A multivariate test for population bioequivalence.

Area of Science:

• Biostatistics
• Pharmacometrics
• Analytical Chemistry

Background:

• Univariate population bioequivalence testing is well-established.
• Existing multivariate equivalence methods often focus solely on mean comparisons.
• There is a need for comprehensive multivariate bioequivalence testing that includes variance.

Purpose of the Study:

• To propose a multivariate generalization of univariate bioequivalence testing criteria.
• To develop a novel aggregate criterion combining mean and variance comparisons.
• To evaluate the performance of the proposed method in finite samples.

Main Methods:

• Developed a multivariate generalization using the trace of the covariance matrix.
• Employed a confidence interval approach for hypothesis testing.
• Constructed two versions of a modified large-sample confidence interval for a linearized multivariate criterion.
• Conducted a simulation study to assess empirical coverage and rejection rates.

Main Results:

• The proposed method integrates mean and variance comparisons into a single criterion.
• Simulation results provide insights into the performance of the confidence intervals and tests.
• The methodology is demonstrated with a practical example of nasal spray bioequivalence.

Conclusions:

• The proposed multivariate approach offers a robust method for assessing population bioequivalence.
• This generalization extends existing bioequivalence testing to more complex scenarios.
• The method is applicable to real-world product equivalence testing, such as for nasal sprays.