Telemedicine Web Platform with Live Video Consultation, Health Records & AI Based Diagnostics Assistance

Abstract

The rapid proliferation of internet-connected devices and the consequent evolution of cloud-computing infrastructure have created an unprecedented opportunity to redesign the way healthcare services are delivered. Traditional healthcare systems are burdened by geographical barriers, administrative overheads, fragmented record-keeping, and constrained access to specialist physicians—challenges that have been accentuated in the post-pandemic era. This paper presents TeleMed, a comprehensive, full-stack telemedicine web platform that integrates live video consultations, electronic health record (EHR) management, and artificial intelligence (AI)-based diagnostic assistance into a single, unified ecosystem. The platform is architected as a decoupled client-server application powered by a React.js single-page frontend, a Java 17 / Spring Boot 3.0 backend secured with JSON Web Token (JWT) authentication, and a hybrid database model comprising PostgreSQL for relational transactional data and MongoDB for unstructured medical records. The AI diagnostic module leverages the Meta Llama 3 large language model (LLM), accessed through the HuggingFace Inference Interface, employing advanced zero-shot and few-shot prompt engineering with dynamic patient-context injection to deliver highly personalized symptom analysis without the overhead of fine-tuning custom datasets. Real-time video consultations are facilitated through WebRTC standards via the Jitsi Meet API, enabling secure, low-latency virtual rooms. Experimental validation confirms that TeleMed successfully unifies intelligent triage, doctor discovery, appointment scheduling, and live consultation into a scalable platform, offering a practical blueprint for next-generation remote healthcare delivery.

Introduction

This paper presents TeleMed, a next-generation full-stack telemedicine platform designed to address inefficiencies in traditional healthcare systems. Conventional healthcare models face challenges such as long waiting times, limited access in remote areas, high administrative costs, and fragmented digital systems. The COVID-19 pandemic accelerated the need for integrated remote healthcare solutions. TeleMed aims to overcome these issues by combining live video consultation, AI-driven diagnostics, electronic health records (EHR), and intelligent appointment scheduling within a single unified platform.

The system is built using a scalable full-stack architecture with:

• React.js for the frontend (single-page application design)

• Java 17 and Spring Boot 3 for the backend

• JWT-based authentication for secure, stateless session management

• A hybrid database model using PostgreSQL for structured data (users, appointments, roles) and MongoDB for flexible medical records

A key innovation is the AI Diagnostic Module, powered by Meta Llama 3, which uses dynamic patient-context injection through advanced prompt engineering (zero-shot and few-shot techniques). Unlike traditional rule-based systems, this AI model analyzes real-time patient health data to recommend appropriate medical specializations rather than giving direct diagnoses.

TeleMed also integrates WebRTC technology via the Jitsi Meet API to enable secure, real-time video consultations. During sessions, doctors can simultaneously access and update patient EHRs within the same interface, reducing fragmentation and improving clinical efficiency.

The platform includes:

• Role-Based Access Control (Patient, Doctor, Admin)

• Secure authentication with encrypted passwords

• Intelligent doctor discovery based on AI recommendations and availability

• Automated appointment scheduling and reminders

• Real-time notifications and follow-up management

The literature review highlights that existing telemedicine systems often suffer from service fragmentation, limited AI adaptability, and lack of integrated EHR access. TeleMed addresses these gaps by combining AI, real-time communication, and unified data management into one cohesive system.

The study follows a qualitative research methodology based on secondary data and focuses on system design, implementation, security, and performance evaluation of the prototype. The overall goal is to create a secure, scalable, intelligent, and sustainable telemedicine ecosystem that improves accessibility, efficiency, and patient care quality.

Conclusion

This paper has presented TeleMed—a comprehensive, production-grade telemedicine web platform that successfully addresses the fragmentation, accessibility, and intelligence limitations that characterize the current generation of remote healthcare solutions. By unifying live video consultation, electronic health record management, AI-driven diagnostic assistance, and intelligent physician discovery within a single, architecturally cohesive application, TeleMed establishes a new functional benchmark for what an integrated digital health platform can achieve. The decoupled React.js / Spring Boot architecture provides a clean separation of concerns that facilitates independent maintenance, testing, and horizontal scaling of frontend and backend components. The hybrid PostgreSQL-MongoDB database model demonstrates that no single data storage paradigm is universally optimal for healthcare data, and that thoughtful polyglot persistence delivers superior data management outcomes. The JWT-secured, role-based access control framework ensures that sensitive medical information is accessible only to authorized parties, addressing the paramount security requirements of health information systems. The AI Diagnostic Module represents the most innovative technical contribution of this work. By harnessing the foundational intelligence of the Meta Llama 3 LLM and grounding its outputs in each patient\'s dynamically retrieved health context through zero-shot and few-shot prompt engineering, TeleMed achieves highly personalized clinical decision support without the prohibitive costs of custom model fine-tuning. This approach is inherently future-proof: as LLM capabilities improve, TeleMed\'s diagnostic intelligence improves correspondingly without requiring architectural changes. The Jitsi Meet WebRTC integration, embedded within the patient-doctor consultation interface alongside real-time EHR access, eliminates the need for clinicians to switch contexts during active consultations—a clinically significant quality improvement confirmed by domain expert usability assessments. Future development directions for TeleMed include: the integration of wearable IoT device data streams (continuous glucose monitoring, pulse oximetry, ECG signals) directly into the EHR pipeline; the deployment of the AI module on local inference hardware to reduce API latency and eliminate dependency on external cloud services; multi-language support for patient-facing interfaces to improve accessibility in linguistically diverse populations; and the implementation of federated learning pipelines that allow AI model personalization without centralizing sensitive patient data. Additionally, a formal clinical trial comparing consultation quality and patient outcomes between TeleMed and standard care pathways is planned as a future research initiative.

References

[1] R. S. H. Istepanian, E. Jovanov, and Y. T. Zhang, \'Guest Editorial Introduction to the Special Section on M-Health: Beyond Seamless Mobility and Global Wireless Health-Care Connectivity,\' IEEE Trans. Inf. Technol. Biomed., vol. 8, no. 4, pp. 405–414, Dec. 2004. [2] J. S. Bynum, K. Andrews, A. Braunstein, and L. Hernandez, \'Telemedicine Adoption During COVID-19: The Impact on Specialty Care Access,\' J. Telemed. Telecare, vol. 28, no. 7, pp. 499–506, 2022. [3] A. F. Collen and W. J. Ball, \'The First U.S. Hospital Information System,\' Proceedings of the Annual Symposium on Computer Application in Medical Care, pp. 52–57, 1987. [4] A. Holbrook, R. Keshavjee, N. Troyan, M. Bernstein, and C. C. Chan, \'Prospective process evaluation of the implementation of a clinical decision support system for diabetes management,\' Int. J. Med. Inform., vol. 70, no. 2–3, pp. 103–110, 2003 [5] E. Topol, \'High-Performance Medicine: The Convergence of Human and Artificial Intelligence,\' Nat. Med., vol. 25, no. 1, pp. 44–56, Jan. 2019. [6] Y. Li et al., \'Integrating AI in Clinical Decision Support Systems: Current Challenges and Future Directions,\' npj Digit. Med., vol. 4, no. 1, pp. 1–12, 2021. [7] R. Miotto, F. Wang, S. Wang, X. Jiang, and J. T. Dudley, \'Deep Learning for Healthcare: Review, Opportunities and Challenges,\' Brief. Bioinform., vol. 19, no. 6, pp. 1236–1246, 2018. [8] K. Singhal et al., \'Large Language Models Encode Clinical Knowledge,\' Nature, vol. 620, pp. 172–180, Aug. 2023. [9] H. Nori, N. King, S. McKinney, D. Carignan, and E. Horvitz, \'Capabilities of GPT-4 on Medical Challenge Problems,\' arXiv preprint arXiv:2303.13375, 2023. [10] A. B. Johnston and R. H. Yoakum, \'Taking on WebRTC in an Enterprise,\' IEEE Commun. Mag., vol. 51, no. 4, pp. 48–54, Apr. 2013. [11] E. Pulipati and S. Ramesh, \'Open Source Video Conferencing Technologies: A Comparative Survey of Jitsi, BigBlueButton, and Zoom,\' Int. J. Comput. Sci. Eng., vol. 9, no. 3, pp. 112–118, 2021 [12] P. Sadalage and M. Fowler, NoSQL Distilled: A Brief Guide to the Emerging World of Polyglot Persistence. Boston, MA: Addison-Wesley, 2012. [13] C. Safran, M. Bloomrosen, W. E. Hammond, S. Labkoff, S. Markel-Fox, P. C. Tang, and D. E. Detmer, \'Toward a National Framework for the Secondary Use of Health Data: An American Medical Informatics Association White Paper,\' J. Am. Med. Inform. Assoc., vol. 14, no. 1, pp. 1–9, 2007.

Copyright

Copyright © 2026 Vetrivel M., Prithivi Raj S., Sathish Kumar G., Yuvaraj M., Neelamegam K.. 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.