CMS-0057-F System Architecture.

Deployment models, core components, integration patterns, and security considerations for solution architects.

Looking for the compliance suite? CMS-0057-F Compliance Suite · Want to see X12 278 to FHIR PAS conversion? PAS tool overview

Deployment Models

On-Premises

  • Your data center
  • Full infrastructure control
  • Air-gapped deployment assessed per customer

Private Cloud

  • AWS / Azure / GCP VPC
  • Your cloud account
  • Isolation and compute tenancy specified for the chosen deployment

Containerized

  • Docker image available for the conversion services (the PAS workflow service is not packaged)
  • Kubernetes deployment assessed per customer
  • Helm charts not currently provided

Core Components

Components of the current demo suite. The PAS workflow service is not packaged; production components are scoped per customer.

Redix PAS API (demo)

FHIR R4 PAS operations ($submit, $inquire) with simulated decisions; not a general-purpose FHIR server

Conversion Engine

AnyToAny Engine (X12 278 to FHIR PAS)

Storage Layer

SQLite demo storage; other production storage integrations require separate design and implementation

Auth Layer

Authentication method, identity provider and operating responsibilities are agreed for the deployment

Integration Patterns

Reference architecture, not a packaged product. The demo shows X12 278 to FHIR PAS conversion, FHIR PAS to X12 278 request and response conversion for online evaluation, and simulated PAS steps on sample data. A FHIR front end over an X12 278 backend needs both directions plus the gateway, authorization and workflow around them, which are scoped per customer.

Provider
EHR / PM System
→
Redix PAS
X12 278 to FHIR PAS conversion (payer side)
→
Payer
PA System

Provider to Payer

  • Submit PA via FHIR PAS (demonstrated: $submit and $inquire with simulated decisions)
  • Convert X12 278 to FHIR PAS (available for supported content)
  • CRD and DTR: synthetic demonstration scenarios; EHR integration, payer rules and the runtime to execute them are customer implementation work

Payer Backend

  • Legacy X12 278 systems
  • FHIR facade over an X12 278 backend: proposed; FHIR PAS to X12 278 conversion is available for online evaluation, and the integration around it is scoped per customer
  • RMap (Redix's intermediate format), with an agreed source adapter and mapping
  • RMap 278 request to FHIR PAS as an alternative path

Authentication

  • API key support (deployment-specific)
  • OAuth 2.0 via an identity provider agreed for the deployment
  • SMART on FHIR: App Launch demonstrated for DTR in a public test environment; production EHR registration and system-to-system security agreed per deployment

Where SMART on FHIR fits. SMART on FHIR is the authorization layer between the EHR and the apps it launches. In a prior authorization workflow it has two roles: App Launch, where the clinician opens the DTR questionnaire from the EHR and the app receives a limited, patient-scoped token; and Backend Services, where systems authenticate to each other without a user (for example, payer API access defined by the Da Vinci guides). CDS Hooks for CRD is a separate interface, and an EHR access token does not authorize calls to payer PAS endpoints. This explains the standard: Backend Services uses preregistered clients, signed authentication assertions and scoped access tokens, while the demo API uses API keys. Backend Services for a deployment is agreed and implemented per customer.

Demonstrated (public test environment)

  • SMART App Launch 2.0 EHR launch from the SMART Health IT sandbox
  • Authorization code with PKCE; patient-read scope only
  • Synthetic patient name and date of birth pre-filled into a Total Knee DTR questionnaire
  • Simulated PAS submission and status
SMART DTR walkthrough →

Customer Deployment Work

  • App registration with each EHR, with client authentication agreed with the EHR vendor
  • System-to-system security for payer APIs (for example SMART Backend Services), agreed per deployment
  • The payer's questionnaire content and pre-fill rules (CQL), plus a runtime to execute the rules, EHR data retrieval and mapping, and acceptance testing
  • HTTPS hosting, token handling and audit arrangements in the agreed environment

What Is / Is Not Redix

Capability Redix Provides Customer Provides
FHIR API Infrastructure Demo API and workflow components (not a packaged production service) Data sources and access policies; building out a production service is implementation work scoped separately, not only configuration
X12 278 Conversion X12 278 to FHIR PAS; FHIR PAS to X12 278 (online evaluation) —
Data Storage SQLite demo storage Other production storage integrations require separate design and implementation
Clinical Decisions — Your PA rules
Existing EDI Systems — Your integration
Identity Provider Authentication method, identity provider and operating responsibilities are agreed for the deployment
Provenance (source X12 transaction identifiers) FHIR Provenance targeting the Bundle in converted X12 278 output (not a complete access or event audit) —
Access and event audit logging Conversion request logging in the Redix API Retention, access-event logging and SIEM integration, defined for the deployment

Security & Data Sovereignty

Deployment requirements and design guidance. The public demo uses API keys over HTTPS with synthetic data; production controls are specified and verified for each deployment.

Security Layer

  • TLS 1.2 or 1.3 required; coverage of internal service and database connections is specified and verified for the deployment
  • API key support (deployment-specific; an API key alone does not establish production PAS/HRex security)
  • OAuth 2.0 via an identity provider agreed for the deployment
  • Conversion request logging; access and event audit logging, retention and SIEM arrangements are defined for the deployment
  • FHIR Provenance (source X12 transaction identifiers) in converted X12 278 output

Data Sovereignty

  • Self-hosted conversion engine — runs in the agreed customer environment
  • External services, data flows and deployment dependencies are defined for each integration
  • Full infrastructure control
  • Air-gapped deployment assessed per customer