Design and Development of an AI-Driven Personalized Microlearning Model for Canine Rehabilitation Education

Objective: Canine rehabilitation requires integrative knowledge and advanced clinical reasoning. However, lecture-based instruction can restrict individualized feedback, equal case exposure, and timely learning opportunities. We developed and evaluated the first-cycle implementation of an AI-driven personalized microlearning model for canine rehabilitation education using a design-based research (DBR) approach. Design: Design-based research Methods: Following iterative DBR cycles, we conducted a needs analysis, which included literature review and expert interviews/surveys. We also specified objectives, content modules, assessments, and adaptive rules, and then developed a proto type focused on postoperative tibial plateau leveling osteotomy (TPLO) rehabilitation, organized into microlearning units. A multidisciplinary expert panel (n = 5) reviewed the content using the Delphi method. Quantitative evaluation included a content validity index (CVI; threshold ≥ 0.80) and constructive alignment audit. The algorithmic logic was tested for coverage and rule conflicts. Qualitative data (interviews, consultation records, and observations) were subjected to inductive content analysis with investigator triangulation and member checking. Results: All modules exceeded the CVI threshold; welfare, safety,and handling (CVI = 0.95) and assessment and outcomes (CVI = 0.93) scored the highest. Thealignment audit identified and corrected gaps between objectives and assessments. Logic testing demonstrated complete coverage without rule conflicts. Qualitative analyses yielded three themes: strong clinical relevance, convenience, and motivational benefits of microlearning, as well as calls for greater AI explainability. Conclusions: The model demonstrated content validity, structural coherence, and field relevance for TPLO rehabilitation education, suggesting a possible way to bridge the theory–practice gap while enabling individualized learning. Future studies should conduct longitudinal effectiveness studies, extend them to additional conditions, and enhance the explainability features to strengthen learner trust and impact clinical outcomes.