Scalable Natural Language Processing to Transform Healthcare

Electronic health records (EHRs) have immense potential to transform medicine both at the point-of-care and through retrospective research. However, current EHRs are burdensome to use, resulting in documentation which is suboptimal for patients, clinicians, and researchers alike. Clinical notes are written in their own jargon-heavy dialect, patient histories can contain hundreds of notes, and there is often minimal labeled data available. In this talk, I will first discuss scalable techniques for clinical information extraction that leverage large language models, how these methods could improve equity in healthcare, and how we might think about the data training these models. Next, I will describe a paradigm for smarter electronic health records that decreases documentation burden, incentivizes the creation of high-quality data at the point-of-care, and aids in information retrieval. I will also briefly describe work exploring clinical text summarization and medical text simplification. I will end with open challenges and opportunities for NLP to impact a variety of healthcare workflows, with a lens towards human-AI interaction, evaluation, and deployability.