Adjustment for competing risk in kin-cohort estimation.

Area of Science:

• Genetics and Epidemiology
• Statistical Genetics
• Cancer Epidemiology

Background:

• Kin-cohort studies analyze genetic mutation risks across relatives.
• Existing methods struggle with multiple, related events and censoring.
• Bias in risk estimation arises when events are interconnected and influence follow-up.

Purpose of the Study:

• To develop a statistical framework for kin-cohort studies to estimate risks of multiple, competing events.
• To address limitations of current methods that assume independence of censoring events.
• To accurately assess the cause-specific hazard functions for genetic mutation carriers and noncarriers.

Main Methods:

• Utilized a competing risk framework to model multiple event occurrences.
• Proposed an extension of a previously developed composite-likelihood approach for estimation.
• Applied the method to estimate ovarian cancer risk from BRCA1/2 mutations, excluding breast cancer cases.

Main Results:

• Demonstrated that cause-specific hazard functions are identifiable from kin-cohort data within the competing risk framework.
• The proposed composite-likelihood extension provides a viable estimation method.
• Successfully illustrated the method's application using data from the Washington Ashkenazi Kin-Cohort Study.

Conclusions:

• The competing risk framework effectively handles multiple, related events in kin-cohort studies.
• The extended composite-likelihood method offers improved and unbiased risk estimation for genetic mutation effects.
• This approach enhances the understanding of genetic risks for complex diseases like ovarian cancer.