Webhooks
Webhooks are HTTP callbacks a platform sends when something happens — in clinical infrastructure, the event stream (encounter completed, prescription signed, order shipped) that lets a product react to care in real time.
- Event-driven callbacks replacing polling
- Care events: encounter completed, Rx signed, order shipped, delivered
- Consumers must verify signatures and process idempotently
- Periodic API reconciliation catches anything missed
Why care runs on events
Clinical work is asynchronous by nature: a clinician signs a chart minutes or hours after intake, a pharmacy ships the next morning, a delivery lands three days later. Polling for those state changes is wasteful and slow; webhooks invert it — the platform pushes an event the moment state changes, and your product updates the patient, triggers the next step, or alerts your team. A typical care loop emits events like encounter.completed, prescription.signed, order.shipped, and order.delivered.
Building webhook consumers that survive production
The engineering rules are standard but non-optional in healthcare: verify signatures so only the platform can call you, process idempotently because retries mean duplicate deliveries, respond fast and do real work async, and reconcile periodically against the API for anything missed. The reward is a patient experience that feels alive — status that updates the moment the clinician decides, not the next time a cron job runs.
Compliance handled, so you can build
Lithos runs the clinicians, pharmacies, and 50-state rules behind your care program — one API.
Frequently asked questions
What clinical events matter most?
The decision and fulfillment moments: encounter completed (the clinician’s verdict), prescription signed, order shipped, and delivery — each drives patient communication.
What if my endpoint is down when an event fires?
Platforms retry with backoff, which is why idempotent processing matters — you will eventually see the event, possibly more than once.
Webhooks or polling?
Webhooks for reactivity, plus periodic reconciliation via the API for completeness — production systems use both.