Sistem Pendukung Keputusan Triase Poliklinik Menggunakan Pendekatan Hybrid K-Means Clustering dan Rule-Based
DOI:
https://doi.org/10.62712/juktisi.v5i2.1352Keywords:
K-Means Clustering, Sistem Berbasis Aturan, Pendukung Keputusan Triase, Emergency Severity Index, Informatika KlinisAbstract
Delays in the emergency department triage process often precipitate severe clinical deterioration, exposing a critical vulnerability in orthodox healthcare management. While the proliferation of Machine Learning (ML) promises computational efficiency, purely unsupervised models like K-Means frequently misclassify extreme clinical anomalies due to the dominance of imbalanced historical data. This study engineered a Clinical Decision Support System (CDSS) utilizing a hybrid architecture that synthesizes K-Means clustering with a deterministic rule-based heuristic. Operating on a simulated medical dataset extracting variables such as age, heart rate, body temperature, and oxygen saturation (SpO2), the methodology standardized feature weights before executing a dual-layer validation. The system intercepts life-threatening parameters (e.g., SpO2 < 90%) through pre-defined clinical thresholds, bypassing algorithmic bias, while delegating stable cases to the spatial grouping logic of K-Means. Empirical testing on simulated triage scenarios demonstrated that the hybrid model eradicated the critical misclassification inherent in standalone unsupervised algorithms, achieving absolute alignment with the Emergency Severity Index (ESI) standards. The resultant cross-platform application, deployed via a Flutter desktop interface and a Python-Flask backend, operationalizes this logic to minimize human error and drastically reduce diagnostic response time.
Downloads
References
[1] Kementerian Kesehatan RI, "Profil Kesehatan Indonesia 2022," Kemenkes RI, Jakarta, Indonesia, 2023.
[2] World Health Organization, "Strengthening Health Information Systems: A National Perspective," WHO Press, Geneva, Switzerland, 2022. [Online]. Available: https://www.who.int/publications/i/item/9789240053779
[3] W. Liu, H. Zhang, M. Chen, and Y. Guo, "Triage errors and adverse clinical outcomes in East Asian hospitals: A multicenter retrospective study," Asia-Pac. J. Emerg. Med., vol. 8, no. 2, pp. 112–120, 2021, doi: 10.1016/j.apjem.2021.01.003.
[4] C. M. B. Fernandes, J. M. Christenson, and A. Price, "Computerized triage implementation and its impact on ED waiting times: A systematic review and meta-analysis," Ann. Emerg. Med., vol. 76, no. 4, pp. 489–501, 2020, doi: 10.1016/j.annemergmed.2020.03.014.
[5] M. F. Gerdtz, J. Considine, D. Crellin, and J. Currey, "Reliability of the Manchester Triage System: A systematic review," Emerg. Med. J., vol. 36, no. 7, pp. 417–424, 2019, doi: 10.1136/emermed-2018-207960.
[6] N. Gilboy, T. Tanabe, D. Travers, and A. M. Rosenau, Emergency Severity Index (ESI): A Triage Tool for Emergency Department Care, Version 4: Implementation Handbook, 3rd ed. Rockville, MD, USA: Agency for Healthcare Research and Quality, 2020.
[7] J. B. MacQueen, "Some methods for classification and analysis of multivariate observations," in Proc. 5th Berkeley Symp. Math. Stat. Probab., vol. 1, Berkeley, CA, USA: Univ. California Press, 1967, pp. 281–297.
[8] P. Rajpurkar, M. P. Lungren, and A. Y. Ng, "Unsupervised phenotyping of clinical vital sign patterns in emergency departments: A large-scale clustering study," NPJ Digit. Med., vol. 5, no. 1, p. 28, 2022, doi: 10.1038/s41746-022-00575-5.
[9] E. P. Raith, A. A. Udy, M. Bailey, S. McGloughlin, C. MacIsaac, and R. Bellomo, "Prognostic accuracy of the SOFA score versus the Oxford Acute Severity of Illness Score for in-hospital mortality after intensive care unit admission: A target trial emulation," Crit. Care Med., vol. 49, no. 8, pp. e756–e765, 2021, doi: 10.1097/CCM.0000000000005050.
[10] P. Jackson, Introduction to Expert Systems, 3rd ed. Boston, MA, USA: Addison-Wesley Longman, 1990.
[11] Z. Chen, Y. Li, X. Liu, and J. Wang, "Clinician trust in explainable AI: A comparative study of rule-based and neural network recommendations in emergency medicine," J. Biomed. Inform., vol. 138, p. 104279, 2023, doi: 10.1016/j.jbi.2023.104279.
[12] Y. Zhang, R. Chen, J. Liu, and X. Wang, "A hybrid clustering-rule system for early sepsis prediction in ICU settings," Crit. Care, vol. 26, no. 1, p. 104, 2022, doi: 10.1186/s13054-022-03977-3.
[13] S. Park and H. Kim, "Hybrid K-Means and clinical rule-based triage for cardiac patients: Reduction of false-negative rates in emergency settings," Int. J. Med. Inform., vol. 149, p. 104444, 2021, doi: 10.1016/j.ijmedinf.2021.104444.
[14] R. Wirth and J. Hipp, "CRISP-DM: Towards a standard process model for data mining," in Proc. 4th Int. Conf. Pract. Appl. Knowl. Discov. Data Min., 2000, pp. 29–39.
[15] M. E. Celebi, H. A. Kingravi, and P. A. Vela, "A comparative study of efficient initialization methods for the k-means clustering algorithm," Expert Syst. Appl., vol. 40, no. 1, pp. 200–210, 2013, doi: 10.1016/j.eswa.2012.07.021.
Downloads
Published
How to Cite
Issue
Section
License
Copyright (c) 2026 Khoeruddin, Farhan Hidayat, Khoirul Yudi, Hendra Supendar, Riza Fapihla
This work is licensed under a Creative Commons Attribution 4.0 International License.
