Elamur Research⌕
☰
home / papers / 2026 / Budget-Aware Routing for Long Clinical Text

Budget-Aware Routing for Long Clinical Text

Paper · 2026 · Khizar Qureshi, Geoffrey Martin and Yifan Peng

Summary

Sending long clinical documents such as discharge notes to large language models is costly. API fees and latency rise with every input token, and hospitals make these calls at very high volume. The authors treat context construction as a separate, training-free module. Given a document and a token budget, it picks a subset of text units that satisfies a knapsack constraint, and only that subset goes to a fixed generator (GPT-4o).

The study varies two things. The first is how text is segmented: sentence, section, window or cluster units. The second is the selector: Lead, shuffled, sliding, hierarchical, GraphCluster, MMR, and a new monotone submodular objective called RCD that combines relevance, facility-location coverage and log-determinant diversity. A lightweight router then chooses a selector based on the budget and document statistics.

Across three test sets (MIMIC discharge notes, abstracts from Cochrane, and L-Eval), no single method comes out ahead in every setting. Lead is strongest under extractive evaluation at small budgets, while MMR helps LLM generation by reducing redundant input. The router comes close to oracle performance, which gives practical guidance for cost-limited clinical systems.

Key points

  • On MIMIC, a 512-token budget (about a quarter of the median note length) gives results on par with feeding the full document.
  • Selector choice affects quality more than unitization. Cluster-based units hurt performance, and the other three segmentation schemes perform about the same.
  • ROUGE saturates when scoring LLM-generated summaries, whereas BERTScore separates methods more clearly.
  • RCD weight sensitivity is smooth rather than brittle, and it becomes more stable as the budget grows. Results on PubMedQA show the framework carries over to classification.
  • Code is released at github.com/stone-technologies/ACL_budget_paper. Stated limitations are a single proprietary generator (GPT-4o) and reliance on automatic metrics only, with no human or clinical-safety evaluation.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

 Institutions

🎓Massachusetts Institute of TechnologyCambridge · United States243 works · 76,506 citations🎓Cornell UniversityIthaca · United States134 works · 65,159 citations🎓Weill Cornell MedicineNew York · United States2 works

 Fields, topics and concepts

📚 Shelves 5The field of science this work belongs to (OpenAlex fields).Computer Science · 18,614 worksArtificial Intelligence · 17,965 worksArts and Humanities · 17,450 worksLanguage and Linguistics · 17,296 worksInformation Systems · 1,730 works

Cite

Cite as Khizar Qureshi, Geoffrey Martin and Yifan Peng. 2026. Budget-Aware Routing for Long Clinical Text. Findings of the Association for Computational Linguistics: ACL 2026.

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/2026.findings-acl.2114

@inproceedings{qureshi2026budget,
  title = {{Budget-Aware Routing for Long Clinical Text}},
  author = {Khizar Qureshi, Geoffrey Martin and Yifan Peng},
  booktitle = {Findings of the Association for Computational Linguistics: ACL 2026},
  year = {2026},
  doi = {10.18653/v1/2026.findings-acl.2114},
  url = {https://doi.org/10.18653/v1/2026.findings-acl.2114}
}

References

24 references, 11 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. American Hospital Association. 2025. Fast facts on U.S. hospitals. https://www.aha.org/statistics/ fast-facts-us-hospitals. Accessed 2025-12-30. ↗
  2. Chen An, Shansan Gong, Kai Zhong, Ruiyi Lang, Shangyue Guo, Zhaoyi Li, Xin Yao, Jie Zhang, Yidong Zhao, Wei Liu, and Xipeng Qiu. 2024. L-Eval: Instituting standardized evaluation for long context language models with real-world tasks. In Proceedings of the 62nd Annual Meeting of the Association for… find
  3. Yushi Bai and 1 others. 2023. LongBench: A bilingual, multitask benchmark for long context understanding. arXiv preprint arXiv:2308.14508.
  4. Jaime Carbonell and Jade Goldstein. 1998. The use of MMR, diversity-based reranking for reordering documents and producing summaries. In Proceedings of the 21st Annual International ACM SIGIR Conference on Research and Development in Information Retrieval, pages 335–336. find
  5. Arman Cohan, Franck Dernoncourt, Dong Suk Kim, Trung Bui, Seokhwan Kim, Walter Chang, and Nazli Goharian. 2018. A discourse-aware attention model for abstractive summarization of long documents. In Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational…
  6. Cheng Hsieh and 1 others. 2024. RULER: Scaling benchmarks for long context language models. arXiv preprint arXiv:2404.06654.
  7. Dong Jiang, Chenghao Liu, Leyang Cui, and 1 others. 2023. LLMLingua: Compressing prompts for accelerated inference of large language models. arXiv preprint arXiv:2310.05736.
  8. Qiao Jin, Bhuwan Dhingra, Zhengping Liu, William Cohen, and Xinghua Lu. 2019. Pubmedqa: A dataset for biomedical research question answering. In Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Proce…
  9. Alistair E W Johnson, Tom J Pollard, Lu Shen, Liwei H Lehman, Mengling Feng, Mohammad Ghassemi, Benjamin Moody, Peter Szolovits, Leo Anthony Celi, and Roger G Mark. 2023. MIMIC-IV, a freely accessible electronic health record dataset. Scientific Data, 10(1):1–13.
  10. George L Nemhauser, Laurence A Wolsey, and Marshall L Fisher. 1978. An analysis of approximations for maximizing submodular set functions. Mathematical Programming, 14(1):265–294. find
  11. OpenAI. 2025. Openai API pricing. https://openai.com/api/pricing/. Accessed 2025-12-30. ↗
  12. S. Trent Rosenbloom, William W. Stead, Joshua C. Denny, Dario Giuse, Nancy M. Lorenzi, Steven H. Brown, and Kevin B. Johnson. 2010. Generating clinical notes for electronic health record systems. Applied Clinical Informatics, 1(3):232–243. find
  13. Abigail See, Peter J. Liu, and Christopher D. Manning. 2017. Get to the point: Summarization with pointer-generator networks. In Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 1073– 1083. find
  14. Eva Sharma, Chen Li, and Lu Wang. 2019. Bigpatent: A large-scale dataset for abstractive and coherent summarization. In Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, pages 2204–2213.
  15. Maxim Sviridenko. 2004. A note on maximizing a submodular set function subject to a knapsack constraint. In Proceedings of the 13th Annual ACM-SIAM Symposium on Discrete Algorithms, pages 1020–1021. find
  16. Greg Kamradt. 2023. Needle in a haystack evaluations for long context models. https://github.com/ gkamradt/LLMTest_NeedleInAHaystack. Accessed 2025-12-30. ↗
  17. Tianyi Zhang, Varsha Kishore, Felix Wu, Kilian Q Weinberger, and Yoav Artzi. 2020. Bertscore: Evaluating text generation with bert. In International Conference on Learning Representations. find
  18. Andreas Krause and Daniel Golovin. 2014. Submodular function maximization. Tractability, pages 71–104. find
  19. Yutong Zhang and 1 others. 2024. Benchmarking large language models for medical evidence summarization. Nature Machine Intelligence, 6. find
  20. Alex Kulesza and Ben Taskar. 2012. Determinantal point processes for machine learning. Foundations and Trends in Machine Learning, 5(2–3):123–286. find
  21. Patrick Lewis, Ethan Perez, Aleksandra Piktus, Fabio Petroni, Vladimir Karpukhin, Naman Goyal, Deniz Kucuk, Mike Lewis, Wen-tau Yih, and Sebastian Riedel. 2020. Retrieval-augmented generation for knowledge-intensive NLP tasks. Advances in Neural Information Processing Systems, 33:9459–9474.
  22. Chin-Yew Lin. 2004. Rouge: A package for automatic evaluation of summaries. Text Summarization Branches Out.
  23. Hui Lin and Jeff Bilmes. 2011. A class of submodular functions for document summarization. Proceedings of the 49th Annual Meeting of the Association for Computational Linguistics.
  24. Yang Liu and Mirella Lapata. 2019. Text summarization with pretrained encoders. In Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP), pages 3730–3740.

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/qureshi-2026-budget-aware-routing-long-clinical/