A Decision Tree for the Rational Selection of Mathematical Models in Drug Dissolution and Release Studies

Background/Objectives: In vitro dissolution tests are an essential tool in pharmaceutical development, allowing for the analysis of biopharmaceutical properties, understanding release mechanisms, and comparing formulations. The wide variety of mathematical models available for interpreting the profiles obtained creates ambiguity and difficulty in their selection and application. The objective of this work was to develop a decision tree algorithm that, based on the initial observation of the experimental profile, rationally guides the selection of the most appropriate mathematical model, considering their advantages and limitations. Methods: A review of classical and recent dissolution/release models was conducted, highlighting their constraints and parameters of pharmaceutical relevance. Based on this information, a decision tree was designed, integrating observational (curve shape, burst or lag time phenomena), statistical (R², AIC), and interpretability criteria. The algorithm was validated using topical and oral representative systems: hydrogels, polymeric films, modular systems (Dome Matrix), and 3D-printed pills. Results: The decision tree allowed reducing the number of candidate models and guiding the selection toward equations consistent with the observed phenomena. The usefulness of the Lumped–Gonzo model, capable of fitting complete profiles and providing physically meaningful parameters, was highlighted compared to classic models such as Weibull or Korsmeyer–Peppas. The methodology proved to be versatile and applicable to different release mechanisms. Conclusions: The proposed algorithm constitutes a flexible and practical tool that facilitates the rational selection of mathematical models. It does not replace the researcher's judgment, but rather complements it, promoting a more efficient use of mathematical modeling in the development of dosage forms.