Absolute risk from double nested case-control designs: cause-specific proportional hazards models with and without...

Area of Science:

• Biostatistics
• Epidemiology
• Survival Analysis

Background:

• Competing risks present challenges in survival analysis.
• Double nested case-control (DNCC) designs offer efficient data collection but require specialized estimation methods.
• Estimating absolute risks and relative hazards accurately is crucial for understanding disease progression and outcomes.

Purpose of the Study:

• To develop and validate efficient estimators for cause-specific proportional hazards models using DNCC data.
• To improve the estimation of absolute risks and relative hazards in the presence of competing risks.
• To leverage complete covariate data from the phase-two sample and partial data from the full cohort.

Main Methods:

• Utilized design-weighted estimators based on inverse sampling probabilities for DNCC data.
• Augmented standard estimators with a term to incorporate additional cohort data, enhancing efficiency.
• Established asymptotic properties and derived consistent variance estimators for the proposed methods.
• Conducted simulations to assess the performance of the estimators in practical sample sizes.

Main Results:

• Augmented design-weighted estimators demonstrated greater efficiency compared to standard design-weighted estimators.
• The proposed asymptotic methods exhibited nominal operating characteristics in simulations.
• Validated the methodology using real-world data from the Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial.

Conclusions:

• The developed augmented estimators provide a more efficient approach for analyzing competing risks in DNCC studies.
• The methods are robust and suitable for practical application in epidemiological research.
• Accurate estimation of absolute risks and relative hazards is achievable with the proposed techniques.