fragility

fragility computes the fragility index for clinical studies with binary outcomes and for pairwise and network meta-analyses to quantify how many minimal event-status modifications are required to change statistical significance.

Key Features:

• Fragility index calculation: Computes the number of minimal event-status modifications required to alter the statistical significance of binary-outcome study results.
• Treatment-group specification: Allows specification of which treatment groups have modified event statuses when computing fragility indices.
• Statistical-test options: Evaluates associations between treatments and outcomes using chosen statistical methods for significance testing.
• Significance-level control: Uses pre-specified significance levels (alpha) when determining whether event-status modifications change significance.
• Meta-analysis support: Extends fragility index computation to conventional pairwise meta-analyses and to network meta-analyses with multiple treatment comparisons.
• Visualization: Provides visualization of fragility indices for individual studies and for meta-analyses.

Scientific Applications:

• Individual study robustness assessment: Quantifies the robustness of clinical study results with binary outcomes by estimating fragility indices.
• Meta-analysis robustness evaluation: Assesses the robustness of pooled estimates across studies in pairwise and network meta-analyses.
• Result sensitivity analysis: Evaluates how minimal changes in event statuses affect statistical significance and the reproducibility of clinical findings.

Methodology:

Determines the minimal number of event-status modifications in specified treatment groups required to change statistical significance according to a chosen statistical test and a pre-specified significance level, and applies the same computation to pairwise and network meta-analyses.