Hospital Triage Optimization: Evaluation of Machine Learning Models for Blood Pressure Estimation to Enhance Emergency Response in Colombia
Abstract
Hospital triage requires rapid and accurate assessment of vital signs, supported by well-trained personnel within a robust emergency medical system. However, in remote regions of Colombia, access to skilled staff and monitoring equipment is limited. This study proposes a machine learning framework to estimate blood pressure using photoplethysmography (PPG) signals, demographic data, and comorbidity information within an automated IoT-based triage system. As no prior machine learning–based solutions have addressed patient health status prediction in isolated Colombian regions, this framework aims to provide a complementary triage system in areas lacking expert support. The system integrates a forearm-worn wearable device with a kiosk to collect data, generating 20 input features encompassing demographic/comorbidity information, summary vital signs, and PPG morphology and variability descriptors. Three regression models - feedforward neural network, XGBoost, and Random Forest - are trained and compared for simultaneous estimation of systolic and diastolic blood pressure. Training uses a public short-record PPG dataset comprising 657 signal segments from 219 subjects with subject-wise cross-validation; external validation is performed on the PhysioNet Pulse Transit Time-PPG database. Tree-based ensemble models outperform the neural network on the main dataset, with XGBoost achieving the best performance for both systolic and diastolic blood pressure. These findings highlight ensemble models as competitive and interpretable alternatives for PPG-based blood pressure estimation, supporting their integration into IoT-enabled triage systems to improve evidence-based patient prioritization, especially in underserved regions.
Keywords: Hospital Triage; Machine Learning; Blood Pressure Estimation; Photoplethysmography (PPG); IoT; XGBoost; Random Forest.
Full Text:
PDFReferences
M. Bednarek-Chałuda, A. Żadło, N. Antosz, and P. Clutter, “Polish Perspective: The Influence of National Emergency Severity Index Training on Triage Practitioners’ Knowledge,” J Emerg Nurs, vol. 50, no. 3, pp. 413–424, 2024.
R. El-Bouri, D. W. Eyre, P. Watkinson, T. Zhu, and D. A. Clifton, “Hospital admission location prediction via deep interpretable networks for the year-round improvement of emergency patient care,” IEEE J Biomed Health Inform, vol. 25, no. 1, pp. 289–300, 2020.
D. Olivia, C. Amrutha, A. Nayak, M. Balachandra, and A. Saxena, “Clinical severity level prediction based optimal medical resource allocation at mass casualty incident,” IEEE Access, vol. 10, pp. 88970–88984, 2022.
N. Gilboy, P. Tanabe, D. Travers, A. M. Rosenau, and others, “Emergency Severity Index (ESI): a triage tool for emergency department care, version 4,” Implementation handbook, vol. 2012, pp. 12–14, 2012.
Colombia. Ministerio de Salud y Protección Social, “Resolución número 5596 de 2015,” 2015, Bogotá. [Online]. Available: https://www.minsalud.gov.co/sites/rid/Lists/BibliotecaDigital/RIDE/DE/DIJ/resolucion-5596-de-2015.pdf
R. C. Wuerz, L. W. Milne, D. R. Eitel, D. Travers, and N. Gilboy, “Reliability and validity of a new five-level triage instrument,” Academic emergency medicine, vol. 7, no. 3, pp. 236–242, 2000.
L. Burgess, K. Kynoch, and S. Hines, “Implementing best practice into the emergency department triage process,” JBI Evid Implement, vol. 17, no. 1, pp. 27–35, 2019.
P. Williams et al., “Efficacy of a triage system to reduce length of hospital stay,” The British Journal of Psychiatry, vol. 204, no. 6, pp. 480–485, 2014.
A. Khorram-Manesh, K. Lennquist Montán, A. Hedelin, M. Kihlgren, and P. Örtenwall, “Prehospital triage, discrepancy in priority-setting between emergency medical dispatch centre and ambulance crews,” European Journal of Trauma and Emergency Surgery, vol. 37, no. 1, pp. 73–78, 2011.
T. Ullah, “Analysis Of Factors That Affect the Implementation Of Triage On Satisfaction Of Patients Family,” Journal of Applied Nursing and Health, vol. 4, no. 1, pp. 140–145, 2022.
I. Kuzmanov, A. M. Bogdanova, M. Kostoska, and N. Ackovska, “Fast cuffless blood pressure classification with ECG and PPG signals using CNN-LSTM models in emergency medicine,” in 2022 45th Jubilee International Convention on Information, Communication and Electronic Technology (MIPRO), 2022, pp. 362–367.
S. Sarkar and S. K. Pahuja, “Current developments in cuff-free non-invasive continuous blood pressure estimation using photoplethysmography,” Biomedical Materials & Devices, vol. 2, no. 2, pp. 743–758, 2024.
M. R. Shaikh and M. Forouzanfar, “Dual-stream CNN-LSTM architecture for cuffless blood pressure estimation from PPG and ECG signals: A PulseDB study,” IEEE Sens J, 2024.
F. Pan, P. He, F. Chen, J. Zhang, H. Wang, and D. Zheng, “A novel deep learning based automatic auscultatory method to measure blood pressure,” Int J Med Inform, vol. 128, pp. 71–78, 2019.
A. S. Alghamdi, K. Polat, A. Alghoson, A. A. Alshdadi, and A. A. Abd El-Latif, “A novel blood pressure estimation method based on the classification of oscillometric waveforms using machine-learning methods,” Applied Acoustics, vol. 164, p. 107279, 2020.
Organización Panamericana de la Salud, “Manual para la implementación de un sistema de triaje para los cuartos de urgencias,” 2011.
L.-H. Wang, K.-K. Sun, C.-X. Xie, M.-H. Fan, P. A. R. Abu, and P.-C. Huang, “Cuffless blood pressure estimation using dual physiological signal and its morphological features,” IEEE Sens J, vol. 23, no. 11, pp. 11956–11967, 2023.
Q. Wan et al., “Toward Real-Time Blood Pressure Monitoring via High-Fidelity Iontronic Tonometric Sensors with High Sensitivity and Large Dynamic Ranges,” Adv Healthc Mater, vol. 12, no. 17, p. 2202461, 2023.
B. Tarifi, A. Fainman, A. Pantanowitz, and D. M. Rubin, “A machine learning approach to the non-invasive estimation of continuous blood pressure using photoplethysmography,” Applied Sciences, vol. 13, no. 6, p. 3955, 2023.
A. Garrett et al., “Simultaneous photoplethysmography and blood flow measurements towards the estimation of blood pressure using speckle contrast optical spectroscopy,” Biomed Opt Express, vol. 14, no. 4, pp. 1594–1607, 2023.
P.-K. Man et al., “Blood pressure measurement: From cuff-based to contactless monitoring,” in Healthcare, 2022, p. 2113.
S. K. Nadar and Y. V Gregory, Hypertension (Oxford Cardiology Library), 3rd ed. OUP Oxford, 2023.
Y. Liang, Z. Chen, G. Liu, and M. Elgendi, “A new, short-recorded photoplethysmogram dataset for blood pressure monitoring in China,” Sci Data, vol. 5, no. 1, pp. 1–7, 2018.
P. Mehrgardt, M. Khushi, S. Poon, and A. Withana, “Pulse Transit Time PPG Dataset,” 2022. [Online]. Available: https://physionet.org/content/pulse-transit-time-ppg/1.1.0/
T. Chen and C. Guestrin, “Xgboost: A scalable tree boosting system,” in Proceedings of the 22nd acm sigkdd international conference on knowledge discovery and data mining, 2016, pp. 785–794.
Refbacks
- There are currently no refbacks.
