Common Feasibility Assessment Mistakes Research Sites Make

Every site runs some version of a feasibility check before agreeing to a study. Most run the same one every time, no matter what changed. The protocol gets skimmed instead of read, the coordinator who will actually run the visits sees the questionnaire after it has already gone back to the sponsor, and the enrollment number gets set by whoever in the room feels most optimistic. None of that looks like a mistake in the moment. It looks like a mistake three months later, when the study needs the numbers the questionnaire promised and the site cannot produce them.

Mistake 1: Softening the review once the relationship feels right

The sponsor is someone the site has worked with before. Or the calendar has a gap and the site badly needs a new study on it. Either way, the review gets shorter. The questionnaire gets filled out to close the deal instead of to test the fit, and the honest answer gets replaced with the one that keeps the relationship warm.

This is the single most common failure and it rarely feels like a mistake while it is happening. It feels like confidence. The problem surfaces later, when the study needs the patients the questionnaire promised and the site's actual population cannot supply them at that rate.

A feasibility review that only produces yes answers is not a review. It is a formality wearing a review's clothes.

In practice this looks like a coordinator flagging a borderline exclusion criterion and being told not to raise it because the study needs the win. Or a business contact quoting an enrollment number on a call before the clinical team has read the protocol at all. Neither person is acting in bad faith. Both are answering a relationship question instead of a capability question.

Mistake 2: Setting the enrollment number from optimism instead of history

Ask a site how many patients it can enroll in six months and the answer usually starts with how many patients with the condition walk through the door in a year. That figure skips over screen fail rate, consent refusal, and the real conversion from eligible patient to randomized patient, three numbers only a site's own history can supply.

Sites that build their projection from last year's screen-to-consent rate on a comparable protocol are running a real feasibility model. Sites that build it from population size alone are running a guess with a decimal point attached.

The number handed to a sponsor should come from what the site has actually done before, not from what sounds achievable in the abstract.

A site with 400 diagnosed patients a year is not a site that can enroll a meaningful share of 400 people. Once screen fail rate, consent refusal, and no-shows are applied, the realistic number is often a tenth of that or less. A site that has never pulled its own historical conversion rate has no way to know which end of that range it actually falls on.

Mistake 3: Ignoring the studies already running at the site

A new protocol can look like a strong fit in isolation and still be a poor fit in practice, because the site is already running two other studies with overlapping eligibility criteria. The new study does not add a fresh patient pool. It splits the one the site already has across three competing enrollment targets.

This gets missed because feasibility questionnaires ask about a site's general capability, not about what else is currently pulling from the same population. Checking current census against the new protocol's key inclusion criteria has to happen before the enrollment number goes back to the sponsor, not after.

Every open study at a site is a competitor for the next one's population until someone actually checks.

Two respiratory studies with overlapping exacerbation-history requirements, or two diabetes protocols both screening on a similar HbA1c band, do not stack. They compete for the same patients in the same months. A site that answers each questionnaire in isolation will not see the overlap until both sponsors are asking why enrollment is behind at the same time.

Mistake 4: Letting one person sign off alone

The person most likely to sign a feasibility questionnaire quickly is often the person least involved in the day-to-day work. A PI focused on the science, or a business contact focused on the relationship, can both answer in good faith without knowing whether the coordinator team has room for another active protocol.

The people who actually run the visits, screen the patients, and manage the sponsor's data requests are frequently the last to see the questionnaire, if they see it before it goes out at all.

Feasibility sign-off should require input from whoever will actually be doing the work, not only from whoever is authorized to sign.

A ten-minute check-in with the coordinator lead before the questionnaire goes back catches most of this. It is a small step that rarely happens, because it is not a required field on the form and no one owns the responsibility to ask.

Mistake 5: Treating staff bandwidth as fixed capacity

Bandwidth usually gets estimated by headcount: three coordinators on staff, so there is room for another study. That math ignores what those three coordinators are already carrying. A coordinator managing two high-visit-burden protocols has less real capacity than one managing a single light protocol, even though both count as one full-time coordinator on the org chart.

A capacity check built on active visit load and upcoming sponsor deadlines, the same kind covered in a full site readiness self-assessment, tells a different story than a staffing count does.

Capacity is a workload calculation. It is not a headcount.

The gap shows up fastest at sites running more than one active protocol per coordinator. Adding a fourth study on top of three already-demanding ones rarely fails because the coordinator lacks skill. It fails because there are only so many hours in a visit day, and the feasibility review never asked how many of them were already spoken for.

Mistake 6: Treating the assessment as a one-time gate

Protocol amendments are common after site selection. When eligibility criteria tighten or the visit schedule changes mid-study, the feasibility answer given at intake no longer describes the study the site is actually running, yet most sites never revisit it.

A site that reads the amendment the way it read the original protocol catches a shrinking eligible pool while there is still time to adjust outreach. A site that files the amendment without rereading it finds out the same thing from a missed enrollment target instead.

Feasibility is not a single checkbox at intake. It is a question worth asking again every time the protocol changes.

Why sponsors notice these mistakes even when a site does not

None of the six mistakes above show up as a line item on the feasibility questionnaire itself. They show up later, on the performance scorecard a sponsor keeps on every site after a study closes: enrollment against projection, screen fail rate, time to first patient in.

A site that overpromised at feasibility posts a wide gap between projected and actual enrollment. A site that missed a capacity conflict posts a slow start. A site that never revisited an amendment posts a shortfall that looks, from the sponsor's side, like a site that could not execute rather than a site that answered one question inaccurately six months earlier.

The feasibility questionnaire and the performance scorecard are the same conversation, measured at two different points in time.

A corrected process, in four steps

None of the fixes above require new software or a bigger team. They require the review to happen in a specific order, with the right people in the room.

What catching these mistakes actually buys a site

None of the six mistakes above is complicated to fix. They persist because they are easy to skip when a study looks good on paper and the relationship feels right. The sites that catch them consistently are the same sites sponsors keep coming back to, because their feasibility answers hold up months later instead of becoming the explanation for a shortfall.

Clinical Enroll has delivered randomization across 30+ indications, and the pattern behind those results starts the same way every time: an honest feasibility answer, checked against a site's real numbers, before anyone commits to a target.