Exact probabilities and confidence limits for binomial samples: applied to the difference between two proportions

Area of Science:

• Biostatistics
• Medical Statistics
• Clinical Trial Analysis

Background:

• Accurate confidence intervals are crucial for interpreting medical binomial parameters.
• Existing methods like Wald and Wilson have limitations in precision and reliability.
• The 2x2 contingency table is a fundamental tool in medical research for comparing proportions.

Purpose of the Study:

• To propose and evaluate an exact probabilities method for computing confidence limits of medical binomial parameters.
• To assess the performance of this new method for the difference between two binomial proportions.
• To compare the proposed method against established techniques like Wald and Wilson intervals.

Main Methods:

• Development of an algorithm based on exact probabilities for confidence limit calculation.
• Application and assessment of the algorithm to a bidimensional parameter: the difference between two binomial proportions.
• Comparative analysis against Wald and Wilson methods, with and without continuity corrections.

Main Results:

• The exact probabilities method demonstrated monotonic behavior in computing confidence limits.
• Experimental errors associated with the exact probabilities method did not exceed the 5% significance level.
• The proposed method showed reliable performance when applied to the difference between two proportions.

Conclusions:

• The exact probabilities method offers a reliable and accurate approach for calculating confidence limits for binomial parameters in medical research.
• This method provides a robust alternative to existing techniques, particularly for the difference between two proportions.
• The findings support the use of the exact probabilities method in clinical trial analysis and medical statistics for improved data interpretation.