Interpretable NLP-Based Analysis of Longitudinal Electronic Health Records for Early Prediction of Parkinson’s Disease: A Retrospective Multicenter Study in India
DOI:
https://doi.org/10.15343/0104-7809.202650e20702026PKeywords:
Parkinson’s Disease, Natural Language Processing, Electronic Health Records, Early Detection, Temporal Embedding, SHAP interpretabilityAbstract
Parkinson’s disease (PD) is often diagnosed after substantial delay in India, limiting timely intervention. Unstructured longitudinal electronic health records (EHRs) may contain prodromal signals, but NLP-based approaches are understudied in Indian settings.We conducted a retrospective multicenter study using de-identified EHRs from AIIMS New Delhi and PGIMER Chandigarh (2015–2025). From 45,000 documents across 15,000 patients with neurological symptoms, 2,500 had confirmed PD. After exclusions and NLP pre-filtering, the final cohort comprised 8,500 patients (1,800 PD-positive, 6,700 controls). PD was defined by MDS clinical criteria; DaTSCAN supported 78% of diagnoses. Mimics were excluded. We developed an NLP-XGBoost pipeline using BioBERT for named entity recognition, negation detection, and temporal BiLSTM embeddings aggregated via attention pooling. A temporal split prevented leakage; a 6-month pre diagnosis washout excluded diagnostic workup notes. SHAP provided interpretability. The model achieved AUC 0.923 (95% CI 0.908 0.937), sensitivity 89.4%, specificity 91.2%, PPV 85.6%, NPV 93.1%, and F1-score 0.874, outperforming logistic regression (AUC 0.847) and CNN (AUC 0.891). Calibration was good (Brier 0.082; slope 1.02), PR-AUC was 0.891, and NRI was 0.32. External validation at SGPGI Lucknow yielded AUC 0.897. PD risk was predicted 2–5 years before diagnosis, with performance across horizons (AUC 0.941 at 6 months to 0.898 at 5 years). SHAP identified slow turning (mean |SHAP| 0.42), speech dysfluency (0.31), and REM sleep complaints (0.28) as top predictors. Rural patients showed bradykinesia signals 18 months earlier. Among 1,240 high-risk patients, 68% converted within 4 years (HR 12.4). Retrospective modeling projected a 41% reduction in diagnostic delay. Interpretable NLP of Indian EHRs offers a scalable, imaging-free prescreening approach with potential to support earlier referral. Prospective validation is warranted to confirm clinical impact.
Downloads
Translation
References
Allwright M, Mundell H, Sutherland G, Austin P, Guennewig B. Machine learning analysis of the UK Biobank reveals IGF-1 and inflammatory biomarkers predict Parkinson’s disease risk. PLoS ONE [Internet]. 2023b May 9;18(5):e0285416. DOI: https://doi.org/10.1371/journal.pone.0285416
Downing GJ, Tramontozzi LM, Garcia J, Villanueva E. Harnessing internet search data as a potential tool for medical diagnosis: Literature review. JMIR Mental Health [Internet]. 2024b Dec 26;12:e63149. DOI: https://doi.org/10.2196/63149
Eliwa EHI, El-Hafeez TA. Particle swarm optimization framework for Parkinson’s disease prediction. PeerJ Computer Science [Internet]. 2025b Sep 11;11:e3135. DOI: https://doi.org/10.7717/peerj-cs.3135
Elshehabi M, Hansen C, Hobert MA, Von Thaler A, Brockmann K, Tallapragada B, et al. Turning slowly predicts future diagnosis of Parkinson’s Disease: A Decade‐Long Longitudinal Analysis. Annals of Neurology [Internet]. 2025b Oct 21;99(1):114–23. DOI: https://doi.org/10.1002/ana.78034
Fan, et al. Development and validation of a machine learning-based diagnostic model for Parkinson’s disease in community-dwelling populations: Evidence from the China health and retirement longitudinal study (CHARLS). Parkinsonism & Related Disorders [Internet]. 2024 Oct 30;130:107182. DOI: https://doi.org/10.1016/j.parkreldis.2024.10718
Fu M, et al. Identifying Common Disease Trajectories of Progressive Supranuclear Palsy with Electronic Health Records. Movement Disorders Clinical Practice [Internet]. 2025 May 9;12(10):1528–38. DOI: https://doi.org/10.1002/mdc3.70119
Stavers-Sosa et al. Protocol for designing a model to predict the likelihood of psychosis from electronic health records using natural language processing and machine learning. The Permanente Journal [Internet]. 2024 Sep 2;28(3):23–36. DOI: https://doi.org/10.7812/tpp/23.139
Hossain MK, et al. Optimizing Alzheimer’s disease prediction through ensemble learning and feature interpretability with SHAP‐based feature analysis. Alzheimer S & Dementia Diagnosis Assessment & Disease Monitoring [Internet]. 2025 Jul 1;17(3):e70162. DOI: https://doi.org/10.1002/dad2.70162
Ibrahim AM, Mohammed MA. A comprehensive review on advancements in artificial intelligence approaches and future perspectives for early diagnosis of Parkinson’s disease. International Journal of Mathematics Statistics and Computer Science [Internet]. 2024 Jan 26;1:173–82. DOI: https:// doi.org/10.59543/ijmscs.v2i.8915
Jiang et al. Machine learning methods for the detection and prediction of cognitive impairment in Parkinson’s disease: a systematic review and meta analysis. Frontiers in Aging Neuroscience [Internet]. 2025 Dec 11;17:1704039. DOI: https://doi.org/10.3389/fnagi.2025.1704039
Johnson et al. Wearable Sensor-Based Assessments for Remotely Screening Early-Stage Parkinson’s Disease. Sensors [Internet]. 2024 Aug 30;24(17):5637. DOI: https://doi.org/10.3390/s24175637
Joshi H. Natural language processing of electronic health records for predicting Alzheimer’s disease. In: Advances in bioinformatics and biomedical engineering book series [Internet]. 2024. p. 141–74. DOI: https://doi.org/10.4018/979-8-3693-6442-0.ch006
Krismer F, Seppi K, Poewe W. The role of neuro-imaging in multiple system atrophy. Journal of Neural Transmission [Internet]. 2025b Jul 12;133(5):875 89. DOI: https://doi.org/10.1007/s00702-025-02964-6
Lefkovitz I, Walsh S, Blank LJ, Jetté N, Kummer BR. Direct Clinical Applications of natural language processing in common neurological disorders: Scoping review. JMIR Neurotechnology [Internet]. 2024b May 22;3:e51822. DOI: https://doi.org/10.2196/51822
Li A, Lian J, Vardhanabhuti V. Multi-modal machine learning approach for early detection of neurodegenerative diseases leveraging brain MRI and wearable sensor data. PLOS Digital Health [Internet]. 2025b Apr 25;4(4):e0000795. DOI: https://doi.org/10.1371/journal.pdig.0000795
Liang, et al. The pyroptosis mediated biomarker pattern: an emerging diagnostic approach for Parkinson’s disease. Cellular & Molecular Biology Letters [Internet]. 2024 Jan 3;29(1):7. DOI: https://doi.org/10.1186/s11658-023-00516-y
Liu Z, et al. The use of fully immersive virtual reality for screening neurodegenerative diseases: A systematic review of behavioral and diagnostic outcomes. Alzheimer S & Dementia Diagnosis Assessment & Disease Monitoring [Internet]. 2026 Jan 1;18(1):e70244. DOI: https://doi.org/10.1002/ dad2.70244
Loo RTJ, et al. Multi-cohort machine learning identifies predictors of cognitive impairment in Parkinson’s disease. Npj Digital Medicine [Internet]. 2025 Jul 26;8(1):482. DOI: https://doi.org/10.1038/s41746-025-01862-1
Ma L, Lin S, Jin J, Wang Z, Wang X, Chen Z, et al. Objective assessment of gait and posture symptoms in Parkinson’s disease using wearable sensors and machine learning. Frontiers in Aging Neuroscience [Internet]. 2025b Aug 8;17:1618764. DOI: https://doi.org/10.3389/fnagi.2025.1618764
Mekkes NJ et al. Identification of clinical disease trajectories in neurodegenerative disorders with natural language processing. Nature Medicine [Internet]. 2024 Mar 12;30(4):1143–53. DOI: https://doi.org/10.1038/s41591-024-02843-9
Pei G, Hu M, He Y, Yang X, Liu H, Jiang B, et al. Disruption of low-frequency narrowband EEG microstate networks in Parkinson’s disease with mild cognitive impairment. Journal of NeuroEngineering and Rehabilitation [Internet]. 2025b Oct 22;22(1):218. DOI: https://doi.org/10.1186/s12984-025 01747-0
Rafaï I, et al. Public hesitancy for AI-based detection of neurodegenerative diseases in France. Scientific Reports [Internet]. 2025 Jul 23;15(1):26849. DOI: https://doi.org/10.1038/s41598-025-11917-8
Rao NGR, et al. Advances in Computational Biology for Diagnosing Neurodegenerative Diseases: A Comprehensive review. Chinese Journal of Applied Physiology [Internet]. 2024 Jan 1;40:e20240008. DOI: https://doi.org/10.62958/j.cjap.2024.008
Soman K, Nelson CA, Cerono G, Goldman SM, Baranzini SE, Brown EG. Early detection of Parkinson’s disease through enriching the electronic health record using a biomedical knowledge graph. Frontiers in Medicine [Internet]. 2023b May 12;10:1081087. DOI: https://doi.org/10.3389/ fmed.2023.1081087
Downloads
Published
How to Cite
Issue
Section
License
Copyright (c) 2026 O Mundo da Saúde

This work is licensed under a Creative Commons Attribution 4.0 International License.




























