A comparative study of confidence intervals to assess biosimilarity from analytical data

Area of Science:

• Biopharmaceutical Analytics
• Statistical Modeling
• Regulatory Science

Background:

• Assessing analytical similarity for biosimilarity relies on hypotheses comparing product attributes.
• Accurate statistical tests must address uncertainties in mean differences and reference product standard deviation.

Purpose of the Study:

• To survey and compare methods for constructing confidence intervals for a linearized biosimilarity hypothesis.
• To evaluate the performance of statistical tests for analytical similarity assessment.

Main Methods:

• Survey of confidence interval construction methods for linearized biosimilarity hypothesis.
• Computer simulation study comparing method performance (test size, power, computational complexity).
• Application of methods to two real-world examples.

Main Results:

• Comparison of recently developed and previously established methods for analytical similarity testing.
• Evaluation of statistical power, test size, and computational efficiency across different methods.
• Demonstration of practical application through case studies.

Conclusions:

• Recommendations are provided for selecting appropriate statistical methods for demonstrating analytical similarity.
• The study facilitates informed choices in biosimilar development and regulatory assessment.
• The linearized hypothesis offers a practical framework for robust biosimilarity testing.