{"id":"epcr-ai-narrative-dictation","slug":"epcr-ai-narrative-dictation","title":"ePCR AI Narrative Dictation","description":"When a clinician has just completed a difficult call, the most complete version of the story is often still in their head. Reconstructing that narrative later from short field notes leads to omissions, inconsistent phras","category":"ai","tags":["ai"],"lastModified":"2026-04-17","source_ref":"content/modules/epcr-ai-narrative-dictation.md","url":"/developers/epcr-ai-narrative-dictation","htmlPath":"/developers/epcr-ai-narrative-dictation","jsonPath":"/api/docs/modules/epcr-ai-narrative-dictation","markdownPath":"/api/docs/modules/epcr-ai-narrative-dictation?format=markdown","checksum":"5303f149fd943795c0aa7582524610b8cdee6eee1743c21ec8baeb6108b661ce","headings":[{"id":"overview","text":"Overview","level":2},{"id":"key-features","text":"Key Features","level":2},{"id":"use-cases","text":"Use Cases","level":2},{"id":"integration","text":"Integration","level":2},{"id":"open-standards","text":"Open Standards","level":2}],"markdown":"# ePCR AI Narrative Dictation\n\n## Overview\n\nWhen a clinician has just completed a difficult call, the most complete version of the story is often still in their head. Reconstructing that narrative later from short field notes leads to omissions, inconsistent phrasing, and avoidable audit queries. At the same time, asking crews to type a long narrative while moving between scene, vehicle, and hospital is not realistic.\n\nThe AI Narrative Dictation module lets clinicians speak a short clinical note and turn it into a structured documentation proposal for review before it is written into the encounter. Instead of replacing clinician judgement, it reduces the friction of getting a high-quality narrative into the record while preserving explicit human acceptance of every field-level change.\n\n```mermaid\nflowchart TD\n    A[Clinician Speaks or Types Note] --> B[Transcript and Clinical Parsing]\n    B --> C[Structured Documentation Proposal]\n    C --> D[Field-by-Field Review]\n    D --> E{Accept Change?}\n    E -->|Yes| F[Encounter Updated]\n    E -->|No| G[Proposal Rejected or Edited]\n    F --> H[Audit of Accepted AI Assistance]\n    G --> H\n```\n\n**Last Reviewed:** 2026-04-17\n**Last Updated:** 2026-04-17\n\n## Key Features\n\n- **Hands-Free Narrative Capture**: Let clinicians capture a short spoken note during or immediately after care rather than waiting to reconstruct it from memory later.\n\n- **Structured Proposal, Not Blind Auto-Fill**: Convert the dictated note into a proposed clinical update that the clinician reviews and accepts, field by field, before it changes the encounter.\n\n- **Clinical Vocabulary Normalisation**: Improve consistency around common clinical abbreviations, pathway names, and shorthand so the final narrative is easier to review and audit.\n\n- **Mobile and Desktop Entry Points**: Support the same dictation workflow in the primary ePCR workspace and the mobile responder experience so crews can use the capability in the context that fits the call.\n\n- **Manual Transcript Fallback**: Allow the same structured proposal flow to work from typed narrative where microphone use is not practical or live transcription is not available.\n\n- **Field-Level Acceptance Workflow**: Preserve clinician control by letting the user accept some proposed changes and reject others instead of treating dictation as an all-or-nothing overwrite.\n\n- **Auditability of AI Assistance**: Record that narrative assistance was used and preserve the accepted change set for later review.\n\n## Use Cases\n\n- **On-Scene Narrative Capture**: A paramedic dictates a summary before leaving scene so the most accurate account is captured while the details are fresh.\n\n- **Back-of-Ambulance Documentation**: During transport, a clinician records a short note describing assessment, interventions, and response while hands remain focused on patient care.\n\n- **Hospital Handover Narrative Clean-Up**: After handover, the crew refines the encounter with a dictated summary rather than typing a long narrative from scratch.\n\n- **Supervisory Review of Complex Cases**: A governance lead reviewing a difficult case sees a more complete narrative because the clinician documented it in natural language at the point closest to care.\n\n- **Field Mobility Workflow**: Mobile crews working away from a full keyboard still get structured documentation quality without waiting to return to base.\n\n## Integration\n\n- **Electronic Patient Care Report Clinical Workspace**: Dictation proposals are reviewed and applied directly against the active encounter.\n\n- **Mobile Responder Workspace**: The same assisted narrative workflow can be exposed in the field when crews are working from the mobile surface.\n\n- **Clinical Audit**: Accepted proposal history gives reviewers more context about how the final narrative was assembled.\n\n- **AI Guardrails and Routing**: Narrative assistance uses the existing platform AI control layer rather than introducing an isolated one-off model workflow.\n\n## Open Standards\n\n- **Media Capture and Streams**: microphone access for browser-based dictation relies on the standard web media capture model.\n\n- **MediaStream Recording API**: client-side voice capture uses the standard browser recording interface rather than a proprietary capture plugin.\n\n- **Vibration API**: mobile dictation feedback can use the standard web haptics interface on supported devices.\n\n- **HTTPS / TLS**: dictated content and proposal review move over the platform's standard secure web transport.\n\n- **JSON**: proposal diffs and clinician-reviewed changes are exchanged in a structured, interoperable format.\n\n- **WCAG 2.2 AA**: the review and acceptance flow can be presented in the same accessibility baseline as the rest of the clinical workspace.\n"}