A mathematical framework for predicting tablet weight variability from blend particle size distribution and tooling...

Area of Science:

• Pharmaceutical Manufacturing
• Materials Science
• Statistical Modeling

Background:

• Tablet weight variability is a critical quality attribute in pharmaceutical manufacturing.
• Current methods for predicting weight variability are primarily empirical.
• A need exists for a predictive framework grounded in first principles.

Purpose of the Study:

• To develop a first-principles statistical model for predicting tablet weight variability.
• To incorporate formulation and tooling parameters into the model.
• To provide a quantitative understanding of factors influencing weight variation.

Main Methods:

• Developed a statistical model based on particle-sampling theory.
• Integrated tablet weight, punch diameter, and blend particle size distribution into the model.
• Validated the model using experimental data from Suglets® spheres across various parameters.

Main Results:

• The model accurately predicts tablet weight variability (R² ≈ 0.93, normalized RMSE ≈ 0.05).
• Confirmed predicted scaling relationships with tablet mass, punch diameter, and particle size.
• Demonstrated model robustness through resampling and cross-validation techniques.

Conclusions:

• The developed framework provides a predictive tool for understanding and controlling tablet weight variability.
• Enables quantitative assessment of empirical relationships and construction of design maps.
• Offers a pathway to optimize formulation and tooling for consistent tablet quality.