Consumer Mobile Health Application Functional Framework, Release 2
2.0.1 - CI build
Consumer Mobile Health Application Functional Framework, Release 2, published by HL7 International / Mobile Health. This guide is not an authorized publication; it is the continuous build for version 2.0.1 built by the FHIR (HL7® FHIR® Standard) CI Build. This version is based on the current content of https://github.com/HL7/cmhaff-ig/ and changes regularly. See the Directory of published versions
| Official URL: http://hl7.org/ehrs/uv/cmhaffr2/Requirements/CMHAFFR2-AIG.4 | Version: 2.0.1 | ||||
| Standards status: Informative | Computable Name: AIG_4_AI_Validation__Performance__and_Equity | ||||
This category is about evaluating an AI function against its documented context of use before deployment, documenting what was and was not evaluated, and establishing whether performance differs materially across the populations the application serves.
| AIG.4#01 | SHALL conditional |
IF the system uses AI functionality THEN the system SHALL evaluate the AI function, before deployment, for performance, safety, reliability, and usability in the documented context of use and against pre-specified acceptance criteria. |
| AIG.4#02 | SHALL conditional |
IF the system uses AI functionality that can affect health, safety, access, or material wellbeing THEN the system SHALL evaluate whether performance differs materially across relevant subgroups, including populations defined by age, sex, language, disability, geography, device type, health condition, or other factors relevant to the intended use, and SHALL address identified unacceptable disparities. |
| AIG.4#03 | SHALL conditional |
IF the system uses AI functionality THEN the system SHALL document the performance measures used, their limitations, the evaluation dataset characteristics, and the conditions under which the AI function was not evaluated or did not meet acceptance criteria. |
| AIG.4#04 | SHOULD conditional |
IF the system uses AI functionality THEN the system SHOULD obtain external or independent validation when the risk assessment identifies elevated potential for health, safety, equity, or consumer-protection harm. |
| AIG.4#05 | MAY conditional |
IF the system uses AI functionality THEN the system MAY publish a performance dashboard or public validation summary that identifies the intended use, relevant performance metrics, and limitations. |