Evaluating competing adverse and beneficial outcomes using a mixture model.

Area of Science:

• Biostatistics
• Epidemiology
• Clinical Research Methodology

Background:

• Competing risks are common in clinical studies, where standard survival analysis can overestimate event incidence.
• Mixture models offer a way to analyze subdistributions in competing risk scenarios.
• Interventions may involve balancing adverse events against beneficial outcomes.

Purpose of the Study:

• To present methods for estimating adverse-benefit ratio curves and subhazard ratios using mixture models.
• To compare the risk of antiretroviral discontinuation/switching against HIV RNA suppression in an HIV cohort.

Main Methods:

• Utilized a mixture model framework for competing risk analysis.
• Developed parametric approaches for estimating adverse-benefit ratio curves and subhazard ratios.
• Considered extensions for event type uncertainty, left truncation, and interval censoring.

Main Results:

• Demonstrated the application of the mixture model to HIV clinical cohort data.
• Estimated the adverse-benefit ratio curve and subhazard ratios for competing events.

Conclusions:

• Mixture models provide a robust framework for analyzing competing risks, particularly for comparing adverse events to beneficial outcomes.
• The presented methods offer valuable tools for clinical research, allowing for nuanced comparisons of event risks.