Comment on draft standard for Patient-Visible and Accountable AI in Clinical Decision Support

[Chris Johnston]

Pleased to share an opportunity to have your say on a new draft standard: CAN/DGSI 138, Patient-Visible and Accountable AI in Clinical Decision Support has now been posted for 60-day Public Review on the DGC website.

The Public Review period will run from August 5, 2026 to October 5, 2026, during which time any interested party may comment on this proposed first edition of the National Standard of Canada by visiting the DGC website.

All comments submitted will be compiled and dispositioned by the Technical Committee at a future Technical Committee #2 meeting to be scheduled following the closing of the public review period.

To comment on the draft standard:

• Visit the DGC website link.

• On the dedicated page for the standard, please click on the Provide Comments tab, where you will have access to the draft standard.

[Joan Tu]

@Chris-Johnston Thank you Chris. I have commented on this draft standard in this public review and believe it would be beneficial for more patients to participate.

[Chris Johnston]

@Joan-Tu
Many thanks Joan!

[Debra Turnbull]

@Joan-Tu
Thank you Joan.

For everyone:
this standard is meant to establish "rules" for comparing Artificial Intelligence systems used in Clinical Decision Support (CDS). These CDS systems are getting slowly incorporated into your doctors' EMRs (electronic medical records).

As a standard, this is what a CDS system will be compared to. People like Accreditation Canada will use such standards as a way of reporting on organizations (hospitals, clinics) about their quality of care.

Some interesting components of the standard:

• the UAII: acts as a "license plate" for the AI module (sect. 6.3)
• dataset provenance: there needs to be a trace of what data was used for training the AI module (sect. 7.1.1)
• AI Decision Summary for patient visibility (sect. 5)

Section 5.4.3 - who was asking about advertising in clinical systems?