Enhanced Electronic Health Records Text Summarization Using Large Language Models
It addresses the need for efficient, clinician-specific summaries in healthcare, though it is incremental as it builds on prior LLM adaptations for clinical tasks.
This project tackled the problem of generating focused summaries from Electronic Health Records for clinicians by fine-tuning the Google Flan-T5 model on an EHR dataset, achieving an Exact Match score of 81.81% and high ROUGE scores up to 96.10%.
The development of Electronic Health Records summarization systems has revolutionized patient data management. Previous research advanced this field by adapting Large Language Models for clinical tasks, using diverse datasets to generate general EHR summaries. However, clinicians often require specific, focused summaries for quicker insights. This project builds on prior work by creating a system that generates clinician-preferred, focused summaries, improving EHR summarization for more efficient patient care. The proposed system leverages the Google Flan-T5 model to generate tailored EHR summaries based on clinician-specified topics. The approach involved fine-tuning the Flan-T5 model on an EHR question-answering dataset formatted in the Stanford Question Answering Dataset (SQuAD) style, which is a large-scale reading comprehension dataset with questions and answers. Fine-tuning utilized the Seq2SeqTrainer from the Hugging Face Transformers library with optimized hyperparameters. Key evaluation metrics demonstrated promising results: the system achieved an Exact Match (EM) score of 81.81%. ROUGE (Recall-Oriented Understudy for Gisting Evaluation) metrics showed strong performance, with ROUGE-1 at 96.03%, ROUGE-2 at 86.67%, and ROUGE-L at 96.10%. Additionally, the Bilingual Evaluation Understudy (BLEU) score was 63%, reflecting the model's coherence in generating summaries. By enhancing EHR summarization through LLMs, this project supports digital transformation efforts in healthcare, streamlining workflows, and enabling more personalized patient care.