The Fine-Gray Model Under Interval Censored Competing Risks Data

Area of Science:

• Biostatistics
• Survival Analysis
• Epidemiology

Background:

• Competing risks data present challenges in accurately modeling event probabilities.
• Interval censoring, where event times are known only within a range, further complicates analysis.
• Existing methods may not adequately address mixed-case interval censoring in competing risks scenarios.

Purpose of the Study:

• To develop and evaluate a semiparametric method for analyzing competing risks data with mixed-case interval censoring.
• To apply the Fine-Gray model with sieve semiparametric maximum likelihood estimation.
• To assess the performance of univariate and multivariate likelihood approaches.

Main Methods:

• Utilized the Fine-Gray model to estimate cumulative incidence functions.
• Employed sieve semiparametric maximum likelihood estimation.
• Investigated both univariate (cause-specific) and multivariate (full data) likelihoods.
• Assessed asymptotic normality and semiparametric efficiency of regression parameter estimators.

Main Results:

• The regression parameter estimator demonstrates asymptotic normality and semiparametric efficiency under both likelihood approaches.
• The spline-based sieve estimator for the baseline cumulative subdistribution hazard converges at a sub-root-n rate.
• Simulation studies confirmed the finite sample performance of the proposed method.

Conclusions:

• The proposed semiparametric method provides a robust framework for analyzing competing risks data with mixed-case interval censoring.
• The method allows for efficient estimation of regression parameters in complex survival data.
• The findings are applicable to epidemiological studies, such as dementia cohort analyses.