Clinresolve Server

Load a patient profile by patient ID, including demographics, conditions, allergies, active medications, and recent labs.

Collect and reconcile evidence for a drug from RxNorm, OpenFDA, DailyMed, RxNav Drug Interactions, PubMed, ClinicalTrials.gov, hospital guidelines, and clinical literature — ranked by Trust Engine scoring with temporal reasoning applied. Conflicts are detected and annotated with the preferred source and rationale.

List all patients in the system with their demographics, active medications, and conditions. Use this to browse available patients before selecting one for a medication check.

Run the full Evidence Reconciliation pipeline for a given drug and patient. Collects evidence from RxNorm, OpenFDA, DailyMed, RxNav, PubMed, ClinicalTrials.gov, HospitalGuidelines, and ClinicalEvidence; detects conflicts; applies Trust Engine ranking and Temporal Reasoning; applies patient context and hospital policies; computes reliability metrics; and produces a deterministic APPROVE / CAUTION / AVOID recommendation. The recommendation is NOT generated by an LLM — use the explain-decision prompt afterwards to get a narrative.

Retrieve the complete audit history of all clinical decisions generated in this session, merged with seeded historical records from the fixture data. Returns decisions sorted newest-first, with full reasoning trails and reliability metrics.

Search PubMed (NCBI) for research articles relevant to a drug or clinical query. Returns article titles, authors, journals, and publication dates. Results are cached for 6 hours.

Retrieve the FDA-approved Structured Product Label (SPL) from DailyMed for a drug. Extracts indications, contraindications, warnings, drug interactions, and adverse reactions. Results are cached for 12 hours.

Search ClinicalTrials.gov for clinical studies relevant to a drug. Returns study title, phase, status (COMPLETED/RECRUITING/TERMINATED), and brief summary. Completed Phase 3/4 trials are marked supportive; terminated/suspended trials are marked caution.

Re-compute structured reliability metrics for a drug+patient combination on demand. Returns 5 metrics: evidenceCompleteness, sourceAgreement, policyCompliance, patientContextCoverage, evidenceFreshness — all in range 0-1.

Collect evidence for a drug and return a head-to-head comparison of all sources, showing their trust scores, stances, and authority rankings side by side. Use this to understand which sources agree/disagree and why.

Deterministically suggest a safer alternative medication for a patient when the requested drug has a CAUTION or AVOID verdict. Looks up the alternatives table from evidence fixtures and runs a quick viability check against the patient's allergy profile.

Re-run the full decision pipeline with a hypothetically modified patient context. Useful for answering "what if" questions: what if the patient didn't have CKD? What if amiodarone was removed from their medication list? What if their eGFR was 90 instead of 35? Supports changes to: conditions (add/remove), medications (remove), labs (override eGFR/INR/creatinine).

Search the audit history for past clinical decisions involving the same drug or patients with similar conditions. Useful for longitudinal care continuity and precedent review.

Returns the real-time health status of all external evidence API providers: RxNorm, OpenFDA, DailyMed, PubMed, ClinicalTrials.gov, and RxNav Drug Interactions. Includes average latency, success/error counts, cache hit rate, and whether mock fallback is active.

Build and return a structured Evidence Graph for a drug + patient combination. The graph contains nodes (Drug, Patient, Condition, LabResult, DrugInteraction, FdaWarning, ClinicalGuideline, ClinicalTrial, ResearchPaper, HospitalPolicy) and typed edges (contraindicated_with, supported_by, conflicts_with, recommends, overrides, requires, associated_with). Each node carries metadata: source, authority, freshness, confidence, evidenceLevel.

Given a drug and patient, identify all evidence conflicts detected during reconciliation and explain which source was preferred and why (based on Trust Engine scoring). Use this to understand WHY the system sided with one piece of evidence over another.