The conventional site qualification questionnaire asks the site about itself. How many investigators? What equipment? What IRB? What CTMS? The answers come back, the site qualifies on paper, and three months later the site has not enrolled — and no one quite understands why, because the credentials were excellent.
The disconnect is that none of those questions investigate the underlying physician-patient relationship at the practice. That relationship is what determines whether the patient walks in for the visit, trusts the recommendation to consent, sits through the additional draw, and comes back for follow-up. Every act in that sequence is a test of the relationship. A site whose physician-patient relationships are deep can sustain those tests easily; a site whose relationships are shallow cannot, regardless of how good the equipment is or how prompt the IRB.
Good site qualification, then, is an exercise in unraveling the relationship — investigating, in operational detail, whether it is the kind of relationship that will hold up against the friction the study is going to introduce. The framework below is how we do it. The five principles operate at different layers of the relationship, from the high-level visit data down to the parking-lot fee that the patient is going to pay on the way to the second draw.
Ground every claim in real patient flow.
The relationship is observable in who actually walks through the door. Eligible-patient counts on paper are a fiction without the visit data behind them.
The first principle. A site that says "we see 800 patients with this diagnosis per year" is making a claim that can be true and operationally irrelevant at the same time. The 800 patients may be on the practice's panel. They may have a diagnosis code in the EHR. The question for the study is how many of them walk into the practice, in the relevant visit type, during the recruitment window. That number — invariably smaller than the panel count — is the empirical measure of whether the physician-patient relationship is producing the visits the study needs.
The corrective is to qualify against actual visit flow. How many patients with this diagnosis walked through the practice in the last 90 days, by visit type, by day of week, by clinician? Claims data and EHR analytics answer this question. The site's self-report does not, because the site is reporting on the relationship as it understands it, and the gap between the understanding and the actuals is exactly where qualification has to look.
On a recent Hep B seromarker study, two qualified sites had similar nominal eligible populations on paper. Real visit-flow analysis revealed they were running at materially different rates: one site's claimed weekly eligibility was four times what the other site delivered, despite both qualifying on credentials. The differentiator was the depth of the physician-patient relationship at each practice — one site's patients came in routinely on cadence; the other site's patients required active outreach to schedule. The flow data exposed the relationship gap before the study did.
Every qualification packet should include a 90-day visit-flow analysis: real patients with the relevant diagnosis or indication walking into the practice during the period that matches your recruitment window. If the analysis is not possible — claims access, EHR access — that is a qualification gap, not a research opportunity.
The qualification funnel is a marketing funnel.
Awareness, walk-in, eligibility, consent, completion. Each step has a conversion rate, and each rate is a different measurement of the relationship.
The second principle. Marketing teams have been building funnels for decades, and the IVD site qualification problem has the same shape. A patient who is going to enroll passes through a sequence: aware the study exists, walks into the practice during the recruitment window, screens eligible, consents, completes the visit and follow-up. Each stage has a conversion rate, and each rate is a different stress test of the physician-patient relationship.
A site whose patient awareness is high but whose consent rate is low has a relationship that brings patients through the door but does not produce the trust required for them to consent to study participation. A site whose awareness is low but whose consent rate is high has the inverse — deep trust where the patients know about the study, but shallow reach across the panel. The two failure modes look the same in aggregate eligibility numbers and look very different through the funnel lens. Qualifying against the funnel surfaces these differences before contract.
For every candidate site, estimate the five conversion rates: awareness (% of relevant patients who know the study exists), walk-in (% of aware patients who arrive in the recruitment window), eligibility (% of walk-ins who pass screening), consent (% of eligible patients who sign), completion (% of consented patients who finish the visit and any follow-up). Sites with low conversions at any single stage are signaling something specific about the relationship at that stage, regardless of their totals.
Audit the physical distance from check-in to consent.
Every step a patient has to walk between arriving at the practice and reaching the consent room is a step at which the relationship is being tested.
The third principle is the most physical. The journey a patient takes between checking in at the front desk and sitting down with the study coordinator is rarely audited as part of site qualification. It should be. Patients who have to walk a long hallway, take an elevator to a different floor, or wait in a separate area for the study are patients with multiple opportunities to leave — and the relationship's strength is what determines whether they take the opportunities or not. A patient with a deep relationship to their physician will tolerate more friction; a patient with a shallow one will leave at the first inconvenience.
The audit is straightforward. Walk the path the patient will walk, on a typical clinic day, with a watch. How many minutes from check-in to the consent room? How many waiting transitions? How many points where a patient could reasonably decide they have other things to do? The shorter and more direct the path, the less of the relationship's strength has to be spent absorbing the friction. The right configuration is the consent room adjacent to the check-in area; the wrong configuration is the consent room three floors away in a separate building.
Site qualification visits should always include a physical walk-through of the patient path from check-in to consent. Time it. Note the transitions. Sites where the path is more than a few minutes or has more than one waiting transition should be flagged as relationship-stress risks regardless of other strengths.
Does the patient come in anyway, or have to be summoned?
Studies that overlay an existing visit type ride on a relationship that is already producing the visit. Summoned visits ask the relationship to do work it may not be sized for.
The fourth principle restates the relationship insight in the language of recruitment design. A site qualifies for a study at the rate at which the visit the study requires is going to happen anyway. If the protocol can be embedded in the patient's existing well-women visit, their pediatric well-child visit, their endocrinology follow-up — the visit is already on the calendar because the relationship is already producing it, and the study runs. If the protocol requires the patient to come in for a study-only visit untethered from clinical care, the relationship has to be strong enough to summon them, and summoning is the failure mode for shallower relationships.
The qualification implication is sharp. Two sites with identical credentials and identical eligible populations can have wildly different enrollment rates depending on whether the protocol matches their existing visit pattern — because the underlying relationships have different absorption capacity for the friction the study introduces. A primary-care practice that already runs annual physicals on a steady cadence is qualified for a protocol that fits that visit. The same practice is poorly qualified for a protocol that requires patients to make a special trip. The qualification has to capture this fit.
For each candidate site, name the visit type the protocol will overlay on. If the answer is "a study-only visit," the site qualifies only at the rate the relationship can sustain summoned visits, which is materially lower than the rate for embedded ones. Plan accordingly, or design the protocol around an existing visit type.
The details that erode the relationship are first-class qualification criteria.
Parking fees, drop-off zones, elevator access, phlebotomist day-of-week availability. Each one is a small tax on the relationship, and the taxes accumulate until the relationship cannot pay them.
The fifth principle is the one most often skipped. The site qualification questionnaire never asks about parking, never asks about drop-off zones, never asks whether the practice's lot recently raised fees from $5 to $50, never asks which days of the week the phlebotomist is on site. These details are not, individually, dealbreakers. They are taxes on the physician-patient relationship that accumulate over the course of the study, and the relationship can only absorb so many before it stops producing visits. A patient who paid $50 for parking last visit will think twice about coming back for the second draw. A patient whose ride dropped them at the wrong door and got a parking ticket will not.
The corrective is to ask about these details directly during qualification. Some of them — phlebotomist day-of-week, draw-room location, parking lot fees — are simple to verify and have meaningful relationship implications. Others — the demeanor of the front-desk staff, the ease of finding the consent room from the elevator — require a site visit to assess. None of them are in the standard qualification questionnaire. All of them belong there, because all of them are taxes on the relationship the study is asking the site to sustain.
The deeper move this enables is the recognition that the CRO's job is to know the ground-level reality of the site in a way the sponsor cannot. The CRO that knows the parking lot fee just went up is the CRO that can tell the sponsor why follow-up enrollment dipped at the lead site this month. That ground-level knowledge is the qualification work that matters, because it surfaces the small forces that erode the relationship before they accumulate into a deficit no recruitment plan can recover.
The site qualification packet should include a section on operational details: parking and drop-off, distance to consent, days of week the relevant clinical staff (phlebotomist, lab tech, study coordinator) are on site, hours of practice, and any recent changes to operations that might affect patient return. Sites where these details are unfavorable should be flagged regardless of credential strength.
What this framework rules out.
The five principles describe how to qualify sites against the underlying physician-patient relationship that produces every visit. They also rule out a few conventions worth naming.
They rule out credential-based qualification as the primary instrument. Credentials are necessary but not sufficient; they tell you the site can run the study, not that the relationship will hold up.
They rule out aggregate eligible-patient counts as predictive of enrollment. The number that matters is the number who walk in, in the relevant visit type, during the relevant window — and that number is a function of the relationship, not the panel size.
They rule out treating qualification as a paperwork exercise. The relationship is the unit. Investigating it requires fieldwork — walking the path, asking about parking, watching the front desk, listening to the coordinator describe what their day actually looks like.
The framework is not closed. When the study outcome matters, you call RDI. Sites qualify on the relationship, not on the paperwork. Unravel the relationship and you'll qualify the right ones.