Virtual healthcare bot (VHC-Bot): a Person-centered AI chatbot for transforming patient care and healthcare workforce dynamics

Sara A. Alsalamah, Shada Alsalamah, Hessah A. Alsalamah, Haytham A. Sheerah, Kurt Luther, Chang-Tien LuNetw. Model. Anal. Health Informatics Bioinform 2025 · DOI

Introduction

The healthcare industry is currently navigating a dual crisis: a projected global shortage of nearly 124,000 physicians in the US alone by 2034, and an increasing reliance on impersonal, rules-based virtual healthcare (VHC) platforms. While Artificial Intelligence (AI) offers a scalable solution, it introduces the risk of "deskilling"—the erosion of clinical judgment as practitioners over-rely on automated tools. The Virtual Healthcare Bot (VHC-Bot) is proposed as a hybrid AI-human platform designed to scale delivery while preserving clinician expertise through a "human-in-the-loop" (HITL) architecture.

System Architecture

VHC-Bot utilizes a robust three-tier architecture to ensure scalability and security. The Presentation Tier uses ReactJS for an intuitive interface; the Application Tier leverages PHP Laravel to integrate with the OpenAI API for natural language processing; and the Data Tier employs MySQL for secure storage of medical records and chat logs.

High-level system architecture of the VHC-Bot platform.

As shown in the architecture diagram, the system bridges the gap between patient queries and professional medical systems, routing information through an AI-driven chatbot that interacts with health information, emergency assistance, and medication management systems.

Operational Workflow and Human-in-the-Loop Integration

A core innovation of VHC-Bot is its commitment to Person-Centered Care (PCC) and the mitigation of deskilling. Unlike fully autonomous systems, VHC-Bot functions as a "Clinical Assistant." It performs preliminary triage—gathering history and distinguishing symptoms—but requires clinicians to review and validate AI-generated summaries.

Detailed workflow of the VHC-Bot showing the interaction between the patient interface, the AI engine, and the clinician-validated feedback loop.

The workflow includes a "Data Enrichment" phase where the system ingests authoritative sources like The Gale Encyclopedia of Medicine and learns from clinician edits. This ensures that the AI remains a supportive tool that keeps the human practitioner engaged in the decision-making process, effectively using the AI as a teaching tool or a second opinion rather than a crutch.

Results and Evaluation

A pilot study involving 7 healthcare professionals and 18 patients (ages 25–60) was conducted to evaluate the prototype against standard technical KPIs and qualitative surveys.

  • Technical Performance: The system achieved 100% uptime with a 0% error rate during testing. Latency was recorded between 300–500ms, which remained well within the range for real-time conversation. Stress testing showed the cloud infrastructure could handle 1,000 requests per minute.
  • Clinical Accuracy: 71.4% of clinicians rated the AI's preliminary diagnoses as "generally accurate," while 94.4% of patients felt their queries were addressed correctly.
  • User Satisfaction: Usability was rated as "Excellent" or "Good" by 100% of providers, with 85.8% noting the tool's importance in managing their current workloads.

Conclusion and Future Directions

While the initial results are promising, the study's small sample size necessitates larger, more diverse clinical trials to ensure generalizability. Future work will focus on developing offline capabilities for low-resource or rural environments and implementing localized data models to account for regional health nuances and language differences. By positioning AI as a collaborative partner rather than a replacement, VHC-Bot demonstrates a path forward for sustainable, high-quality virtual healthcare that empowers both patients and providers.

Summary updated

BibTeX
@article{alsalamah2025virtual,
  author = {Sara A. Alsalamah and Shada Alsalamah and Hessah A. Alsalamah and Haytham A. Sheerah and Kurt Luther and Chang-Tien Lu},
  title = {{Virtual healthcare bot (VHC-Bot): a Person-centered AI chatbot for transforming patient care and healthcare workforce dynamics}},
  journal = {Netw. Model. Anal. Health Informatics Bioinform.},
  year = {2025},
  doi = {10.1007/s13721-025-00537-x},
}
Figure