소아 저체중 한의 진료 CDSS 프레임워크 : LPA-GBTM-HTE의 통합적 접근

A CDSS Framework for Pediatric Underweight in Korean Medicine: Integrating LPA, GBTM, and HTE

Background: Pattern identification and treatment in Korean medicine embodies the essence of personalized medicine. However, difficulties in objectification and quantification have limited the development of modern Clinical Decision Support Systems (CDSS). This study proposes a statistical methodological framework for implementing data-driven pattern identification in Korean medicine management of pediatric underweight. Methods: We searched PubMed, Google Scholar, and RISS for literature on Latent Profile Analysis (LPA), Group-Based Trajectory Modeling (GBTM), and Heterogeneous Treatment Effects (HTE). Key literature was selected according to inclusion and exclusion criteria, and we narratively reviewed the theoretical foundations, clinical applications, and applicability to Korean medicine for each methodology. Results: LPA can objectify pattern identification by discovering patient subgroups from multidimensional symptom patterns at initial consultation. GBTM quantifies prognosis by modeling heterogeneous trajectories of treatment response over time. HTE analysis supports optimal prescription decisions by estimating differential treatment effects according to patient characteristics. The integrated LPA-GBTM-HTE framework, connecting these methods sequentially, can serve as an interpretable analytical engine for CDSS that reproduces the clinical logic of pattern identification and treatment in Korean medicine based on data. Conclusion: This study provides a theoretical foundation for statistical methodologies in precision Korean medicine. Future studies are needed to validate the framework using real clinical data and develop prototype CDSS.