yourself.

The mechanical checks

These are quick, they are always done, and they find a surprising amount.

Is there a signed consent for every participant.

Is it dated before any trial-specific procedure. The most common consent finding in clinical research is a procedure performed before the consent date, frequently a screening blood test or an imaging study done as part of a clinic visit and later counted as a trial assessment.

Is it the version approved at that date. Cross-checked against the ethics committee approval history. A participant consented on a version approved three weeks later, or on a superseded version, both appear here.

Is it signed by the participant and by the person taking consent, and is that person on the delegation log as permitted to take consent, from a date before they did.

Where a representative, witness or interpreter was involved, is that documented and is the basis established.

Was a copy given to the participant, which is required and frequently unevidenced.

The substantive checks

These take longer and reveal more.

Re-consent history against amendment history. For each substantive amendment, which participants were re-consented and when, compared against when the change was implemented at that site. This finds the failure described in the previous lesson reliably.

Consent dates against enrollment patterns. Several consents on the same day at the same site, at times suggesting a few minutes each, raises a question about the process.

Consent in special populations. How capacity was assessed, how representatives were established, how assent was obtained, and whether any of it is documented beyond a tick.

Observation of a consent conversation, where the inspection timing allows. This is the most revealing thing available and the least commonly possible.

Asking a participant, where permitted, what they understood. Rare and decisive.

What each finding actually means

An inspector is not collecting paperwork defects. They are forming a view, and the finding is a route to it.

A procedure before consent means the site's process does not reliably separate research from care, which raises a question about everything else it does.

A superseded version means version control failed, which raises a question about every other controlled document.

Consent taken by an undelegated person means the delegation process is not working, which raises a question about who else did what.

No capacity documentation beyond a tick means the site cannot demonstrate any assessment occurred for any participant, which is a finding about the whole population rather than about one file.

Re-consent not performed means participants continued under a consent that no longer described their situation, which is the most serious consent finding available short of consent being absent.

What to check yourself

Everything above, on a schedule, before anyone else does.

Practically, a periodic self-check across the site's participants: consent dates against first trial-specific procedure dates; consent versions against the approval history; signatories against the delegation log with dates; re-consent status against the amendment history; and documentation of capacity, representation, assent and witnessing where applicable.

It is a few hours per site per quarter. It finds the mechanical failures while they are still correctable, and its greater value is that where an inspector's finding does arise, a dated record of the site checking itself changes what the finding means: a site that examines its own consent is a different organization from one that does not, and an inspector reads it that way.

In practice

A coordinator runs a quarterly consent self-check across 47 participants.

The findings: two participants whose screening blood test preceded their consent by one day, because the test was part of a routine clinic visit and was later used as a trial assessment; one consented on a version superseded two days earlier; and four with no documentation of the interpreter who was present.

None is catastrophic. All four are reportable as deviations, all four are corrected in process, and the screening finding leads to a real change: the site now confirms consent date before any sample is labeled as a trial sample.

Eighteen months later an inspection finds one similar instance. The site produces six quarterly self-checks showing it had identified the pattern, reported it, and changed its process.

The finding stands, and it is a single instance in an organization that demonstrably monitors itself, which is a different conclusion from a single instance found by an inspector in a site that had never looked.

Key takeaways