Suggested (g)(10) test kit workflow improvement for test 9.14 - SMART Launch with Fine-Grained Scopes (SMART v2.0.0)

[Justin-Ramm]

Observed Workflow Issue

In a typical run-through of the g10 test kit, test 10 Single Patient API is run using a practitioner-level access token from test 3 EHR Practitioner App, and includes data from multiple patients - particularly for systems certifying to USCDI v3, which requires data from multiple patients to realistically meet all data requirements.

Currently Test 9.14 relies on data from test 10 Single Patient API to compare against. However, test 9.14 expects a workflow where the user selects the limited scopes to be tested. From the test description: "the system under test can display more scopes to the tester during authorization. In this case, it is expected that the tester will only approve the appropriate scopes in each group as described above." To mimic a realistic workflow, this user is likely to be a single patient, as it would be atypical to ask a provider to select individual scopes as part of an app that is pre-configured to integrate with their EHR system workflows.

This presents a challenge when going through a full (g)(10) test kit workflow, because the results returned for a single patient client in test 9.14 cannot be accurately compared against multi-patient results of test 10. That means that the system under test needs to:

1. Re-run test 1 Standalone Patient App to reset auth context that will be used by test 10
2. Re-run test 10 Single Patient API on a single patient
   • Parts of this test will now fail because only a single patient is included (ex: deceasedDateTime missing in test 10.2.09)
3. Run test 9.14 to accurately compare against single-patient results from test 10
4. Re-run test 3 EHR Practitioner App to reset auth context that will be used by test 10
5. Re-run test 10 to ensure all tests pass once again.

Test 10 can take a long time to run. If a system under test runs through the test kit in order, they may have to run test 10 three times throughout the course of the workflow if accounting for the steps above. This is very inefficient.

Suggested Improvement

It might make more sense to relocate test 9.14 higher in the workflow sequence; perhaps alongside test 2 Limited Access App, which feels technically and conceptually similar. Regardless of what might be done here, it would be generally helpful if the pre-requisite tests that 9.14 relies on for comparison were separated out from test 10 to avoid having to wait for that entire test to re-run when only a small proportion of data from it is required for comparison. Test 10 takes longer than any other test in the series (and that trend will likely continue as USCDI expands in future versions), so Inferno should strive to avoid scenarios that require systems to re-run this test multiple times during the overall g10 test kit workflow.

[Jammjammjamm]

the system under test can display more scopes to the tester during authorization. In this case, it is expected that the tester will only approve the appropriate scopes in each group as described above.

Having the user manually select/deselect scopes is not an expected part of this workflow, but a possible one, depending on exactly how the system under test behaves and how Inferno is registered with it.

The only tests which require that users manually select/deselect scopes are 2 and 9.15.