Controlled drug delivery aims to achieve an effective drug concentration in the action site for a desired period of time, while minimizing side effects. In this contribution, biodegradable poly(3-hydroxybutyrate) films were evaluated as a reservoir platform for dexamethasone controlled release. These systems were morphological and physicochemically characterized. In vitro release assays were performed for five different percentages of drug in the films and data were fitted by a mathematical model developed and validated by our research group. When the profiles were normalized, a single curve properly fitted all the experimental data. Using this unique curve, the dissolution efficiency (DE), the time to release a given amount of drug (tX% ), and the mean dissolution time were calculated. Furthermore, the dissolution rate, the initial dissolution rate (a%) and the intrinsic dissolution rate were determined. The a% mean value was 1.968 × 10-2% released/min, t80% was about 14 days, and the DE was 59.6% at 14 days and 66.5% at 20 days. After 2 days, when approximately 40% of the drug was released, the dissolution rate decreased about 60% respect to the initial value. The poly(3-hydroxybutyrate) platforms behaved as an appropriate system to release and control the dexamethasone delivery, suggesting that they could be an alternative to improve drug therapy.

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.