Research theme 10
NLP-Enabled Clinical-Trial Matching
Trial extraction, eligibility interpretation, structured trial data, patient-context matching and traceable evidence.
Stated area of interest Named by KPCIRC as an area of scientific interest. Not a claim of an active funded project.
Trial eligibility criteria are written as free text, differ between protocols and change over time. Natural-language processing can help extract and structure them, so that a clinician can see which criteria may fit, which conflict and which are unknown.
Questions of interest
- How accurately can eligibility criteria be extracted and normalised?
- How should unknown or missing criteria be shown to clinicians?
- How can matches stay traceable to the protocol version and retrieval date?
Approaches
- Criteria extraction and normalisation
- Three-state matching: potentially satisfied, contradicted, unknown
- Recording registry source and retrieval date
Limitations
- A computational match is not confirmed eligibility, enrolment or benefit.
- Recruitment status and site activity must be verified with the trial team.
Projects and outputs
No verified projects or publications are listed for this theme yet.