Transformer-Based Health Record Summarisation with Auditable Source Attribution: A Resilience-Oriented Study Protocol
DOI:
https://doi.org/10.66687/xs4f5y45Keywords:
Artificial Intelligence, Machine Learning, Medical Informatics, Decision Support Systems, Clinical, Reproducibility of ResultsAbstract
Clinical summarisation systems can sound convincing while omitting decisive events or attributing facts to the wrong encounter. This gap has consequences for interpreting performance outside carefully curated evaluation settings. The planned investigation seeks to measure whether source-linked transformer summaries reduce unverifiable statements without sacrificing important clinical detail. Relevant methodological and data-quality literature informs the design of the sampling and validation strategy [3–7]. The protocol specifies a blinded assessment of de-identified clinical narratives paired with evidence passages and structured omission/error rubrics. Comparators will include ordinary abstractive summaries, retrieval-supported summaries and provenance-constrained summaries. The technical pipeline must remain auditable: prior work on storage operations provides system-level context [1], while a second source identifies infrastructure conditions relevant to data continuity [2], rather than patient-level outcomes. The resilience component introduces missing data, unexpected interruptions and recovery dependencies as controlled disturbances. Statistical and computational sources help define the reproducibility checks and comparison procedures [8–12]. The primary outcomes are claim-level factual accuracy, critical omission rate, evidence traceability and clinician review time. Potential biases include ambiguous documentation, conflicting observations, retrieval gaps and variations in note length. Outcomes would be assessed before and after each disturbance, including time needed to restore valid operation. Additional research informs the prespecified stress tests and limits on interpretation [13–17]. The intended contribution is an auditable research protocol rather than an unsupported claim of effectiveness; related work frames its transferability limits [18–23].
References
[1] Mohammed Nazir. (2023). Performance-Telemetry-Based Anomaly Detection for Early Failure Identification in Enterprise Storage Systems. Journal of Multidisciplinary Research and Integrated Sciences, 3(1), 59–70. https://jmris.com/index.php/jmris/article/view/2023V1P6
[2] Mohammed Nazir. (2021). RTO and RPO Optimization Using Asynchronous Replication in Multi-Petabyte Enterprise Storage Environments. Journal of Computing, Intelligence and Information Sciences, 2(1). https://jciis.com/index.php/jciis/article/view/2020-01-04
[3] Azam, M. A., Ansari, I., Haque, G. M. M., & Jahid, A. (2026). Leveraging Health Information Systems and Predictive Analytics to Improve Patient Outcomes: A Data-Driven Approach. The American Journal of Medical Sciences and Pharmaceutical Research, 8(3), 45–70. https://theamericanjournals.com/index.php/tajmspr/article/view/7512
[4] Esteva, A., Kuprel, B., Novoa, R. A., et al. (2017). Dermatologist-level classification of skin cancer with deep neural networks. Nature, 542, 115–118. https://www.nature.com/articles/nature21056
[5] Johnson, A., Pollard, T., & Mark, R. (2016). MIMIC-III Clinical Database, version 1.4. PhysioNet. https://physionet.org/content/mimiciii/1.4/
[6] McKinney, S. M., Sieniek, M., Godbole, V., et al. (2020). International evaluation of an AI system for breast cancer screening. Nature, 577, 89–94. https://www.nature.com/articles/s41586-019-1799-6
[7] Begum, M. H., Talukder, S. I., Ahmed, M. J., Khan, M. F. I., & Azam, M. A. (2025). XAI-Driven Multimodal Deep Learning for Early Sepsis Prediction in ICU. Journal of Science Technology and Social Transformation, 1(1), 30–41. https://jstst.info/j/article/view/25
[8] Nahid, M. A. U., Azam, M. A., Ratul, A. S., et al. (2026). Machine Learning Meets XAI: Grad-CAM Visualization for Enhanced Lung and Colon Cancer Detection. Proceedings of the IEEE 2nd International Conference on Quantum Photonics, Artificial Intelligence & Networking. https://ieeexplore.ieee.org/document/11546191/
[9] Jumper, J., Evans, R., Pritzel, A., et al. (2021). Highly accurate protein structure prediction with AlphaFold. Nature, 596, 583–589. https://www.nature.com/articles/s41586-021-03819-2
[10] Liu, X., Faes, L., Kale, A. U., et al. (2020). Reporting guidelines for clinical trial reports for interventions involving artificial intelligence: the CONSORT-AI extension. Nature Medicine, 26, 1364–1374. https://www.nature.com/articles/s41591-020-1034-x
[11] Singhal, K., Azizi, S., Tu, T., et al. (2023). Large language models encode clinical knowledge. Nature, 620, 172–180. https://www.nature.com/articles/s41586-023-06291-2
[12] Alagappan, R. et al. (2019). Protocol-Aware Recovery for Consensus-Based Storage. USENIX ATC '19. https://www.usenix.org/conference/atc19/presentation/alagappan
[13] Lundberg, S. M., & Lee, S.-I. (2017). A Unified Approach to Interpreting Model Predictions. Advances in Neural Information Processing Systems, 30. https://arxiv.org/abs/1705.07874
[14] Vaswani, A., Shazeer, N., Parmar, N., et al. (2017). Attention Is All You Need. Advances in Neural Information Processing Systems, 30. https://arxiv.org/abs/1706.03762
[15] Li, Q. et al. (2023). More Than Capacity: Performance-oriented Evolution of Pangu in Alibaba. USENIX FAST '23. https://www.usenix.org/conference/fast23/presentation/li-qiang-deployed
[16] Maneas, S. et al. (2022). Operational Characteristics of SSDs in Enterprise Storage Systems: A Large-Scale Field Study. USENIX FAST '22. https://www.usenix.org/conference/fast22/presentation/maneas
[17] Zhou, W. et al. (2025). 3L-Cache: Low Overhead and Precise Learning-based Eviction Policy for Caches. USENIX FAST '25. https://www.usenix.org/conference/fast25/presentation/zhou-wenbin
[18] Guo, C. et al. (2017). On Calibration of Modern Neural Networks. Proceedings of Machine Learning Research, 70, 1321–1330. https://proceedings.mlr.press/v70/guo17a.html
[19] Jiao, B., Goel, A. & Wang, A.-I.A. (2025). Silhouette: Leveraging Consistency Mechanisms to Detect Bugs in Persistent Memory-Based File Systems. USENIX FAST '25. https://www.usenix.org/conference/fast25/presentation/jiao
[20] Xu, J. et al. (2021). Anomaly Transformer: Time Series Anomaly Detection with Association Discrepancy. arXiv:2110.02642. https://arxiv.org/abs/2110.02642
[21] Khan, M. F. I., Begum, M. H., Rahman, M. A., Limon, G. Q., Azam, M. A., & Masum, A. K. M. (2025). A comprehensive review of advances in transformer, GAN, and attention mechanisms: Their role in multimodal learning and applications across NLP. International Journal of Science and Research Archive, 15(1), 454–459. https://ijsra.net/content/comprehensive-review-advances-transformer-gan-and-attention-mechanisms-their-role
[22] Rahman, M. A., Bhuiyan, M. T. A., Azam, M. A., et al. (2026). TransFuse: An Interpretable Transformer Ensemble Framework for Bangla Smishing Detection. Proceedings of the IEEE 2nd International Conference on Quantum Photonics, Artificial Intelligence & Networking. https://ieeexplore.ieee.org/document/11546480/
[23] Bhujel, K., Haque, G. M. M., Ansari, I., Azam, M. A., & Jahid, A. (2026). Business Analytics Maturity and Competitive Advantage: Evidence from Data-Driven Enterprises. The American Journal of Management and Economics Innovations, 8(3), 1–23. https://theamericanjournals.com/index.php/tajmei/article/view/7579
Published
Issue
Section
License
Copyright (c) 2026 Arpit Shrimankar
This work is licensed under a Creative Commons Attribution 4.0 International License.
Authors retain copyright in their published work. Unless an individual Version of Record identifies different terms, JMRIS articles are distributed under the Creative Commons Attribution 4.0 International License (CC BY 4.0): https://creativecommons.org/licenses/by/4.0/. This permits sharing and adaptation, including commercial use, with appropriate attribution, a link to the license, and indication of changes. Third-party material may be subject to separate terms. Existing articles retain the license originally granted to their Version of Record; a policy update does not retrospectively replace that license.
