HealthPilot · Clinical evidence surveillance · Patent pending
Every day, new research redraws the map of medicine. MedVoyager explores it continuously — converting newly published studies into computable evidence, charting them to the patients in your EMR who stand to benefit, and surfacing care-plan deltas for clinician review.
The evidence-to-practice gap
1 study/ 26 seconds
Biomedical literature grows continuously. No clinician can monitor it against the specific characteristics of every patient on their panel.
~17 yearsaverage lag
Research findings historically take years to reach routine practice. Patients continue on plans that newer evidence indicates could be improved.
0 systemsclose the loop
Trial matchers match patients to trials. Retrieval tools surface papers. Guideline engines lag the literature. None of them emit an actionable, patient-specific change of plan — and none learn from what clinicians accept.
The expedition
Five co-operating engines, numbered as in our patent filing. Demand-driven and bounded — MedVoyager evaluates only the evidence-cohort pairings that matter, never the whole universe at once.
Newly published research within your surveillance scope is converted into a computable evidence object: population, intervention, comparator, outcomes, exclusions, adverse events, effect size, and an evidence-quality grade — normalized to SNOMED CT, ICD, LOINC and RxNorm.
Live EMR data becomes patient phenotype vectors — diagnoses, labs and trends, medications, comorbidities, prior treatment response — clustered into cohorts that share a condition and an outcome gap.
For each evidence-cohort pairing, an applicability score weighs population match, exclusion safety, disease state, treatment history, baseline risk, comorbidity fit and evidence strength. Unsafe matches are gated out.
Where new evidence is a clinically material deviation from guideline or current plan, MedVoyager emits a typed care-plan delta — lifestyle, dietary, diagnostic, or therapeutic — never an executable order.
A clinician's accept updates the organization's decision substrates and a coverage ledger. The next matching patient gets the recommendation immediately — the same evidence is never re-litigated for the same cohort.
Clinician review
MedVoyager is decision support, not autopilot. Each care-plan delta arrives as a structured, evidence-linked card — with the affected cohort, the source study, safety checks, and what's still missing before you act. Accept, reject, or defer; choose the scope; the system remembers.
Provenance EV-2107 · reviewed decisions propagate via coverage ledger
Expedition logs
LOG 11.11 · GASTROENTEROLOGY
A 58-year-old with confirmed diverticulosis and recurrent pain — three visits in six months, managed with generic fibre advice because guidelines only cover acute diverticulitis. MedVoyager matched her phenotype to a meta-analysis subgroup and proposed a structured SUDD pathway: baseline symptom score, cyclic rifaximin with fibre, scheduled follow-up.
Delta type: therapeutic + monitoring · subject to clinician acceptance
LOG 11.12 · RESPIRATORY
A 64-year-old with COPD still exacerbating on triple therapy, eosinophils 340–380 cells/µL. The BOREAS trial (NEJM 2023) reported ~30% fewer exacerbations with add-on dupilumab in exactly this phenotype. MedVoyager surfaced the delta the week the study was indexed — years before guidance caught up.
Delta type: diagnostic confirmation + specialist referral + therapeutic consideration
Safeguards
MedVoyager never diagnoses, prescribes, or changes treatment. It proposes; clinicians dispose. A hard contraindication withholds a recommendation for that patient even when the cohort pairing is pre-accepted.
Every recommendation carries a persistent provenance handle: the source paper, the extracted evidence object and version, the matching logic, the score components, and who accepted what, when. Ask "why was this shown?" at any time — including retrospectively.
Condition-scoped ingestion, event-driven triggers, and a coverage ledger mean MedVoyager evaluates only genuinely relevant, previously unresolved pairings — not every patient against every paper.
Organization-level learning uses de-identified acceptance signals. Protected health information never leaves your boundary to teach a global model.
Early access
We're onboarding design-partner health systems and welcoming interest from clinicians and investors. Tell us who you are — we'll reach out with a walkthrough.
You're on the early-access list. We'll be in touch shortly with a walkthrough of MedVoyager.