Model Context Protocol in EU Telemedicine: AI Integration with Clinical Safety
Model Context Protocol (MCP) allows patients to connect AI assistants like Claude to their Hi-Doctor account using OAuth 2.1. Learn how conversational data entry operates alongside strict EU doctor oversight.
DIRECT ANSWER
An online consultation lets an EU-licensed doctor review your case and, if appropriate, issue a prescription that is valid across the EU.
Less than 4 hours on average
Prescription issued in less than 4 hours on average if approved.
EU-licensed doctors
Every case is reviewed by a licensed clinician.
Valid across the EU
Approved prescriptions can be used at any EU pharmacy.
In this article
HOW IT WORKS
From consultation to collection
- 1
Step 1
Consultation
Answer a secure medical questionnaire online.
- 2
Step 2
Doctor reviews
An EU-licensed doctor personally assesses your case.
- 3
Step 3
Prescription in less than 4 hours on average
If approved, your prescription is issued in less than 4 hours on average.
- 4
Step 4
Collect at any EU pharmacy
Use an approved prescription at a pharmacy anywhere in the EU.
Model Context Protocol (MCP) is an open standard that allows artificial intelligence assistants to connect directly and securely with external patient software accounts. At Hi-Doctor, MCP enables patients to complete consultation questionnaires conversationally or review past consultations, orders, and prescriptions through an assistant like Claude, using OAuth 2.1 authentication. Crucially, the AI assistant acts solely as an interface for data entry and retrieval, while every clinical decision and prescription assessment remains strictly in the hands of an EU-licensed doctor.
Key takeaways
- Model Context Protocol (MCP) connects patient accounts to AI assistants via OAuth 2.1 with PKCE, eliminating shared API keys or exposed credentials.
- An AI assistant can help patients draft questionnaire answers, track weight-loss progress, and inspect past orders, but it cannot prescribe, diagnose, or assess medical eligibility.
- Every clinical assessment is reviewed asynchronously by an EU-licensed doctor who decides whether a treatment and REMPE e-prescription are appropriate.
- If a doctor declines a consultation request, the full consultation fee is automatically refunded.
What is Model Context Protocol and why does it matter for digital healthcare?
For years, interacting with digital healthcare meant filling out static web forms or navigating complex account dashboards. Model Context Protocol, launched as an open standard, changes how software applications share structured context with large language models. Rather than re-entering information across multiple forms or clicking through multiple screens, MCP allows an AI client to interact directly with backend services through a standardized tool architecture.
In telemedicine, this technology translates to conversational accessibility. A patient can complete a detailed questionnaire intake conversationally in their native language, going back to change any answer and reviewing everything before submitting the finalized data to the medical platform.
Hi-Doctor's hosted MCP server exposes 50 granular tools grouped strictly by permission level, ensuring patients maintain total sovereignty over which account data an assistant can view or edit.
How does conversational intake differ from traditional web forms?
Traditional telemedicine platforms rely on rigid, multi-page forms that can feel clinical and repetitive. MCP introduces a structured yet flexible bridge between the patient and the underlying database. The assistant presents medical questions one at a time, allowing the patient to go back and change any answer and review the complete questionnaire summary before submitting.
| Dimension | Traditional Web Questionnaire | MCP Assistant Integration |
|---|---|---|
| Authentication | Standard browser session / password login | OAuth 2.1 authorization-code flow with PKCE |
| Interaction Model | Form fields, radio buttons, static text boxes | Conversational, step-by-step natural language guidance |
| Data Modification | Manual entry field by field | Conversational step-by-step entry with pre-submission review |
| Clinical Gatekeeper | EU-licensed doctor assessment | EU-licensed doctor assessment (identical clinical boundary) |
How does the asynchronous clinical workflow operate?
Whether a patient submits a consultation via the web dashboard or through an MCP-connected assistant, the backend processing and clinical evaluation remain identical. The AI assistant never assesses risk, never calculates eligibility, and never issues medication prescriptions. Once submitted, the complete questionnaire enters the queue for review by an EU-licensed doctor.
- Patient connects assistant via OAuth 2.1
- Patient answers questionnaire conversationally
- Final data submitted to Hi-Doctor backend
- EU-licensed doctor reviews medical intake
- Doctor Clinical Decision
- If the doctor approves — REMPE PDF prescription issued in under 24 hours, and the patient presents it at an EU pharmacy
- If the doctor declines — consultation fee refunded in full, with medical guidance
What are the non-negotiable security and clinical boundaries?
Integrating artificial intelligence into medical platforms requires rigid boundary enforcement. Under Hi-Doctor's technical architecture, clinical decision tools—such as approving consultations, altering dosage instructions, or writing prescriptions—are completely absent from the MCP tool surface. The assistant literally lacks the programmatic capacity to perform clinical actions.
Patients retain complete control through OAuth 2.1 permissions. Ungranted scopes mean corresponding tools are hidden from the assistant entirely. Furthermore, for financial transactions, the system never processes raw card data through the assistant; payment requests return a secure link directly to Stripe's hosted checkout.
Zero clinical tools are granted to AI models: 100% of prescription decisions are made by registered EU physicians operating under European medical regulations.
How Hi-Doctor can help
Hi-Doctor provides 100% online, asynchronous consultations with registered EU doctors across defined care categories: Hair Growth, Weight Loss, Sexual Health, Premature Ejaculation, Chlamydia, Urinary Tract Infections, Emergency Contraception, and HIV PrEP. Patients can complete their medical questionnaires via our traditional web platform or through a connected MCP AI assistant.
Consultations for Hair Growth and Weight Loss are structured as ongoing subscriptions at €25 per review, while all other categories carry a €35 single consultation fee. An EU-licensed doctor evaluates every submission and, when clinically appropriate, issues an official Spanish REMPE e-prescription PDF in under 24 hours. The REMPE e-prescription standard is designed for cross-border recognition across EU member states with all required fields. Outside Spain, a patient presents the REMPE PDF to a pharmacist, who reviews the document and may dispense a locally marketed equivalent with the same active ingredient. Pharmacists retain the right to refuse, so dispensing is never guaranteed. If the reviewing doctor determines that treatment is not medically suitable, no prescription is issued and the consultation fee is refunded in full.
COMMON QUESTIONS
Frequently asked questions
No. An AI assistant cannot diagnose conditions, approve consultations, issue prescriptions, or adjust medication dosages. Its role is strictly limited to helping you complete questionnaires, view account status, log weight-loss progress, and communicate asynchronously with our team. Every medical decision is made solely by an EU-licensed doctor.
