dsSurvival

dsSurvival performs privacy-preserving survival analysis by enabling federated meta-analysis of Cox proportional hazards models within the DataSHIELD framework to estimate hazard ratios without sharing individual-level data.

Key Features:

• Privacy-Preserving Framework: Operates under a federated meta-analysis approach in which only anonymized aggregated data are exchanged between institutions.
• Interactive Modeling: Supports exploratory and interactive modeling while avoiding manual intervention from data providers at each site.
• Meta-Analysis of Cox Regression Models: Tailored for meta-analysis of Cox proportional hazards models and automates integration of results from individual studies.
• Calculation of Hazard Ratios: Provides functionality to calculate hazard ratios for covariate effects in survival analyses.

Scientific Applications:

• Multi-site survival studies: Enables large-scale biomedical survival analyses that combine data from multiple sources to achieve sufficient statistical power.
• Collaborative federated meta-analysis: Supports cross-institutional meta-analysis of Cox models to estimate covariate effects while keeping patient-level data local.

Methodology:

Performs local analysis at individual study sites, exchanges anonymized aggregated summary statistics, and integrates results using meta-analytic techniques within the DataSHIELD framework.