Research Publications

Published work on the safe integration of AI and large language models into clinical workflows.

Journal Article · December 2025

Safe-By-Design Clinical LLMs: Formal Risk Models and Silent-Trial Protocols for Workflow Integration

Charan Tej Gayapu · Macaw International Journal of Advanced Research in Computer Science and Engineering · Vol. 11, No. 2, pp. 8–15

Addresses the integration of large language models into clinical workflows with an emphasis on safety: formal risk modeling and silent-trial protocols that allow AI systems to be evaluated in live clinical environments without influencing care, before workflow deployment.

Large Language ModelsClinical AI Safety Risk ModelsSilent TrialsWorkflow Integration
Sample Entry · 2025

Transformer-Based Automated MedDRA Coding of Adverse Event Verbatim Terms

Sample publication — replace with your own paper details (authors, journal, DOI, citations)

Fine-tuned clinical language models for preferred-term assignment with a human-in-the-loop workflow for regulatory compliance.

NLPMedDRAPharmacovigilance
Sample Entry · 2024

Metadata-Driven Automation of SDTM Mapping: A Scalable Framework for Submission Readiness

Sample publication — replace with your own paper details

A reusable metadata repository and rules engine that generates SDTM datasets and define.xml, reducing mapping cycle time from weeks to days.

CDISCSDTMAutomation
Sample Entry · 2024

Linking EHR and Trial Data for Real-World Evidence Generation

Sample publication — replace with your own paper details

A privacy-preserving record-linkage methodology connecting EHR-derived endpoints to trial outcomes.

RWEEHRData Linkage
Sample Entry · 2023

Machine Learning–Enhanced Key Risk Indicators for Risk-Based Monitoring

Sample publication — replace with your own paper details

Gradient-boosted KRI models that predict site-level data quality issues earlier than threshold-based approaches.

RBMMachine LearningSite Quality
Sample Entry · 2022

Data Management Challenges in Decentralized Clinical Trials: Wearables, eCOA, and Beyond

Sample publication — replace with your own paper details

A framework for ingesting, standardizing, and quality-controlling high-frequency sensor data streams within CDISC-conformant pipelines.

DCTWearableseCOA