Prism / Clinical Workflow Lab
Keep the whole
story in view.
An activity is only part of an account. I built Prism to connect what happens before, during and afterwards with the original words, the next question and the draft.
The question
What gets lost
between the note and the report?
Support from another person. A pause in the middle. The next task left until later. Prism keeps that context attached to an everyday activity, so the person reviewing the account can see what is recorded and what still needs asking.
01 / Try Prism
Follow one account through.
Inspect → Ask → Review
Prism / The corner shop
Fictional case · In this page only
Original words
S1 / Questionnaire
“I usually go to the corner shop on a Tuesday.”
Self-report · fictional
The original wording stays attached to the draft. Editing the draft does not rewrite this source.
The short account
“I go to the corner shop.”
An activity can sound simple until you look at the support, pauses and variation around it.
A guided sample of Prism. Fictional text stays in this page; refreshing starts a new session.
02 / The design decisions
The source stays.
The draft can change.
- Preserve what was actually said
- Each passage has an identity and keeps its original wording. A follow-up answer becomes a new source linked to its question; the draft remains separately editable.
- Make the missing detail visible
- Questions point to something specific in the record. A pause with no duration prompts an enquiry, rather than a conclusion about what someone can do.
- Treat review as something that can expire
- Editing a paragraph reopens its review. Adding evidence reopens the whole draft. The export carries the original sources and unresolved questions with it.
Under the surface
This portfolio sample uses Prism’s existing fictional case fixtures and source/export engine, with an interface adapted for a short demonstration. Source records, question state, editable paragraphs and review state are kept separate.
The sample runs locally in the page and uses authored questions. The wider prototype has optional provider integrations; they are not connected here. Refreshing clears the session. Export is an explicit download of the reviewed text.
This is a workflow experiment, not a validated assessment system. It does not infer capacity, select descriptors or recommend an outcome. Time savings and clinical benefit have not been measured.
The same question, through teaching