Quantifying surprise in clinical care: Detecting highly informative events in electronic health records with foundation models
This work addresses the challenge of detecting critical events in clinical care for healthcare providers, but it appears incremental as it builds on existing foundation models.
The paper tackles the problem of identifying highly informative events in electronic health records using a foundation model-derived method, which flags anomalous events that rule-based approaches miss and shows these events are significant for predicting downstream patient outcomes, with a fraction of events safely droppable.
We present a foundation model-derived method to identify highly informative tokens and events in electronic health records. Our approach considers incoming data in the entire context of a patient's hospitalization and so can flag anomalous events that rule-based approaches would consider within a normal range. We demonstrate that the events our model flags are significant for predicting downstream patient outcomes and that a fraction of events identified as carrying little information can safely be dropped. Additionally, we show how informativeness can help interpret the predictions of prognostic models trained on foundation model-derived representations.