Analysis of multiple 2x2 tables with site-specific periodontal data

Area of Science:

• Periodontal medicine
• Biostatistics
• Clinical trial analysis

Background:

• Periodontal data often involves multiple observations per patient, leading to positive correlations.
• Standard statistical methods assuming independence are inappropriate for such correlated data.
• Existing regression models for correlated data are complex, and simpler methods have limitations regarding site-specific covariates.

Purpose of the Study:

• To present two novel statistical methods for analyzing site-specific periodontal data organized in multiple 2x2 tables.
• To offer a simplified approach to analyzing correlated periodontal data, addressing limitations of existing methods.
• To demonstrate the utility of these methods using clinical trial data on scaling and root planing effects.

Main Methods:

• Modification of the established Mantel-Haenszel methods.
• Application to multiple 2x2 tables representing site-specific periodontal observations.
• Utilizing data from a clinical trial investigating scaling and root planing impacts.

Main Results:

• The proposed methods provide a viable way to analyze complex, correlated periodontal data.
• The modified Mantel-Haenszel techniques offer a more accessible approach compared to intricate regression models.
• Illustrative analysis using clinical trial data demonstrates the practical application of these methods.

Conclusions:

• The developed methods offer a practical and simplified statistical framework for analyzing site-specific periodontal data.
• These techniques enhance the ability to interpret results from clinical studies involving correlated periodontal measurements.
• The study contributes improved analytical tools for periodontal research and clinical trials.