Clinical XLNet: Modeling Sequential Clinical Notes and Predicting Prolonged Mechanical Ventilation
This work addresses the underutilization of clinical notes in predictive modeling for healthcare, specifically targeting mechanical ventilation prediction, though it appears incremental as an adaptation of XLNet to clinical data.
The authors tackled the problem of predicting prolonged mechanical ventilation by developing Clinical XLNet, a new text representation for sequential clinical notes that leverages temporal information, and demonstrated that it consistently outperforms the best baselines in experiments.
Clinical notes contain rich data, which is unexploited in predictive modeling compared to structured data. In this work, we developed a new text representation Clinical XLNet for clinical notes which also leverages the temporal information of the sequence of the notes. We evaluated our models on prolonged mechanical ventilation prediction problem and our experiments demonstrated that Clinical XLNet outperforms the best baselines consistently.