[Applying multilevel models in evaluation of bioequivalence (I)]

Area of Science:

• Pharmacokinetics and Pharmacodynamics
• Statistical Modeling
• Drug Development

Background:

• Bioequivalence studies are crucial for generic drug approval.
• Current FDA methods rely on the Method of Moments.
• Evaluating bioequivalence requires assessing average, population, and individual bioequivalence.

Purpose of the Study:

• To explore the application and value of multilevel models in bioequivalence evaluation.
• To compare multilevel models with existing FDA methods using a real-world antihypertensive drug study.
• To assess the feasibility and applicability of multilevel models for different bioequivalence metrics.

Main Methods:

• Utilized a 2 x 4 cross-over experimental design.
• Applied multilevel models to analyze variance components (total, between-subject, within-subject).
• Compared results with the FDA's standard Method of Moments for antihypertensive drug bioequivalence.

Main Results:

• Variance components estimated by multilevel models closely matched those from the FDA Method of Moments.
• Multilevel models demonstrated feasibility and applicability in assessing average bioequivalence (ABE), population bioequivalence (PBE), and individual bioequivalence (IBE).
• Both multilevel models and FDA criteria yielded consistent bioequivalence evaluation results.

Conclusions:

• Multilevel models are a viable and flexible tool for bioequivalence evaluation in cross-over designs.
• This approach allows for more detailed variance decomposition, aiding in ABE, PBE, and IBE assessment.
• Multilevel modeling presents a novel and practical method for bioequivalence studies.