[2604.03498] Resource-Conscious Modeling for Next- Day Discharge Prediction Using Clinical Notes

Computer Science > Artificial Intelligence arXiv:2604.03498 (cs) [Submitted on 3 Apr 2026] Title:Resource-Conscious Modeling for Next- Day Discharge Prediction Using Clinical Notes Authors:Ha Na Cho, Sairam Sutari, Alexander Lopez, Hansen Bow, Kai Zheng View a PDF of the paper titled Resource-Conscious Modeling for Next- Day Discharge Prediction Using Clinical Notes, by Ha Na Cho and 4 other authors View PDF Abstract:Timely discharge prediction is essential for optimizing bed turnover and resource allocation in elective spine surgery units. This study evaluates the feasibility of lightweight, fine-tuned large language models (LLMs) and traditional text-based models for predicting next-day discharge using postoperative clinical notes. We compared 13 models, including TF-IDF with XGBoost and LGBM, and compact LLMs (DistilGPT-2, Bio_ClinicalBERT) fine-tuned via LoRA. TF-IDF with LGBM achieved the best balance, with an F1-score of 0.47 for the discharge class, a recall of 0.51, and the highest AUC-ROC (0.80). While LoRA improved recall in DistilGPT2, overall transformer-based and generative models underperformed. These findings suggest interpretable, resource-efficient models may outperform compact LLMs in real-world, imbalanced clinical prediction tasks. Subjects: Artificial Intelligence (cs.AI) Cite as: arXiv:2604.03498 [cs.AI]   (or arXiv:2604.03498v1 [cs.AI] for this version)   https://doi.org/10.48550/arXiv.2604.03498 Focus to learn more arXiv-issued DOI via DataCite (pen...