Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This paper models patient healthcare utilization trajectories using a Transformer Hawkes Process, combining Transformer-based history encoding with Hawkes process dynamics to jointly predict event type and time-to-event. To combat class imbalance common in healthcare data, they introduce an inverse square-root class weighting scheme during training. Experiments on real-world data show improved performance in predicting future care needs and identifying high-risk patients.
Class imbalance in healthcare data can be tamed with a simple inverse square-root weighting scheme, boosting the predictive power of Transformer Hawkes Processes for patient care trajectories.
Patient healthcare utilization consists of irregularly time-stamped events, such as outpatient visits, inpatient admissions, and emergency encounters, forming individualized care trajectories. Modeling these trajectories is crucial for understanding utilization patterns and predicting future care needs, but is challenging due to temporal irregularity and severe class imbalance. In this work, we build on the Transformer Hawkes Process framework to model patient trajectories in continuous time. By combining Transformer-based history encoding with Hawkes process dynamics, the model captures event dependencies and jointly predicts event type and time-to-event. To address extreme imbalance, we introduce an imbalance-aware training strategy using inverse square-root class weighting. This improves sensitivity to rare but clinically important events without altering the data distribution. Experiments on real-world data demonstrate improved performance and provide clinically meaningful insights for identifying high-risk patient populations.