The analysis of bivariate truncated data using the Clayton copula model

Area of Science:

• Biostatistics
• Virology
• Epidemiology

Background:

• Quantitative human immunodeficiency virus (HIV) RNA measurements in infected individuals are often left-censored, falling below assay detection limits (DL).
• Understanding the relationship between plasma and semen viral loads is crucial for HIV management and transmission studies.

Purpose of the Study:

• To develop and validate a statistical methodology for analyzing bivariate truncated data, specifically for HIV RNA viral loads.
• To assess the Clayton model assumption and estimate dependence parameters for left-censored bivariate data.

Main Methods:

• Development of an empirical goodness-of-fit test for bivariate truncated data.
• Utilization of truncated tau to estimate the dependence parameter within the Clayton model.
• Application of the proposed methodology to both truncated and fixed left-censored bivariate data.

Main Results:

• The developed methodology effectively handles bivariate data with left-censoring and truncation.
• The empirical goodness-of-fit test successfully checks Clayton model assumptions for this data type.
• Statistical inference was drawn from an HIV dataset, demonstrating the practical utility of the approach.

Conclusions:

• The proposed statistical approach provides a robust framework for analyzing censored viral load data.
• This method enhances the understanding of plasma and semen viral load relationships in HIV-infected individuals.
• The findings support improved statistical inference in virological studies with detection limit challenges.