tooluniverse-clinical-trial-design

Provide the indication, drug name, and target phase, and the skill runs six research paths covering patient population sizing, biomarker prevalence, comparator selection, endpoints, safety monitoring, and regulatory pathway. It outputs a 14-section report with a 0-100 feasibility score and go/no-go recommendation.

The analysis queries OpenTargets for disease data, ClinVar, gnomAD, and COSMIC for biomarker prevalence, DrugBank and FDA Orange Book for comparator drugs, ClinicalTrials.gov for precedent trials, FAERS for adverse events, and PubMed for epidemiology and regulatory literature.

The score is a weighted composite of five dimensions: patient availability at 30%, endpoint precedent at 25%, regulatory clarity at 20%, comparator feasibility at 15%, and safety monitoring at 10%. Scores of 75 or above indicate high feasibility, 50-74 moderate, and below 50 low.

Yes, it handles biomarker-selected oncology trials by cross-checking mutation prevalence across ClinVar, gnomAD, and COSMIC, and rare disease trials by evaluating orphan drug eligibility, natural history comparators, and multi-year enrollment projections.

DrugBank may not include very new or experimental compounds, and ClinicalTrials.gov searches can return limited results. This is a data availability limitation rather than a code issue, and novel compounds may score low simply due to lack of historical data.

The analysis depends on published and database evidence, so it cannot access proprietary enrollment data or unpublished FDA feedback. Pre-IND meetings and site feasibility surveys are still required to confirm projections before protocol development.