Statistical approaches for analyzing immunologic data of repeated observations: A practical guide

• Practical guide for analyzing repeated cross-sectional measurements.
• Immunologic parameters on the same subject account for their correlations.
• Step by step analytic methods account for heterogeneity among human subjects.
• Present statistical concepts and their respective advantages and shortcomings.
• Link to data structures, documented code, interpretation.

Translational research not only encompasses transitioning from animal to human models but also must address the greater heterogeneity of humans when designing and analyzing experiments. Appropriate study designs can address heterogeneity through a priori data collection, and taking repeated measures can improve the power and efficiency of a study to detect clinically meaningful differences. Although common in other areas of biomedical research, modern statistical methods using repeated measurements on the same subject and accounting for their potential correlations are not widely utilized in immunologic studies. To highlight these analytic issues, we present a practical guide to understanding and applying analytic methods from commonly used T-tests without adjusting for multiple comparisons to mixed models with subject-specific adjustments for correlations using our data on Toll-like receptor-induced cytokine production in monocytes from young and older adults.