Skip to content

Clarification requested for 9.23 fine-grained scopes baseline when Single Patient API uses additional_patient_ids #750

@vs1785

Description

@vs1785

I’m looking for clarification on how 9.23 SMART App Launch with fine-grained scopes is intended to be executed when 10 Single Patient API (US Core 6.1.0) is run with multiple patients.

Read-only test run:
https://inferno.healthit.gov/suites/g10_certification/8HsUwWP6RzC/#9.23.1/view

What I ran:

  • I ran the full suite through 10 Single Patient API (US Core 6.1.0).
  • For 10, I used:
    • 1 launch patient
    • 3 additional_patient_ids
  • 10 Single Patient API (US Core 6.1.0) completed successfully.
  • My understanding from the suite wiring is that 10 uses the generic smart_auth_info exported by the earlier EHR Practitioner App, so the baseline run is using user/*.rs scope.
  • Later, 9.23.1 Granular Scopes 1 performs a standalone launch with patient-level granular scopes.

Observed behavior:

  • 9.23.1 appears to reuse baseline requests generated during 10.
  • Those saved requests include requests for all 4 patients from the earlier Single Patient API run.
  • During 9.23.1, Inferno then replays requests involving all 4 patients while using a patient-scoped granular token, which leads to failures.

Questions:

  1. Is 9.23 intended to be run only after a prerequisite Single Patient API baseline that uses a single patient and no additional_patient_ids?
  2. If Single Patient API is run with multiple patients to satisfy MUST SUPPORT coverage, what is the expected way to make 9.23 pass?
  3. Is the expected workflow to use a separate Inferno session for 9.23, rerun the prerequisite Single Patient API with only the launch patient, and then run 9.23?
  4. For 9.3 Token Revocation, there was no prompt in the test flow telling me to revoke the token during execution. Is the expectation that the tester manually revokes the token outside Inferno before submitting the 9.3 inputs?
  5. For ONC certification, is it expected that the entire g10 test kit passes in one continuous run, or is it acceptable to use separate runs/sessions for scenarios in section 9 that require different setup?

Thanks.

Metadata

Metadata

Assignees

No one assigned

    Labels

    question answeredThis issue has been investigated and replied by Inferno team. No further action needed.

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions