Simple hybrid predictive model for point-of-care pharmaceutical tablet disintegration

Ensuring the consistent disintegration performance of immediate-release oral solid dosage (OSD) forms throughout their shelf life is critical for therapeutic efficacy, yet predicting point-of-care performance from manufacturing data remains a significant challenge. Current predictive approaches often rely on empirical correlations or high-dimensional black-box models that lack mechanistic insight and longitudinal validity. To address this, a simple hybrid model that predicts longitudinal disintegration times t_disint,T solely from at-line porosity measurements f_at-line is established. The framework is anchored in United States Pharmacopoeia <701> standards to ensure regulatory compliance and reproducibility, while maintaining an architecture that is agnostic to the specific porosity characterisation technique. Utilising a five-year real-time stability dataset with different formulations and tablet geometries, global models are established and demonstrate exceptional predictive accuracy (R²>0.92) across all longitudinal time points. The operating range lies in f_at-line∈[0.1,0.3], which is the typical porosity range of pharmaceutical OSDs. The model's longitudinal validity is predicated on the physicochemical stability of the formulation, allowing the model parameters to serve as quantitative indicators in stability assessments. The simplicity and adaptability of this mechanistically grounded model offer a pragmatic path to demonstrating dosage form efficacy at the point-of-care, providing a robust alternative to traditional, destructive end-point testing. By relating at-line structural attributes to long-term performance, this model facilitates Quality by Design in development, seamless technical transfer, and robust real-time release testing. Ultimately, this framework ensures that the rigorously designed quality standards are demonstrably preserved until the moment of patient administration.