CLIPS

CLIPS provides context-dependent retrieval and selection of clinical trial protocol elements to support objective clinical trial protocol design.

Key Features:

• Context-Dependent Protocol Element Selection: Enables selection of protocol elements based on contextual relevance rather than simple keyword matching.
• Comprehensive Database: Built from data extracted from 184,634 clinical trials and 13,210 frame structures of clinical trial protocols.
• Semantic Information: Incorporates a variety of semantic information to enable more precise filtering during searches.
• Connected-Tree Selection Logic: Uses a connected tree format in which the selection of one element influences available subsequent element options.
• Enhanced Predictive Performance: Validation results indicate improved performance over existing databases in predicting phenotypic features.

Scientific Applications:

• Clinical Trial Protocol Design: Provides structured, context-aware selection of protocol elements to compose protocol frameworks.
• Phenotypic Feature Prediction: Supports prediction of phenotypic features to inform protocol-related decisions.

Methodology:

Constructing a detailed database from existing clinical trial protocols (184,634 trials and 13,210 frame structures), extracting semantic information, and applying context-based filtering and retrieval of protocol elements.