Clinical Communication Design System and Prototype Specifications

This is a research prototype specification, not production UI or medical advice. Examples use placeholders and intentionally contain no real patient data or clinical thresholds.

Component library specification

Components own intrinsic meaning and state. Parent compositions own priority, order, density, and role/task context. No component may infer clinical severity from styling tokens alone.

Wireframe A: clinician longitudinal summary

┌ PATIENT / ENCOUNTER / LOCATION ───── data through HH:MM ──┐
│ IDENTITY CHECK     ALLERGIES/CONSTRAINTS     [scope]       │
├────────────────────────────────────────────────────────────┤
│ NOW: assessed severity + basis + assessor + time           │
│ CHANGED: 3 material changes | 1 contradiction | 2 pending  │
├───────────────────────┬────────────────────────────────────┤
│ ACTIVE PROBLEMS       │ ACTIONS / OWNERS / CONTINGENCIES   │
│ state • confidence    │ state • due • acknowledgement      │
│ support ↔ contradict  │ escalation condition               │
├───────────────────────┴────────────────────────────────────┤
│ TRAJECTORIES  [vitals] [labs] [medications] [events]       │
│ annotated timeline with source and intervention markers    │
├────────────────────────────────────────────────────────────┤
│ EVIDENCE / SOURCE RECORDS / AUDIT                           │
└────────────────────────────────────────────────────────────┘

Wireframe B: referral packet

┌ REFERRAL QUESTION + URGENCY BASIS + REQUESTING CLINICIAN ┐
│ What decision/help is requested? What is the deadline?    │
├───────────────────────────────────────────────────────────┤
│ ONE-LINE SITUATION          │ SAFETY CONSTRAINTS          │
│ COURSE + WHAT CHANGED       │ allergies / precautions     │
├─────────────────────────────┴─────────────────────────────┤
│ WORKUP: completed / pending / unavailable / contradictory │
│ TREATMENTS TRIED: response + reason stopped               │
│ RELEVANT EVIDENCE: source-linked                          │
├───────────────────────────────────────────────────────────┤
│ FOLLOW-UP OWNER • communication state • contingency       │
└───────────────────────────────────────────────────────────┘

Wireframe C: patient result explanation

┌ YOUR RESULT ─ collected DATE ─ reviewed/unreviewed state ┐
│ Plain-language name     [number line + labeled zones]     │
│ What this result can mean / cannot establish alone       │
├──────────────────────────────────────────────────────────┤
│ WHAT HAPPENS NEXT       │ WHEN TO SEEK HELP              │
│ named action + timing   │ verified instructions/contact │
├─────────────────────────┴────────────────────────────────┤
│ Questions to ask • source report • accessible data table │
└──────────────────────────────────────────────────────────┘

High-fidelity reference language

Visual tokens

The values are prototype starting points. A high-fidelity prototype must demonstrate light/dark, 200% zoom, narrow viewport, keyboard, screen reader, grayscale, and print states before usability study.

Interaction behavior

• Initial focus and reading order follow clinical priority and DOM semantics.
• Expansion controls announce state and concealed item count.
• Keyboard commands are optional accelerators with visible alternatives.
• Updating data never steals focus; material changes are announced proportionately.
• Destructive or high-consequence actions require explicit confirmation designed from hazard analysis.

Anti-pattern catalog

Decision framework

For each proposed communication element, record:

1. Reader, role, expertise, environment, device, and task.
2. Decision/action supported and harm if missed or misunderstood.
3. Source-of-truth, freshness, reliability, and transformation.
4. Competing representations, including a low-complexity baseline.
5. Why an item is included, ordered, emphasized, summarized, or hidden.
6. Accessibility and equity risks.
7. Failure modes: omission, false prominence, ambiguity, stale state, unit error, automation bias.
8. Outcome measures and noninferiority safety bounds set with clinical/safety experts.
9. Governance owner, monitoring plan, rollback, and invalidation trigger.

Release gate

A pattern is not “CCE validated” until representative scenario testing demonstrates predefined performance, no unacceptable critical-error signal, accessibility conformance, subgroup review, and clinician/patient safety approval appropriate to intended use.