TrialBench

TrialBench provides a comprehensive suite of 23 AI-ready datasets covering multi-modal input features and 8 crucial prediction challenges in clinical trial design.

Key Features:

• Multi-Modal Datasets: 23 meticulously curated AI-ready datasets with multi-modal input features for clinical trial prediction.
• Prediction Challenges: 8 prediction tasks encompassing trial duration, patient dropout rate/event, serious adverse event, mortality event, trial approval outcome, trial failure reason, drug dose, and design of eligibility criteria.
• Validation Methods: Basic validation methods provided for each task to ensure dataset usability and reliability.

Scientific Applications:

• Trial Duration Prediction: Predicting the duration of clinical trials to inform planning and resource allocation.
• Patient Dropout Prediction: Predicting patient dropout rate and dropout events to optimize retention strategies.
• Safety Event Prediction: Predicting serious adverse events and mortality events to assess trial safety risks.
• Trial Outcome Prediction: Predicting trial approval outcome and trial failure reason to guide go/no-go decisions.
• Dose Optimization: Predicting drug dose to support dosing regimen selection.
• Eligibility Criteria Design: Supporting the design of eligibility criteria to refine patient selection.

Methodology:

Datasets were meticulously curated from complex clinical trial data requiring medical expertise, with basic validation methods provided for each prediction task.