Use of tolerance intervals for assessing biosimilarity

Area of Science:

• Pharmacology
• Biostatistics

Background:

• Biosimilars are biological products highly similar to reference drugs, with no clinically meaningful differences.
• Manufacturing processes for biosimilars can introduce variations due to their biological origin.
• Current biosimilarity assessments may overlook critical variability data by focusing solely on mean differences.

Purpose of the Study:

• To propose a novel statistical approach for assessing biosimilarity.
• To emphasize the importance of considering population variability in biosimilar comparisons.
• To introduce tolerance intervals as a more stringent method for biosimilarity evaluation.

Main Methods:

• Utilizing tolerance intervals and associated hypothesis testing for biosimilarity assessment.
• Accounting for the entire population of clinical outcomes for both test and reference products.
• Illustrating the proposed method with a real-world example.

Main Results:

• The proposed tolerance interval approach provides a more stringent assessment than confidence intervals.
• This method is particularly effective when mean differences are small but variability differs.
• It ensures a comprehensive evaluation of clinical outcome populations.

Conclusions:

• Tolerance intervals offer a robust framework for evaluating biosimilarity by incorporating variability.
• This approach enhances the rigor of biosimilar comparisons, especially in sensitive cases.
• The method ensures that biosimilars are truly comparable to reference products across their entire outcome distributions.