Assessing and visualizing fragility of clinical results with binary outcomes in R using the fragility package

Area of Science:

• Biostatistics
• Clinical Research Methodology
• Health Informatics

Background:

• Growing concerns regarding research reproducibility and replicability.
• The fragility index quantifies the robustness of statistical significance in clinical studies with binary outcomes.
• Assessing result reliability is crucial for informed clinical decision-making.

Purpose of the Study:

• To introduce an R package named "fragility" for calculating and visualizing the fragility index.
• To provide user-friendly functions for assessing the robustness of individual studies and meta-analyses.
• To demonstrate the application of the "fragility" package with practical examples.

Main Methods:

• Overview of methods for assessing the fragility index in individual studies, pairwise meta-analyses, and network meta-analyses.
• Development and introduction of the "fragility" R package.
• Illustration of package usage through worked examples.

Main Results:

• The "fragility" R package offers accessible tools for calculating and visualizing the fragility index.
• The package supports the assessment of fragility for single studies, pairwise meta-analyses, and network meta-analyses.
• Demonstrated ease of use and practical applicability of the "fragility" package.

Conclusions:

• The "fragility" package simplifies the complex task of assessing and visualizing result robustness.
• Enhanced accessibility to fragility index calculations aids clinicians in evaluating study reliability.
• Facilitates better understanding and application of research reproducibility measures in clinical practice.