How to build a GLP-1 telehealth platform
A builder's guide to a GLP-1 telehealth platform: compliance, multi-tenant clinics, eligibility checks, licensing by jurisdiction, chart review, pharmacy routing, cold chain and billing.
A GLP-1 telehealth platform is six systems working together: patient intake and eligibility, prescriber licensing by jurisdiction, physician chart review, pharmacy routing, cold-chain tracking and billing, all running per clinic with strict data isolation. Build the compliance and the multi-tenant model first. Adding clinics later is then configuration, not rewrites.
1. Decide the compliance model before the first screen
For US patients that means HIPAA: a business associate agreement with every vendor that touches patient data, encryption, access control and an audit log. For UK patients it adds UK GDPR and a data processing agreement. On our platform these were in the architecture from day zero, on AWS HIPAA infrastructure. Retrofitting them is the most expensive mistake in this category.
2. Make every clinic a tenant
GLP-1 care is delivered by many clinics under one network. Give each clinic its own isolated data, roles and settings on shared infrastructure. Our build uses a multi-tenant PostgreSQL schema with complete tenant data isolation and role-based access control. Test the isolation: a user in one clinic must never be able to read another's patients, and you should prove it with automated checks.
3. Intake and eligibility
Use a structured questionnaire on a web app that can carry a clinic's brand. Check eligibility before a physician sees the case: BMI against the protocol (30 or above, or 27 with a comorbidity) and a screen for contraindications such as MEN2, MTC and pancreatitis. This keeps physician time for cases that can proceed. Store the intake as FHIR R4 so it can be exchanged later.
4. Match the prescriber to the patient's jurisdiction
A prescriber must be licensed where the patient lives: a US state board licence or UK GMC registration. Route each case by the patient's jurisdiction against live provider registries and the local rules for asynchronous care before it reaches a queue. A deterministic match is easier to audit than a manual one.
5. Give physicians a fast, defensible chart review
Asynchronous review works when the physician gets everything on one screen: vitals, biomarkers and a suggested titration step. We designed around a 45-second review target. The decision is signed with a cryptographic signature that follows 21 CFR Part 11, and the whole action goes to an append-only audit log. Let the physician escalate any case to a video visit.
6. Route the prescription to a pharmacy
Signed orders go to a pharmacy over NCPDP SCRIPT EDI, or straight to a 503A compounding partner over REST. Choose the pharmacy on inventory, licensing and proximity rather than a fixed partner, so one slow pharmacy does not stall the network.
7. Track the cold chain
GLP-1 medicine is temperature sensitive. We watch the 36 to 46°F window through the whole shipment with carrier webhooks and in-box loggers, and alert when a package trends toward the limit, not after it breaches.
8. Settle billing at capture
Our split billing settles at the point of payment. That keeps the books of the platform and each clinic in step and makes disputes visible quickly. Take card data through a tokenising provider, with PCI-DSS controls.
9. Automate the glue
Appointment reminders, lab-result notifications and encounter syncing are repetitive and error-prone by hand. We run them in an n8n automation pipeline, so staff time goes to patients.
What it took for us
A dedicated pod of 33 engineers, architects and QA over an initial six-month engagement, then ongoing scaling. The platform now runs 38 clinic tenants (10 operating at scale) and 12,400+ patients. The full story, with architecture diagrams, is in the case study, and the service overview is at telehealth software development. To size your own build, try the MVP cost estimator.
