Visual Notes
Narrative critical-care nursing documentation, restructured into a source-verifiable body-system chart — so the patient story reads at a glance instead of being reassembled from paragraphs.
- 44.2%
- lower perceived workload
- 41
- critical-care nurses
- P<.001
- across all six subscales
Pilot within-subject study, NASA-TLX · Zaidi et al., JMIR Nursing
Workspace
Notes in, chart out
Narrative shift notes are segmented by body system, extracted to a schema, checked for coverage, then drawn as a swim-lane trajectory.
Written notes
One admission · narrative SOAP notes
Ready to convert. Paste a case, then run the pipeline.
Visual note
Body-system lanes across the admission
No visual note yet
Paste a series of nursing notes on the left and run the conversion. The body-system chart renders here.
Method
Nothing on the chart without a receipt
The pipeline is built so that a fabricated finding is structurally impossible rather than merely unlikely, and so that “did we lose anything?” has a number for an answer.
- 01
Segment
Notes split on record boundaries, then on the nurse's own section headers — NEURO, RESP, CV, GI, GU, ID, ENDO. Deterministic, no model involved.
- 02
Extract
Each segment is mapped to a frozen body-system schema. The lane taxonomy is fixed for every case, so two charts are always comparable.
- 03
Ground
Every extracted fact carries the note id and character offsets of the text it came from, and is byte-checked against the source. Unverifiable items never reach the chart.
- 04
Cover
Sentences with no linked fact are fed back for a second pass. Retention stops being a judgement call and becomes a measured percentage.
- 05
Render
Validated JSON is drawn as a swim-lane trajectory. Same input, same chart — the renderer is deterministic and diff-testable.
Research
Visual versus written nursing notes
Forty-one registered critical-care nurses reviewed the same patient admission twice — first as conventional written SOAP notes, then as a structured visual note — rating perceived workload on the raw NASA‑TLX after each condition.
Total workload fell from 54.95 to 30.68, a mean reduction of 24.27 points. All six subscales improved, with the largest effects in frustration and temporal demand. As a fixed-order pilot on a single case, the result is preliminary — order effects cannot be ruled out.
Zaidi M, Tanguay A, Zaidi N, Sial Q, Gazarian PK, Thakral M, Walfre F, Dykes PC. JMIR Nursing (preprint 102954).