A Smart Patient Monitoring System for Home Healthcare

Abstract

For patients getting care at home, home health- care necessitates ongoing observation and prompt medical assis- tance. In order to facilitate efficient coordination between home nurses, physicians, administrators, and ambulance services, this study suggests an AI-driven home healthcare monitoring system. While doctors document clinical observations and treatment in- formation, home nurses routinely update patients’ vital signs and health condition. To find unusual medical conditions, the gathered data is evaluated. The system creates alerts in an emergency to prompt medical professionals, nurses, and ambulance personnel to act quickly. The suggested approach facilitates effective home- based healthcare monitoring, improves care coordination, and increases patient safety.

Introduction

The increasing demand for home healthcare, driven by aging populations and chronic illnesses, has highlighted the limitations of traditional systems, which rely on manual monitoring and delayed communication. These shortcomings can lead to poor coordination, delayed emergency response, and increased risks to patient safety.

To address these challenges, the proposed system introduces a smart home healthcare monitoring platform that integrates digital technologies, AI, and real-time data analysis. The system connects multiple stakeholders—administrators, doctors, home nurses, and ambulance services—into a unified platform for efficient coordination and continuous patient monitoring.

The methodology combines structured data (vital signs, test results) with unstructured clinical notes using machine learning and large language models (LLMs). It extracts symptoms, evaluates severity, and predicts health risks (such as cardiac issues) using models like Random Forest, Gradient Boosting, and XGBoost.

The system also includes real-time alert mechanisms that detect abnormal conditions and notify relevant parties for quick emergency response. Additional features include image-based monitoring, text-based clinical analysis, and role-based dashboards for efficient data management.

Results show that the system improves early risk detection, enhances communication among healthcare providers, reduces delays in emergencies, and supports better decision-making. Overall, it offers a scalable, efficient, and intelligent solution that improves patient safety, care coordination, and the quality of home-based healthcare services.

Conclusion

In order to facilitate early clinical risk prediction and better patient care in home healthcare settings, this project effectively presents an intelligent, AI-enabled healthcare monitoring system. The system efficiently processes both structured clinical data and unstructured medical narratives to produce significant health insights by combining machine learning, natural language processing, and large language model (LLM) approaches. By automating symptom extraction and condition assessment, the suggested method tackles a major issue in healthcare systems: accurate interpretation of complex clinical notes. The system uses LLM-based feature extraction and sophisticated text preprocessing to extract clinically relevant symptoms, severity levels, and temporal patterns from nar- rative data. A thorough multi-modal representation of the patient’s health is produced by combining these extracted insights with structured patient data, such as vital signs and test results. When compared to conventional systems that only use structured data, this integration improves the precision and dependability of risk prediction. The role-based archi- tecture of the suggested system, which consists of modules for administrators, physicians, home nurses, and ambulance services, is one of its main advantages. This architecture guarantees efficient coordination among healthcare stakeholders, safe access control, and well-organized data administration. Administrators may oversee staff assignments and system op- erations, doctors can record clinical notes and treatment data, and nurses can update patient vitals and progress. This kind of coordinated contact facilitates prompt therapeutic interventions and increases workflow efficiency. The system’s prediction models, which include ensemble and tree-based classifiers, show consistent accuracy in recognizing heart-related risks and dysphagia. The system’s potential to assist decision-making in actual healthcare situations is demonstrated by the evalua- tion findings, which show satisfactory accuracy and F1-score. Furthermore, the incorporation of real-time warning and noti- fication methods facilitates prompt reaction to critical patient circumstances, hence mitigating any delays in the provision of care. All things considered, the suggested approach shows great promise for raising the standard of healthcare services by facilitating the early identification of health hazards, enhancing provider coordination, and assisting with data-driven clinical decisions. Future improvements, including adding more illness prediction models, sophisticated deep learning methods, and wearable device data, have a solid foundation thanks to the modular and scalable design. In conclusion, our study presents a practical and intelligent approach toward enhancing patient safety, efficiency, and outcomes in current healthcare systems.

References

[1] T.-D. Le, R. Noumeir, J. Rambaud, G. Sans, and P. Jouvet, ”Adap- tation of autoencoder for sparsity reduction from clinical notes representation learning”|| IEEE Journal of Translational Engineering in Health and Medicine, vol. 11, pp. 469–478, Feb. 2023, doi: 10.1109/JTEHM.2023.3241635. [2] G. Artioli, L. Deiana, M. Bertuol, G. Casella, M. Guasconi, C. Foa`, R. De Simone, L. Sarli, and A. Bonacaro, ”Evaluating a nurse- led narrative interview intervention with cancer patients with a first diagnosis: A feasibility study”|| Heliyon, vol. 10, no. 11, Article e31802, May 2024, doi:10.1016/j.heliyon.2024.e31802. [3] W. Tang and Y. Gao, ”Influence of standardized patient combined with narrative nursing teaching on experimental teaching of surgical nursing: A quasi-experimental study”|| Int. J. Nurs.Sci., vol. 7, no. 2, pp. 100601, Jun. 2025. doi: 10.1016/j.ijnss.2025.06.011. [4] T. Bandholm, R. S. Husted, A. Troelsen, and K. Thorborg, ”Changing the narrative for exercise-based prehabilitation: Evidence-informed and shared decision making when discussing the need for a total knee arthroplasty with patients”|| Osteoarthritis and Cartilage Open, vol. 7, no. 2, p. 100601, Jun. 2025. doi: 10.1016/j.ocarto.2025.100601 [5] L. Junyu, L. Yuan, W. Jianjian, Z. Zhanchou, and Z. Liping, ”Effective- ness of narrative nursing on depression patients with suicide attempt: A preliminary randomized controlled trial”|| Acta Psychologica, vol. 256, p. 105039, Jun. 2025, doi: 10.1016/j.actpsy.2025.105039. [6] Harerimana, Kim and Jang, ”HSGA:A Hybrid LSTM-CNN Self- Guided Attention to Predict the Future Diagnosis From Dis- charge Narratives”|| IEEE Access, vol.11, pp.70184-70194, 2023, doi:10.1109/ACCESS.2023.3320179. [7] L. D´?az, A. Chaple Gil, A. Von Marttens, J. Basualdo, C. Sotomayor, A. Vera Becerra, V.Beltra´n, G. Jorquera, R. Caviedes, and E. Ferna´ndez, ”The clinical efficacy of intravascular laser irradiation of blood (ILIB): A narrative review of randomized controlled trials,|| Photodiagnosis and Photodynamic Therapy”, vol. 53, p. 104618, Jun. 2025. doi: 10.1016/j.pdpdt.2025.104618.

Copyright

Copyright © 2026 Meghna Manoj, Naaz Abdul Jaleel, Smeya Sajith, Vyshna Shaji, Ms Divya C. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.