Patient Journey Trajectories for Disease Progression Prediction and Sub Typing using LLMs
While traditional LLMs considers only diagnoses and patient age, we extend the feature space to several multimodal records, namely demographics, clinical characteristics, vital signs, smoking status, diagnoses, procedures, medications, and laboratory tests, by applying a novel method to unify the frequencies and temporal dimensions of the different features.
We show that additional features significantly improve model performance for various downstream tasks such disease progression prediction and subtyping in different diseases. To ensure robustness, we interpret model predictions using an adaptation of expected gradients, which has not been previously applied to transformers with EHR data and provides more granular interpretations than previous approaches such as feature and token importances.
Furthermore, by clustering the model representations of oncology patients, we show that the model has an implicit understanding of the disease and is able to classify patients with the same cancer type into different risk groups. Given the additional features and interpretability, our LLMs can help make informed decisions about disease trajectories, diagnoses, and risk factors of various diseases.
Thirupathi Pattipaka
Director, DataScience & AI at Novartis Pharma AG
Thiru leads a portfolio of Data Engineering products, Machine Learning products, generative AI and large language models (LLM) to reduce R&D costs, improve operational efficiency, and optimize patient outcomes that enhance clinical research, personalized medicine, patient access, and HCP segmentation. With over 10 years of experience in healthcare data science, he has expertise in building scalable solutions to solve a variety of healthcare AI applications across the drug development process with different databases like clinical operations, claims, EHR, genomics, imaging, and multi-modal data.
