Skip to main content

Always Generate Documentation Using the Clinician-Selected Diagnosis

Feature Request

One feature that would greatly improve confidence in AI-generated documentation is ensuring that the note is always written using the diagnosis selected by the treating clinician.

Occasionally, the AI appears to reinterpret the client's presentation through a different diagnosis than the one I have assigned. While diagnostic suggestions may be useful in certain contexts, progress notes should not automatically shift to a different diagnostic formulation.

The treating clinician is responsible for establishing the diagnosis. Once that diagnosis has been selected, the AI should generate the documentation through that diagnostic lens unless the clinician intentionally changes it.

If the AI identifies information that may warrant reconsidering the diagnosis, it could optionally provide a suggestion or prompt for clinician review, but it should not change the framework of the documentation itself.

This would improve clinical accuracy, consistency, and confidence in AI-generated notes while respecting clinician judgment.

Posted by Kathryn Eden ·
2 comments

This request was merged into another request

Comments2

  • Ian Singer

    •

    Jul 8

    If Upheal is going to implement a feature like this, important that it be optional, as many therapists do not use diagnoses and some of us (e.g. Clinical Social Workers and Psychotherapists in Ontario) are not permitted to make diagnoses.

    • Kathryn Eden

      •

      Jul 8

      I completely agree that it should be optional. My concern is actually a bit different: Upheal is already inserting a diagnosis into the progress note and then using that diagnosis to generate AI interpretations of client progress (e.g., “improved depression” or “improved stress reactivity”), even when those conclusions aren’t supported by the clinician’s documented diagnosis or the session itself. The diagnosis should come from the clinician, not the AI , and any diagnosis-based interpretations should only be generated after the therapist has selected or confirmed the diagnosis. Otherwise, the note can become clinically inaccurate and require significant editing.