The protocol says a patient needs a specific diagnosis, treatment history, and evidence of progression.
That sounds searchable. Then you open the EHR.
The diagnosis is easy enough to find, but the treatment history is buried in oncology notes. Evidence of progression may be sitting in an imaging report, while another exclusion criterion shows up months earlier in a specialist’s documentation.
The information is there. It just isn’t packaged for clinical-trial recruitment.
That’s why finding candidates takes more than running a search.
A clinical-trial protocol tells a research team who can participate. It doesn’t tell a recruitment platform exactly how to find those patients inside years of medical records.
Someone still has to interpret what each criterion means in practice.
Take prior therapy. It may sound like a straightforward eligibility requirement, but answering it could mean reviewing medication history, treatment dates, physician notes, and the sequence in which therapies were given.
The harder cases are the ones where the answer exists in the EHR, just not where a conventional search knows to look.
Pathology narratives, procedure reports, specialist notes and scanned documents can all contain information that won’t appear neatly in a diagnosis code or standard database field.
In urology, the diagnosis may get a patient into the initial search, but it doesn’t present the full picture.
Protocols can hinge on details that aren’t obvious from a problem list alone: symptom severity, prior procedures, medication use, lab values, imaging findings, or specific language buried in a specialist’s note. Those qualifiers are often what determine whether a patient is worth reviewing.
That’s the gap a simple diagnosis search can’t close. It may return a large pool of patients who look relevant on paper, while the details that actually shape eligibility are scattered elsewhere in the record.
For the recruitment team, that means the real screening work starts after the names appear.
AI works through far more patient information than a recruiter could reasonably review record-by-record. However, it still needs the right clinical logic behind it.
That starts with understanding the protocol well enough to know what evidence to look for, where that evidence is likely to appear, and how different pieces of the record relate to one another.
Third Opinion brings clinical-trial expertise into that process rather than treating recruitment as a pure data-search exercise. Our approach is designed to help surface relevant evidence across the longitudinal EHR so clinical teams can focus their attention on patients who warrant deeper review.
That doesn’t eliminate human judgment and nor should it.
AI can help find the evidence, organize it and reduce repetitive chart review. Clinical professionals still decide what the information means for a particular patient and protocol.
Clinical-trial recruitment begins when the protocol is translated into a practical understanding of the patient evidence that needs to be found.
That’s a much more demanding task than searching for a diagnosis.
Sponsors and sites evaluating recruitment technology should look beyond how quickly a platform can return names. The more useful test is whether the approach understands the protocol well enough to find the right evidence across the EHR.
A long patient list isn’t the same thing as a strong recruitment pipeline.
Third Opinion helps research-focused practices enroll more patients, onboard more studies, and fulfill industry data requests while giving patients the opportunity to participate in potentially life-saving treatments. Founded by physicians, Third Opinion is built for patients and researchers.
Contact Third Opinion for more information.