Patient Recruitment · 2026

The Hidden Cost of a Good Enough Recruitment Percentage

Someone presents a number — 78% of target, say — and the room's response is somewhere between relief and resignation. Not great, not a crisis. Filed away. That reaction is understandable, and it's almost always the wrong one.

The hidden cost of a good enough recruitment percentage is what it stops people from asking — because a comfortable number gets filed away, not investigated, while an alarming one gets a “why” almost automatically.

What is “the hidden cost of a good enough percentage”, exactly?

It's the cost of treating a single number as the finding, rather than as the prompt for a real question. A percentage is a comparison against an expectation — useful, but it collapses an enormous amount of information into one figure: sourcing, contact speed, follow-up discipline, accessibility, pre-screening depth, handover quality, tracking and attribution, all folded into a single ratio that can land at exactly the same value for two studies with nothing else in common.

That's the hidden cost specifically: not that the percentage is wrong, but that it's a poor filter for where attention is needed. It systematically under-flags moderate, fixable problems and over-flags severe ones that were probably already visible without a score at all.

Why does a single percentage tell you less than it feels like it does?

Two studies can both report 78% of expected enrolment and be in completely different positions. One might be a rare-disease study with a genuinely small eligible population, running slightly behind an ambitious original target but with every part of its pathway functioning well. The other might be a common-condition study with a large, accessible patient pool, sitting at the same percentage because referrals are arriving and then stalling somewhere in the gap between enquiry and first contact, with nobody tracking where they're actually being lost.

The number is identical. The situations aren't. A review process built around the number alone treats them as the same problem, which means it's very likely to prescribe the wrong response to at least one of them.

How does the same score end up hiding two different situations?

Because the percentage is calculated at the very end of the pathway, after every upstream stage has already had its effect. By the time a number reaches a report, it no longer carries any information about which stage produced it — a shortfall from weak referral transfer, slow first contact, thin follow-up, or a narrow accessibility window all arrive at the review table looking exactly the same: one number, below expectation.

This is precisely the pattern data-science research calls “aggregation blindness”. A 2025 analysis of KPI failure modes found that a credit-scoring model with strong overall performance masked significantly worse outcomes for younger applicants in low-income regions — a structural problem invisible in the aggregate dashboard, and only visible once the metric was broken apart by subgroup. A recruitment percentage works the same way: broken apart by pathway stage, it can point at something specific; left whole, it can't.

What does the evidence say about aggregate metrics hiding real problems?

The clinical-trial-specific version of this pattern was documented as far back as 2012. Ken Getz's analysis for Applied Clinical Trials found that while topline industry enrolment figures looked broadly steady year over year, 11% of investigative sites enrolled zero patients — a fact the industry average alone gave no hint of. The same analysis found regional non-activation rates ranging from 7-9% in Western Europe and Asia Pacific to 20% in Latin America, a spread that a single global percentage collapses into invisibility. Getz's data also showed that 59% of activated sites met or exceeded their enrolment target, meaning a study sitting “below target” overall could still be outperforming most of its peers.

Why is “good enough” often the most dangerous read?

A percentage in a comfortable-looking range does more damage, on average, than a genuinely alarming one — because an alarming number gets investigated, and a “fine” number doesn't. A study reporting 45% of expected enrolment triggers scrutiny almost automatically; someone asks why. A study reporting 78% often doesn't, even though a 22-point gap against expectation, in a study where every referral counts, can represent a meaningfully different outcome than the number suggests on its own.

This asymmetry isn't unique to clinical recruitment. The sociologist Diane Vaughan documented the same mechanism studying the 1986 Challenger disaster, coining the term “normalization of deviance” — a long incubation period in which early warning signs are misinterpreted, ignored, or missed entirely, because nothing about the outcome looked bad enough to force a closer look. The concept has since been applied well beyond aerospace, including in clinical medicine, precisely because the underlying pattern shows up wherever a comfortable number substitutes for a real question.

What changes when the frame shifts from a score to a pathway question?

The more useful question isn't “is this number good or bad” — it's “where in the pathway is this gap actually coming from, and how significant is it given what this study is trying to do”. Instead of one number, the conversation becomes: is the shortfall concentrated in identification and sourcing, in the referral-to-contact window, or in what happens after a referral is made? Is the gap large or small relative to how difficult this specific population was always going to be to recruit?

Those questions produce a genuinely different kind of finding than a score does. The percentage still exists and is still useful as supporting context; it's just not the headline, because the headline should be the thing that would most change the outcome if it were examined. This is the same logic behind closing the loop on outcome data mid-study — a review only changes anything if it produces a specific, actionable finding, not a restated score.

What should you do differently when you see a “fine” percentage?

• Treat a fine-looking percentage as a prompt to check the pathway stage by stage, not as a reason to stop looking.
• When a number is presented, ask what's driving it before deciding whether it's acceptable — the same score can mean genuinely different things.
• Weight review time by the significance of the gap relative to the study's own difficulty, not by how alarming the raw number looks.
• Ask which specific part of the pathway a shortfall traces back to — pre-screening depth, an accessibility window, contact speed, or attribution — rather than treating “enter enrolment is behind” as a single, undifferentiated fact.

A structured recruitment diagnostic is built around this principle deliberately — the report leads with what's actually happening in the pathway, and the percentage appears only as supporting context, never as the headline finding. The same distinction runs through why studies under-enrol.

Frequently asked questions