Every recruitment plan starts with a number: how many patients we can identify. Almost none of them budget for the number that actually matters: how many of those patients will enroll.
That gap between “identified” and “enrolled” is where trial timelines quietly break. And in patient recruitment, it’s the gap that separates vendors who hand off a list from partners who deliver randomizations. It isn’t a marketing problem to be fixed at the top of the funnel – it’s a conversion problem that lives at the bottom, and most of the industry isn’t built to solve it.
Identification is easy to sell. Enrollment is hard to own.
An algorithm can surface thousands of “eligible” candidates from structured EHR data in an afternoon. That’s the easy part – and it’s the part most recruitment vendors sell. The problem is that eligibility on paper and enrollability in practice are two different things.
A patient who matches on ICD codes and lab values may still decline once they understand the visit burden. They may fail a screening question that never made it into the feasibility model. They may simply never pick up the phone. Structured data and historical assumptions don’t survive contact with real patients – and the candidate who looked perfect in the dataset walks away.
Identification produces a list. Enrollment requires a relationship. A recruitment partner earns its fee in the space between those two – by staying on the phone, working the pre-screen, and shepherding a candidate from “interested” to “consented.” Vendors that stop at the lead handoff are selling the easy half of the job and leaving the hard half to the site.
The gap is widening by design
Protocols keep getting more complex, which makes this harder every year. The median number of eligibility criteria per trial has climbed 58% in two decades – from 31 in 2001–2005 to 49 in 2016–2020 – and oncology, neurology, and infectious disease protocols routinely exceed 60 criteria.
Every criterion added at the design stage narrows the pool of patients who can actually enroll, not just the pool who look eligible on screen. So recruitment “failures” are usually symptoms of decisions made months earlier, in the protocol. A recruitment partner worth having flags that risk before the study opens – not after a campaign underdelivers and everyone goes looking for someone to blame.
That’s also why throwing more media spend at a struggling study rarely works. You can’t buy your way past a protocol that was never enrollable at the projected rate. The answer isn’t a bigger ad budget – it’s a partner who pressure-tested feasibility against real patients before a dollar went to media.
What recruitment should look like
If eligible-doesn’t-equal-enrollable is the root cause, then the vendor’s job isn’t a better ad. It’s building the whole engine for conversion:
-
Measure cost-per-randomized, not cost-per-lead. The lead is a vanity metric a vendor hides behind. Randomization is the number the sponsor is actually paying for – and the one a real partner puts its name on.
-
Pressure-test feasibility against real patients, not just data. Live pre-screening surfaces the drop-off points – burden, eligibility surprises, consent friction – before they blow up the timeline. A dataset can’t do that. A call center can.
-
Own the handoff. The gap between “interested” and “consented” is where most candidates disappear. Slow response and pre-screen fatigue quietly kill more enrollments than any ad ever created. Someone has to own that stretch, and it shouldn’t be the already-stretched site.
-
Flag protocol risk early. When feasibility looks shaky, saying so before the study opens is worth more than any tactic after it stalls.
The bottom line
The industry has gotten very good at finding patients and still struggles to enroll them – because those are different jobs, and most vendors only do the first one. The recruitment partners worth keeping are the ones built for the second: measured on randomizations, accountable through consent, and honest about feasibility before the study opens rather than after it stalls.
Identification gets you a spreadsheet. Enrollment gets you a completed trial. Choose your partner accordingly.