Analysis of Secondary Outcomes in Nested Case-Control Study Designs

• Activity Number: 92
• Type: Invited
• Date/Time: Sunday, August 4, 2013 : 8:30 PM to 10:30 PM
• Sponsor: Korean International Statistical Society
• Author(s): Ryung S. Kim*+
• Companies: Albert Einstein College of Medicine
• Keywords: Incidence density sampling ; Nested case-control design ; Secondary outcomes ; Approximate Jackknife standard error

Abstract

It is widely perceived that inference about hazard ratios for secondary outcomes can be made in case-cohort studies but not in nested case-control studies. This widely spread perception is still one of the main reasons studies choose case-cohort designs over nested case-control design. There has been a complete lack of secondary outcomes analysis in the literature based on nested case control studies using Cox's proportional hazards model. However some existing methods for primary outcomes in nested case-control studies can also be used for secondary outcomes by defining the risk sets with the secondary outcomes. This is the first work demonstrating with numerical and real examples that valid inferences can be made about hazard ratios for secondary outcomes under Cox model in nested case-control studies. We used an inclusion probability method originally proposed by Samuelson and a new approximate jackknife standard error that is readily computable by existing software. The numerical example demonstrates that, when sample size is sufficient, the method yields valid type 1 errors and coverage rates for hazard ratios with respect to secondary outcomes. Interestingly, the statistical