Symptom-based classification of Parkinson’s disease using retrospective electronic medical records

This cross-sectional study of electronic medical records (EMRs) was conducted to investigate symptom patterns associated with Parkinson's disease (PD) duration and severity and to explore a symptom-based classification of PD. Data from 334 patients with idiopathic PD who visited a single hospital in South Korea between 2011 and 2021, were used to extract a range of PD-associated symptoms and traditional East-Asian medicine (TEAM) symptoms. Frequency analysis was performed to identify key clinical symptoms by PD progression and severity. Patients with moderate disease severity exhibited different symptom patterns across disease-duration categories, including differences in dyskinesia, speech disorder, floating pulse, and yellow tongue coating. Hierarchical clustering based on PD-related and traditional East Asian medicine symptoms identified 5 distinct EMR-based subtypes. EMR subtype 1 was characterized by the mildest disease severity and shortest disease duration, whereas the remaining subtypes showed distinct symptom profiles and a generally greater overall symptom burden. This information may be useful for identifying clinically distinguishable EMR-based PD subtypes and symptom patterns associated with disease duration and severity. These preliminary findings provide a basis for future studies investigating whether such patterns may inform individualized approaches to PD management.