Dirichlet process mixture models for the analysis of repeated attempt designs

Area of Science:

• Statistics
• Biostatistics
• Longitudinal Data Analysis

Background:

• Longitudinal studies often involve multiple attempts to collect data post-baseline.
• Recording the success of these attempts is crucial for understanding missing data assumptions.
• Existing models for such designs have limitations, including concerns about parametric model misspecification and lack of sensitivity analysis.

Purpose of the Study:

• To propose a novel statistical approach for longitudinal studies with repeated measurement attempts.
• To address limitations of previous parametric models and enhance sensitivity analysis for missing data.
• To provide a robust method for analyzing data where measurement success varies among subjects.

Main Methods:

• Utilized Bayesian nonparametrics to model the observed data distribution, minimizing model misspecification concerns.
• Developed a new approach for the identification and sensitivity analysis of missing data.
• Re-analyzed data from a clinical trial for severe mental illness patients with repeated attempts.

Main Results:

• The proposed Bayesian nonparametric approach offers improved handling of missing data in longitudinal studies.
• The novel sensitivity analysis method provides more reliable inference when data collection is challenging.
• Simulations were conducted to evaluate the properties and performance of the new approach.

Conclusions:

• The new method effectively minimizes issues related to model misspecification in longitudinal studies.
• It provides a valuable tool for robust statistical inference in the presence of missing data.
• This approach enhances the analysis of complex data collection scenarios, as demonstrated in a severe mental illness clinical trial.