Multiple imputation methods for estimating regression coefficients in the competing risks model with missing cause of...

Area of Science:

• Biostatistics
• Survival Analysis
• Epidemiology

Background:

• Competing risks models are crucial for analyzing time-to-event data with multiple failure types.
• Missing cause of failure data presents a significant challenge in statistical modeling.
• Accurate estimation is vital for understanding disease progression and treatment effects.

Purpose of the Study:

• To develop a robust statistical method for estimating regression coefficients in competing risks models with missing failure cause data.
• To address the proportional hazards assumption for the cause of interest.
• To provide a consistent and asymptotically normal estimator for regression coefficients and their variance.

Main Methods:

• Proposed a novel method utilizing multiple imputation to handle missing cause of failure.
• Employed maximum partial likelihood estimation on imputed datasets.
• Combined estimators from multiple imputed datasets for improved consistency and asymptotic normality.
• Derived a consistent estimator for the asymptotic variance.

Main Results:

• The proposed method yields consistent and asymptotically normal estimates for regression coefficients.
• Simulation studies confirm the method's effectiveness in finite samples.
• The approach was successfully illustrated using a real-world breast cancer dataset.

Conclusions:

• The developed method provides a statistically sound approach for analyzing competing risks data with missing failure causes.
• This technique enhances the reliability of regression coefficient estimation in survival analysis.
• The findings have practical implications for epidemiological and clinical research, particularly in oncology.