ideasIdea
Trial matching inside the electronic record at the moment a treatment is chosen
When an oncologist opens the order screen to prescribe a new line of treatment, the record would show the trials this patient may fit, with the nearest open site and a one-click referral.
Eligibility criteria are encoded in a structured, computable form (mCODE/FHIR profiles for stage, biomarkers, prior lines, performance status) and matched against the patient's record inside the EHR order-set workflow, not in a separate portal. Alerts fire only at decision points (new line of therapy, progression documented) to avoid fatigue. Several AI matching tools exist; the missing piece is embedding at the point of decision with a referral action.
Hypothesis
Point-of-decision matching will double the proportion of eligible patients who are offered a trial in participating practices, measured by chart review, compared with practices using the same matching tool in a stand-alone portal.
Rationale
Most patients are never told about a trial; the failure is at the moment of prescribing, when the physician's attention is on the standard option. Decision-support that appears at that moment changes prescribing behaviour in other fields (antibiotic stewardship, anticoagulation).
What would test it
Cluster-randomised trial across 20 community oncology practices sharing one EHR vendor: embedded alert versus portal-only, primary outcome trial offer rate documented in notes, secondary enrolment rate.
Maturity
early clinical
Who has to act
engineering
Cost to try
Medium ($1M to $50M)
Years to first evidence
3
Bottlenecks it attacks
- Trials enrol too few, too slowly · Fewer than one in ten adults with cancer joins a trial. Trials close for lack of patients, not lack of ideas.
- Data silos · Records, scans, genomes and outcomes sit in separate systems that cannot talk. Every patient's experience is lost to the next.