LocalControl

LocalControl implements non-parametric bias-correction methods for observational studies to improve comparative safety and effectiveness assessments, including survival analysis with competing risks and censoring and the generation of bias-corrected personalized treatment-outcome predictions.

Key Features:

• Non-Parametric Bias Correction: Implements a suite/family of non-parametric methods to correct biases and confounding in observational data.
• Survival Analysis Support: Accommodates time-to-event analyses including competing risks and censoring.
• Personalized Predictions: Generates bias-corrected individualized predictions of treatment outcome differences.
• Heterogeneity Analysis: Supports examination of heterogeneity in treatment effect sizes across patient subgroups.
• R Package Implementation: Provided as an R package for analysis of observational study data.

Scientific Applications:

• Comparative Safety and Effectiveness Research: Enables comparative safety and effectiveness analyses in medicine and life sciences using observational data.
• Observational Studies with Multiple Confounders: Applicable to studies across fields that involve outcomes subject to multiple confounders.

Methodology:

Implements a family of non-parametric methods for bias correction, handles competing risks and censoring in survival analysis, and produces bias-corrected personalized predictions and subgroup heterogeneity analyses.