Regression analysis of clustered interval-censored data with informative cluster size

Area of Science:

• Biostatistics
• Epidemiology
• Parasitic Diseases

Background:

• Interval-censored data arise in asymptomatic disease progression, requiring clinical evaluation.
• Standard methods assume independence, which is violated in clustered data.
• Informative cluster size, where cluster size relates to outcomes, biases conventional analyses.

Purpose of the Study:

• To propose a joint model for clustered interval-censored failure data with informative cluster size.
• To analyze the relationship between cluster size and interval-censored nest-extinction times in lymphatic filariasis.

Main Methods:

• A joint modeling approach is developed to handle associations within clusters and dependency on cluster size.
• A proportional hazards model with random effects is used for failure times.
• A mixed ordinal regression model is applied to cluster size.

Main Results:

• The joint model effectively addresses the association among failure times within the same cluster.
• The model accounts for the dependency of failure times on cluster size, mitigating bias.
• Simulation studies validate the finite sample properties of the proposed estimators.

Conclusions:

• The proposed joint model provides a robust framework for analyzing clustered interval-censored data with informative cluster size.
• This approach is crucial for accurate analysis in studies like lymphatic filariasis, where cluster size influences disease progression and clearance rates.