Language Models for Longitudinal Clinical Prediction
This work addresses the challenge of early-stage Alzheimer's monitoring for clinicians, but it is incremental as it applies an existing method to a new domain with minimal training data.
The paper tackled the problem of predicting clinical outcomes from longitudinal patient data by adapting frozen large language models without fine-tuning, achieving accurate and reliable performance in neuropsychological assessments for early-stage Alzheimer's monitoring.
We explore a lightweight framework that adapts frozen large language models to analyze longitudinal clinical data. The approach integrates patient history and context within the language model space to generate accurate forecasts without model fine-tuning. Applied to neuropsychological assessments, it achieves accurate and reliable performance even with minimal training data, showing promise for early-stage Alzheimer's monitoring.