Clustered data analysis under miscategorized ordinal outcomes and missing covariates

Area of Science:

• Statistics
• Biostatistics
• Data Analysis

Background:

• Ordinal data analysis is crucial in many scientific fields.
• Real-world data frequently exhibit missing covariate information and outcome miscategorization.
• Existing models may not adequately address these common data complexities.

Purpose of the Study:

• To develop and analyze a clustered ordinal model accommodating missing covariates.
• To address the challenge of miscategorized data using surrogate variables.
• To investigate the influence of age, sex, and food habits on plaque deposit using orthodontic data.

Main Methods:

• A general model structure is proposed to incorporate information from surrogate variables.
• A novel two-step estimation approach is introduced for model parameter estimation.
• The model's flexibility allows it to handle both missingness and miscategorization.

Main Results:

• The proposed model effectively handles clustered ordinal data with missing covariates.
• The two-step estimation method provides a robust way to estimate model parameters.
• Simulation studies validate the model's performance and applicability.

Conclusions:

• The developed clustered ordinal model offers a flexible solution for analyzing complex datasets.
• The proposed methodology is suitable for various applications, including orthodontic research.
• The approach effectively mitigates the impact of missing data and miscategorization on analysis.