Related Experiment Videos
Recommendation for confidence interval and sample size calculation for the Pearl Index
C Gerlinger1, J Endrikat, E A van der Meulen
1Schering AG, Berlin, Germany.
Summary
This study clarifies the statistical calculation of the Pearl Index for contraceptive effectiveness, proposing the Poisson model as superior for clinical trials. It provides definitions and sample sizes to meet European regulatory guidelines.
Area of Science:
- Pharmacovigilance
- Biostatistics
- Reproductive Health
Background:
- The European Agency for the Evaluation of Medicinal Products (EMEA) guideline requires confidence intervals for the Pearl Index.
- Current interpretation of the Pearl Index as a statistical parameter needs clarification.
- Existing guidelines lack precise definitions for Pearl Index calculation.
Purpose of the Study:
- To compare Bernoulli and Poisson models for Pearl Index calculation.
- To provide unambiguous definitions and formulas for Pearl Index and Method Failure Pearl Index.
- To determine sample sizes for clinical investigations of steroid contraceptives.
Main Methods:
- Statistical modeling: Bernoulli and Poisson models.
- Comparison of models for calculating Pearl Index and its upper confidence limit.
- Development of precise statistical formulas for overall and Method Failure Pearl Index.
Main Results:
- The Poisson model is more suitable due to its ability to handle incomplete treatment cycles.
- Unambiguous definitions and statistical formulas for Pearl Index calculations are provided.
- Required sample sizes to comply with EMEA guidelines are presented.
Conclusions:
- The Poisson model offers a more robust statistical approach for evaluating contraceptive effectiveness.
- Standardized definitions and methods improve the reliability of Pearl Index calculations.
- This work supports regulatory compliance in the clinical investigation of contraceptives.