Analysis of secondary outcomes in nested case-control study designs.

Area of Science:

• Epidemiologic research methods
• Biostatistics
• Public health

Background:

• Case-cohort designs are often preferred for analyzing multiple outcomes.
• Nested case-control designs traditionally focus on a single primary outcome.
• Evaluating secondary outcomes in nested case-control studies presents analytical challenges.

Purpose of the Study:

• To demonstrate the validity of analyzing secondary outcomes in nested case-control studies.
• To introduce a method combining inclusion probability weighting and jackknife standard errors.
• To compare the performance of nested case-control designs with case-cohort designs for secondary outcomes.

Main Methods:

• Utilized inclusion probability weighting for secondary outcome analysis.
• Employed an approximate jackknife standard error for statistical inference.
• Conducted simulation studies to assess type 1 error and coverage rates.
• Applied the method to data from the Cardiovascular Health Study.

Main Results:

• The proposed method provides valid inferences for secondary outcomes in nested case-control designs.
• Simulation studies confirmed appropriate type 1 error and coverage rates with sufficient sample size.
• Statistical power was comparable to case-cohort designs when outcomes were positively correlated.
• The analysis identified associations between C-reactive protein and congestive heart failure incidence.

Conclusions:

• Nested case-control studies can effectively analyze secondary outcomes using the proposed weighting and jackknife method.
• This approach enhances the utility of nested case-control designs in epidemiologic research.
• The method offers a statistically sound alternative for multi-outcome analysis without the need for a separate subcohort.