mHealth v2 Platform architecture

Stanford's mHealth platform resides on Google Cloud Platform (GCP).

The new mHealth platform (v2) is designed with a serverless architecture and relies on a combination of third-party services, client-side logic, and service-hosted remote procedure calls (Functions as a Service). This serverless platform eliminates the need for server software and hardware management, making operational management easier with reduced operational cost and scaling costs. Client mobile applications use Firebase SDK libraries to communicate with backend services which are under GCP VPC service controls.

For study participants, they are first authenticated with password-less login powered by Cloud Identity. In the password-less login process, patients go through onboarding & consent in the app, then input their email address. After that, they check their email to finish login. In this way, patients can sign in securely without the need to provide (or remember) their password. They then are able to send survey data from the mobile apps to Firestore and save sensor data or videos to cloud storage. Data change events from the FireStore will trigger the Cloud Functions to write audit logs. Eventually, the data collected from mobile apps can be used by Nero or AutoML for data analysis. Firestore security rules are set up to make sure that the user can only access his/her own data.

For researchers, they are authenticated via Identity Platform to login to the dashboard web app. They then can view the survey data under their own study. Firestore security rules are set up to ensure that they won’t see data from other studies.

The mHealth v2 platform is a big addition to the existing v1 platform, which has already been used by a few studies. The existing platform is hosted in the GKE cluster, providing APIs for registering participants, saving mobile app data and data processing for researchers.