Elementary methods of cohort analysis with several exposure levels

Area of Science:

• Epidemiology
• Biostatistics
• Statistical Methods

Background:

• The Mantel-Haenszel method is a standard for analyzing dichotomous exposure data.
• Existing methods are limited when dealing with multiple exposure levels.

Purpose of the Study:

• To extend the Mantel-Haenszel rate ratio for dichotomous exposures to situations with multiple exposure levels.
• To develop unbiased estimating functions for improved statistical analysis.

Main Methods:

• Utilized the unbiased estimating function approach.
• Derived closed-form Mantel-Haenszel rate ratio estimators.
• Developed dually consistent variance and covariance estimators.
• Provided a quasi-score-based confidence interval.

Main Results:

• Successfully extended the Mantel-Haenszel rate ratio to multiple exposure levels.
• Obtained closed-form estimators for rate ratios.
• Established variance and covariance estimators for the extended method.
• Developed a confidence interval for common rate ratios.

Conclusions:

• The proposed extension provides a robust method for analyzing rate ratios with multiple exposure levels.
• The new estimators and confidence intervals enhance the precision of epidemiological studies.
• The methodology is also adaptable to common rate difference calculations.