How OrbitLink Built the Patient-Generated Data Pipeline Behind a $10M Moonshot Grant
Built the PGD sync and Open Wearables integration for a UC Berkeley/UCSF open-source clinical research exchange.
Every wearable speaks its own proprietary format. Getting patient data (sleep, heart rate, spirometry, glucose) into a clinical-grade exchange meant a separate one-off integration per device, with no consent flow, no provenance tracking, and no standards compliance.
- Patient-facing sync flow: invite link, consent, OAuth device connection, scheduled sync as FHIR Observations
- Open Wearables integration as a unified device front end (Oura first), with webhook idempotency and sleep-record merge
- Contributions to omh-shim (proprietary JSON to Open mHealth / IEEE 1752.1), including upstream schema fixes
- Role-based access control, initial SAML/SSO, and deployment automation
- Live university clinical-research deployment at UCSF
Research teams and clinicians increasingly want patient-generated data: sleep from an Oura ring, spirometry from a home device, glucose from a CGM. The problem is that every wearable speaks its own proprietary format. Historically that meant a separate one-off integration per device, each with its own auth quirks, data shapes, and failure modes, and none of it standards-compliant or consented in a way clinical research can use.
JupyterHealth, an open-source initiative backed by UC Berkeley, UCSF, The Commons Project, 2i2c, and Duke, is building the fix: the JupyterHealth Exchange (JHE), a FHIR-native exchange where consented patient data lands in one clinical-grade place. OrbitLink's Travis Sauer has been a contributing member of the JupyterHealth software team since March 2025 and is the primary developer and architect of its patient-generated data sync and Open Wearables integration.
The work is the unglamorous plumbing that makes patient data actually usable. OrbitLink built the patient-facing sync flow: a patient receives an invite link, consents to share specific data types, connects their device accounts through OAuth, and their data syncs reliably into JHE as FHIR Observations carrying Open mHealth payloads. OrbitLink integrated the open-source Open Wearables platform as a unified front end for device APIs, handled webhook idempotency and duplicate-record merging, and contributed to omh-shim, the library that converts proprietary device JSON into the Open mHealth / IEEE 1752.1 standard, including fixes to broken schemas upstream. Provenance metadata was added to the data models so every data point records its source and permission. On the platform side, OrbitLink built role-based access control, the initial SAML single sign-on integration, and deployment and CI/CD automation, then supported a live university clinical-research deployment at UCSF.
The outcome: what used to be a separate integration per device is now one standards-based pipeline. When UCSF's respiratory-health study needed patient spirometry data (FVC and FEV1), the schemas were added, validated, and wired as supported scopes, and the study could pull clinical-ready data from the exchange.
The stakes grew in April 2026, when Edge Medicine, the initiative built on JupyterHealth, was named one of eight projects internationally to receive a Laude Moonshot seed grant, now competing for a ten million dollar multi-year Moonshot lab in frontline healthcare. The wearable and patient-generated data pipeline OrbitLink builds is required infrastructure for that demo. Because everything is open source and open standards, any institution can deploy the same stack without vendor lock-in, and the work is publicly verifiable on GitHub.
Contributing member of the JupyterHealth software team since March 17, 2025. Primary developer and architect of the Open Wearables integration and PGD sync. Listed on the JupyterHealth team at jupyterhealth.org (Platform Development, Software).
One standards-based pipeline replaced per-device integrations. UCSF respiratory study pulls patient spirometry data in clinical-ready, validated form. Work runs in a live university clinical-research deployment. JupyterHealth is the platform behind Edge Medicine, one of eight projects awarded a Laude Moonshot seed grant (April 2026), now competing for a $10M multi-year Moonshot lab; the PGD/Open Wearables pipeline is required infrastructure for the September 2026 demo.
Travis is the lead developer on the JupyterHealth Exchange, with well over a hundred merged pull requests across the project. He built core pieces of our native FHIR server and the patient facing data flows, and his work holds up under review. He is one of the few engineers I bring onto the problems that actually have to work in production.
JupyterHealth as a project aims to bridge the gap from data produced by consumer devices to researchers and clinicians. We have been working to improve the robustness of the JupyterHealth Exchange via software security best practices, and developing interfaces to allow more varied pathways for data to be available for analysis from different devices and user situations.
- github.com/jupyterhealth/jupyterhealth-exchange | github.com/jupyterhealth/omh-shim | github.com/travis-sauer-oltech | jupyterhealth.org (team graphic) | docs.openwearables.io | cdss.berkeley.edu/news/edge-medicine-initiative-awarded-laude-moonshots-seed-grant | laude.org/moonshots | whitebrick.com (04/15/2026 news item)