A study-search tool is only useful if it puts enrollable patients in front of your screeners, not if it returns long lists of trials.
November 6, 2026
7
min read
By
Julio G. Martinez-Clark, CEO, bioaccess®
Search is not the problem sites have
It is not that sites cannot find studies. Public registries, sponsor portals, contract research organization (CRO) feasibility requests, and broker emails already surface more opportunities than a team can evaluate. The problem is fit: which studies match the patients you actually see, which protocols your investigators can execute with your current staff and equipment, and which referrals turn into screened — and then randomized — participants.
I see this from both sides. I am chief executive officer (CEO) of bioaccess, a first-in-human (FIH) CRO that runs early-stage trials anywhere from Canada down to Argentina, and CEO of Amavita Research Services, a Miami cardiovascular trial site with 18,000+ active cardiology patients and zero United States Food and Drug Administration (FDA) Form 483 findings — a Form 483 being the form FDA inspectors use to list inspection observations. Running both a CRO and a site teaches you that a long list is not a pipeline. A pipeline is a short list you can staff, screen, and enroll.
A search tool solves the easy part. A matching tool solves the hard part. Your job is to force the vendor to prove which one they built.
What a referral-to-screen rate actually tells you
Define the term before a vendor redefines it. A referral is a specific person, or a study opportunity tied to a specific person or cohort, sent to your site for review. A screen is a person who completes formal screening under a protocol, after informed consent. The referral-to-screen rate is referrals who reached screening, divided by total referrals, over a defined period.
That ratio exposes what search results hide. First, real-world eligibility. A referral that never screens usually failed for a reason you need to name: wrong diagnosis or disease stage, medication exclusion, lab value out of range, distance, or visit burden — or a principal investigator (PI), the physician responsible for the trial at your site, declined the study. If most referrals die before screening, the tool is matching keywords, not your population.
Second, operational cost. A low rate by source shows where coordinators spend prescreening hours on charts that were never viable. Third, honesty. Screening is not enrollment, so also ask for screen-to-randomization rates — but screening is where a vendor's claim first meets your clinic's reality.
Ask for the rate in writing, with numerator, denominator, and time window. A percentage without those is marketing.
Why "by source" is the whole point
Most vendors who report outcomes report them in aggregate: all referrals, all sites, all indications, blended into one reassuring number. That is where weak channels hide.
Demand the rate broken out by source — the channel that produced the referral: a registry, a sponsor direct referral, a recruitment campaign, a physician-referral network, a CRO feasibility channel, a database or electronic health record (EHR) query, an EHR being the digital record system your clinic uses for patient care, or the vendor's own pool. Require the same breakout by therapeutic area and for sites similar to yours.
Channels behave differently. One source may send fewer referrals but a far higher share who screen, while another sends volume that produces almost nothing. You cannot negotiate price or staffing without knowing which is which. A vendor who only tracks clicks or "studies matched" cannot produce this table — that tells you what their product optimizes for.
Set your own threshold before the demo. Decide what rate, by source, would justify the fee given your coordinator cost per prescreen hour. There is no universal number: a rare-disease site and a high-volume cardiology site should set different bars.
Enrollable fit beats theoretical fit
Search tools blur the difference between a study that fits on paper and one your site can genuinely enroll. Paper fit means the indication and age range match. Enrollable fit asks more: Do you have patients in the required disease stage, on or off the background therapies the protocol allows? Can you complete the required imaging, procedures, or overnight stays? Does your clinical trial management system (CTMS), the software sites use to manage trial operations, visits, and finances, and your staffing support the visit schedule? Will your institutional review board (IRB), the committee that reviews and approves research involving human participants, let you activate before enrollment closes?
A serious vendor should explain how their matching handles exclusion criteria, not just inclusion criteria. Ask them to walk through one study of yours that enrolled well and one that failed, and show how their tool would have scored referrals for each. If it would have sent the same volume for both, it is counting, not learning.
Questions to ask in the vendor demo
Bring this list. Do not let the demo stay on the vendor's script.
- How do you define a referral, a screen, and a randomized participant, in writing?
- What is your referral-to-screen rate by source for sites in my therapeutic area and of my size over the last 12 months? Show numerator and denominator.
- What is your screen-to-randomization rate by the same sources?
- How many referred studies did sites decline after feasibility, and for what recorded reasons?
- How do you verify a referred patient meets key exclusion criteria before sending them?
- How do you de-duplicate referrals across channels, so we do not pay twice for the same person or study?
- What data do you feed back from our CTMS or screening log to improve matching — and what happens to it if we cancel?
- Can we run a pilot limited to one or two sources, with success criteria we set in advance?
- What counts as a billable referral, and what happens when a referral is ineligible on first chart review?
- Can you provide two site references in our specialty whom we can call without you on the line?
Notice what is not on this list: database size, sponsor count, or search speed. Those are inputs. You are buying an outcome.
Data to demand before you pay
Before any annual contract, ask for a one-page data pack: referral-to-screen and screen-to-randomization rates by source for comparable sites; median time from referral to first screening visit; the top five reasons referrals failed to screen, by source; and the share that were duplicates or already known to the site. Ask how the data was collected — self-reported, pulled from a system integration, or estimated — because estimates should be priced as estimates.
If the vendor calls this data proprietary, propose aggregated, de-identified results for your segment, shared under a nondisclosure agreement (NDA), a legal contract that keeps shared information confidential. A vendor who will not share outcomes under an NDA is asking you to take all the purchase risk.
Put the measurement in the contract: which rates will be reported, by source, how often, and what remedy — credit, pause, or termination — applies if performance falls below the agreed pilot threshold.
Red flags that should end the conversation
Slow down sharply if you hear: results reported only as search volume or "matches delivered"; refusal to break outcomes out by source; a blended success rate with no numerator or denominator; pressure to sign an annual deal before a pilot; billable-referral definitions that count a name you already had; no clear answer on data ownership after cancellation; or references limited to logos, with no site you can call.
The subtler red flag is a vendor who treats your site as a customer for studies rather than a clinic responsible for patients. Every bad referral costs a coordinator an hour and sometimes a patient a false hope. Respect for that cost shows up in the data a vendor tracks.
You do not need a longer list of studies. You need fewer, better referrals that screen and enroll. Vet for that, by source, and the right vendor will be easy to spot — they will already be measuring the same thing you are.
Julio G. Martinez-Clark is chief executive officer (CEO) of bioaccess\u00AE, a first-in-human contract research organization running early-phase trials anywhere from Canada down to Argentina, and of Amavita Research Services, a Miami cardiovascular research site he has operated for seven years with zero United States Food and Drug Administration (FDA) Form 483 inspection findings. He founded Global Trial Accelerators\u2122, a podcast for medtech and biotech founders navigating first-in-human studies.