Medication Extraction and Entity Linking using Stacked and Voted Ensembles on LLMs

Abstract

Medication Extraction and Mining of its related attributes play an important role in healthcare NLP research due to its practical applications in hospital settings, such as their mapping into standard clinical knowledge bases (SNOMED-CT, BNF, etc.).In this work, we investigate state-of-the-art LLMs in text mining tasks on medications and their related attributes such as dosage, route, strength, and adverse effects.In addition, we explore different ensemble learning methods (STACK-ENSEMBLE and VOTING-ENSEMBLE) to augment the model performances from individual LLMs.Our ensemble learning result demonstrated better performances than individually fine-tuned base models BERT, RoBERTa, RoBERTa-L, BioBERT, BioClin-icalBERT, BioMedRoBERTa, ClinicalBERT, and PubMedBERT across general and specific domains, with statistical significance testing (p=0.048).Finally, we build up an entity linking function to map extracted medical terminologies into the SNOMED-CT codes and the British National Formulary (BNF) codes, which are further mapped to the Dictionary of Medicines and Devices (dm+d), and ICD (Clinical Coding).We host the fine-tuned models and desktop applications at https://github.com/ pabloRom2004/Insight-Buddy-AI-App

Authors 2

1. Lifeng Han Aachen LIACS & LUMC

   Leiden University · University of Manchester · Leiden University Medical Center

   Affiliation as printed

   LIACS & LUMC , Leiden University , NL

   The University of Manchester , UK

2. Goran Nenadić

   University of Manchester

   Affiliation as printed

   The University of Manchester , UK