clogitRV

clogitRV performs conditional logistic regression with a robust variance estimator to analyze nested, matched case-control studies with multiple biomarker measurements per individual.

Key Features:

• Conditional Logistic Regression (CLR): Applies CLR to matched case-control data to estimate marginal relationships between biomarker measurements and event risk.
• Robust Variance Estimation: Incorporates a robust variance estimator to provide reliable inference accounting for within-individual and within-match variability.
• Comprehensive Use of Biomarker Measurements: Leverages all available biomarker measurements per individual rather than using only the first or averaged values.
• Explicit Matching Adjustment: Explicitly accounts for matching in nested matched case-control study designs.
• Comparative Evaluation: Compares performance against using only the first measurement, averaged measurements, and Generalized Estimating Equations (GEE).
• Simulation-based Assessment: Uses simulations to evaluate statistical power and compare methods.

Scientific Applications:

• CLUE cohort analysis: Applied to the CLUE cohort to assess biomarker associations with disease risk.
• sCD27 and non-Hodgkin lymphoma (NHL): Identified a significant association between increased sCD27 levels and NHL.
• Temporal inference of association: Evaluated the strength of the sCD27–NHL association over time until diagnosis to support interpretation of elevated sCD27 as an effect rather than a cause of NHL.

Methodology:

Implements conditional logistic regression including all biomarker measurements with a robust variance estimator, compares against first-measurement, averaged-measurement, and GEE approaches, and assesses performance via simulation studies.