Enhancing Pharmacovigilance and Real-world Evidence Generation in Tardive Dyskinesia through Electronic Health Record Analytics and Artificial Intelligence Methods

Authors

• Deepthi Krishna, University of LynchburgFollow

Specialty

Industry/Pharma

Advisor

Lawrence Herman

Abstract

The objective of this literature review is to evaluate the role of artificial intelligence (AI)-enabled electronic health record (EHR) analytics and digital phenotyping in improving pharmacovigilance and real-world evidence (RWE) generation for the early detection of tardive dyskinesia (TD). A literature search was conducted in Pubmed and arXiv for peer-reviewed studies published within the past five years using key terms related to TD, AI, EHR analytics, natural language processing (NLP), machine learning, and digital phenotyping. The reviewed literature demonstrates that AI technologies, including NLP, machine learning and large language models (LLM) substantially improve adverse drug event identification by extracting clinical insights from unstructured EHR text. Studies consistently show that AI-assisted models identify more TD cases than structured diagnostic coding alone, highlighting widespread underrecognition in routine practice. Emerging video-based AI phenotyping further demonstrates promise for objective movement analysis to complement EHR surveillance. Overall, AI-enabled EHR analytics bridge critical gaps in TD recognition by detecting clinical nuances overlooked by standard coding. Integrating these tools into routine clinical workflows can facilitate timely therapeutic intervention, enhance pharmacovigilance, and mitigate the long-term burden of TD for patients treated with dopamine receptor-blocking agents (DRBAs)

Recommended Citation

Krishna D. Enhancing Pharmacovigilance and Real-world Evidence Generation in Tardive Dyskinesia through Electronic Health Record Analytics and Artificial Intelligence Methods. University of Lynchburg Doctoral Project Assignment Repository. 2026; 8(3).