Analysis of Clinical Cohort Data Using Nested Case-control and Case-cohort Sampling Designs. A Powerful and...

Area of Science:

• Epidemiology
• Biostatistics
• Clinical Trial Design

Background:

• Epidemiological studies often use subsampling for large datasets due to resource limitations for exposure measurement.
• Clinical studies, particularly those investigating risk factors, typically favor full cohort analyses.

Purpose of the Study:

• To bridge the gap between epidemiological and clinical study designs regarding power and resource considerations.
• To compare the statistical power of full cohort analysis versus nested case-control and case-cohort designs.

Main Methods:

• Comparison of power formulas for various sampling designs and full cohort studies.
• Application of nested case-control, case-cohort, random subsample, and full cohort analyses to a data example.
• Calculation of standard errors for regression coefficients and required covariate information for each design.

Main Results:

• Power formulas for nested case-control and case-cohort designs are directly linked to the power of cohort studies via Schoenfeld's formula.
• Sampling designs result in relatively small losses in precision of parameter estimates compared to substantial resource savings.

Conclusions:

• Nested case-control and case-cohort designs are viable alternatives for clinical studies with low event rates, unlike random subsamples.
• Power calculations can effectively quantify the trade-off between statistical power and the number of patients analyzed when using sampling designs versus full cohorts.