Synthesizing evidence from clinical trials with dynamic interactive argument trees.

Area of Science:

• Medical Informatics
• Clinical Epidemiology

Background:

• Evidence-based medicine relies on high-quality evidence, primarily randomized clinical trials, for clinical decisions.
• Systematic reviews aggregate trial evidence to recommend optimal treatments but struggle to keep pace with growing publications.
• Computational approaches are needed to support timely systematic reviews with the latest evidence.

Purpose of the Study:

• To develop a method for synthesizing clinical trial evidence on-demand.
• To create a hierarchical argument structure for comparing therapies based on clinical endpoints.
• To implement and evaluate a web tool for exploring synthesized evidence.

Main Methods:

• An argumentation-based method was developed to synthesize evidence from clinical trials.
• The method arranges evidence hierarchically to recommend superior therapies.
• A web tool was implemented, allowing users to adjust evidence points and endpoint preferences.

Main Results:

• The method successfully generated conclusions comparable to existing systematic reviews in two use cases.
• A survey and usability analysis with medical professionals indicated the tool's value.
• The tool was perceived as a valuable aid for clinical decision-making and complementing guidelines.

Conclusions:

• The proposed argumentation-based method effectively supports the synthesis of clinical trial evidence.
• A limitation is the reliance on a manually populated knowledge base.
• Future work includes using natural language processing to automate information extraction from publications.