Skip to content

Clinical Workflow Actions

UC-ONC-070: Generate Quick Patient Summary

Purpose: Generate an at-a-glance summary when opening a patient record, showing cancer type, stage, current treatment, and key metrics.

Property Value
Actor Summary Generation Service
Trigger Patient record opened
Priority P0

Main Success Scenario:

1. Retrieve patient oncology record
2. Generate summary with:
   - Cancer type & stage
   - Treatment intent (curative/palliative)
   - Line of therapy
   - Performance status (ECOG)
   - Current regimen & cycle day
   - Prior therapies (one-line)
   - Key biomarkers with date
   - Latest scan response
   - Key lab snapshot
   - Critical comorbidities/allergies
3. Display in top panel within 2 seconds

Alternative Flows:

Alt-1: Incomplete Data - Some summary fields missing - Display available data with "Incomplete" flags - Suggest data completion actions

Observability: - Metric: onc_summary_generated, onc_summary_load_time_ms - Log: {"event": "patient_summary", "patient_id": "P1234", "complete": true, "load_time_ms": 450}

Acceptance Criteria: 1. [ ] Summary loads in < 2 seconds 2. [ ] Covers all 10 key elements 3. [ ] Flags incomplete data


UC-ONC-071: Track Changes Since Last Visit

Purpose: Highlight what changed since the patient's last clinic visit for quick follow-up preparation.

Property Value
Actor Change Tracking Service
Trigger Follow-up visit initiated
Priority P0

Main Success Scenario:

1. Identify last visit date
2. Query all new data since last visit:
   - New lab results
   - New imaging studies
   - Dose modifications
   - Adverse events logged
   - New consultations/procedures
3. Generate "What Changed" summary
4. Flag significant changes (e.g., rising tumor markers, new lesions)

Alternative Flows:

Alt-1: First Visit - No prior visit found - Display as "First Visit - Baseline" - Show all available data as new

Observability: - Metric: onc_changes_tracked, onc_significant_changes_flagged - Log: {"event": "changes_since_visit", "patient_id": "P1234", "days_since": 21, "new_items": 8, "significant_flags": 2}

Acceptance Criteria: 1. [ ] Detects all new items since last visit 2. [ ] Highlights significant changes 3. [ ] Supports first-visit scenario


UC-ONC-072: Aggregate Multi-Source Patient Data

Purpose: Aggregate patient data from multiple sources (EMR, PACS, lab portals, outside records, WhatsApp images) into unified view.

Property Value
Actor Data Aggregation Service
Trigger Patient data request
Priority P0

Main Success Scenario:

1. Query connected data sources:
   - Hospital EMR/HIS
   - PACS/Radiology system
   - Laboratory Information System
   - Pathology lab portal
   - Outside hospital records
   - Patient-uploaded documents (WhatsApp, files)
2. Normalize data formats
3. Deduplicate overlapping records
4. Merge into unified patient timeline
5. Tag source for each data element

Alternative Flows:

Alt-1: Source Unavailable - One or more sources timeout/fail - Proceed with available sources - Display warning for unavailable sources - Retry in background
Alt-2: Patient-Uploaded Images - Receive WhatsApp/scanned images - Queue for OCR processing - Extract structured data - Link to patient record

Observability: - Metric: onc_sources_queried, onc_records_merged, onc_source_failures - Log: {"event": "data_aggregation", "sources_queried": 5, "records_merged": 23, "deduplicated": 3}

Acceptance Criteria: 1. [ ] Supports 6+ data source types 2. [ ] Handles source failures gracefully 3. [ ] Tags data provenance


UC-ONC-073: Draft Clinic Note with Citations

Purpose: Auto-generate draft clinic note with inline citations to source documents.

Property Value
Actor Note Generation Service
Trigger Consultation initiated
Priority P1

Main Success Scenario:

1. Gather current visit context (chief complaint, vitals)
2. Pull relevant patient data:
   - Recent labs with values
   - Recent imaging findings
   - Current treatment status
   - Adverse events
3. Generate structured clinic note:
   - Chief Complaint
   - History of Present Illness (with citations)
   - Assessment
   - Plan
4. Embed clickable citations to source documents
5. Present for physician review/edit

Alternative Flows:

Alt-1: Insufficient Data - Key sections cannot be populated - Generate partial note - Highlight sections requiring manual input

Observability: - Metric: onc_clinic_note_drafted, onc_note_citations_count - Log: {"event": "clinic_note_draft", "patient_id": "P1234", "sections_complete": 4, "citations": 12}

Acceptance Criteria: 1. [ ] Generates structured SOAP-style note 2. [ ] All facts have source citations 3. [ ] Requires physician review before saving


UC-ONC-074: Screen Clinical Trial Eligibility

Purpose: Quick screening of patient eligibility for active clinical trials based on cancer type, stage, biomarkers, and prior treatments.

Property Value
Actor Trial Matching Service
Trigger User requests trial check
Priority P1

Main Success Scenario:

1. Extract patient criteria:
   - Cancer type, histology, stage
   - Biomarker status (HER2, EGFR, PD-L1, etc.)
   - Prior lines of therapy
   - Performance status
   - Key lab values
2. Query active trial database
3. Match against inclusion/exclusion criteria
4. Rank trials by fit score
5. Display potential matches with match reasoning

Alternative Flows:

Alt-1: No Matches Found - No eligible trials - Display alternative options (off-label, compassionate use) - Suggest criteria that if changed would enable matches

Observability: - Metric: onc_trial_screens, onc_trial_matches_found - Log: {"event": "trial_screening", "patient_id": "P1234", "trials_checked": 45, "matches": 3}

Acceptance Criteria: 1. [ ] Screens against active trial database 2. [ ] Explains match/mismatch reasoning 3. [ ] Non-prescriptive (informational only)


UC-ONC-075: Flag Missing/Overdue Tests

Purpose: Identify and flag tests that are missing or overdue based on treatment protocol.

Property Value
Actor Protocol Compliance Checker
Trigger Patient record accessed or scheduled check
Priority P0

Main Success Scenario:

1. Identify patient's current treatment protocol
2. Load protocol-defined test schedule:
   - Labs (CBC, LFT before each cycle)
   - Imaging (every 2-3 cycles)
   - Tumor markers (per protocol)
3. Check against patient's actual test dates
4. Flag:
   - Missing: Test never done
   - Overdue: Test due but not completed
5. Display in patient summary and create alerts

Alternative Flows:

Alt-1: Custom Protocol - Non-standard protocol detected - Use physician-defined schedule if available - Otherwise, flag for manual review

Observability: - Metric: onc_missing_tests_flagged, onc_overdue_tests_flagged - Log: {"event": "test_compliance", "patient_id": "P1234", "missing": ["ECHO"], "overdue": ["CBC"]}

Acceptance Criteria: 1. [ ] Protocol-aware scheduling 2. [ ] Distinguishes missing vs overdue 3. [ ] Generates actionable alerts


UC-ONC-076: Capture Patient-Reported Outcomes (PROs)

Purpose: Capture and track patient-reported outcomes including symptoms, quality of life, and toxicities.

Property Value
Actor PRO Capture Service
Trigger PRO questionnaire submitted
Priority P1

Main Success Scenario:

1. Present validated PRO instrument (e.g., EORTC QLQ-C30, PRO-CTCAE)
2. Capture patient responses
3. Score questionnaire
4. Compare against previous PRO submissions
5. Flag worsening symptoms (> threshold change)
6. Display trends in patient summary

Alternative Flows:

Alt-1: Incomplete Submission - Patient abandons questionnaire - Save partial responses - Calculate partial score if possible - Prompt for completion at next visit

Observability: - Metric: onc_pro_submitted, onc_pro_symptom_flagged - Log: {"event": "pro_captured", "patient_id": "P1234", "instrument": "PRO-CTCAE", "flagged_symptoms": ["fatigue", "nausea"]}

Acceptance Criteria: 1. [ ] Supports standard PRO instruments 2. [ ] Tracks trends over time 3. [ ] Flags significant worsening


UC-ONC-077: Manage Clinical Task List

Purpose: Create, assign, and track clinical tasks and reminders for oncology care.

Property Value
Actor Task Management Service
Trigger Task created or patient event
Priority P1

Main Success Scenario:

1. Create tasks from various sources:
   - Tumor board recommendations
   - Missing test alerts
   - Follow-up scheduling
   - Manual entries
2. Assign to team member (physician, nurse, coordinator)
3. Set due date and priority
4. Track completion status
5. Escalate overdue tasks

Alternative Flows:

Alt-1: Task Reassignment - Assigned user unavailable - Allow reassignment with audit trail - Notify original and new assignee

Observability: - Metric: onc_tasks_created, onc_tasks_completed, onc_tasks_overdue - Log: {"event": "task_status", "task_id": "T456", "action": "completed", "days_to_complete": 2}

Acceptance Criteria: 1. [ ] Supports multiple task sources 2. [ ] Role-based assignment 3. [ ] Overdue escalation


UC-ONC-078: Compare Sequential Imaging Studies

Purpose: Side-by-side comparison of sequential imaging studies with lesion measurement overlay.

Property Value
Actor Imaging Comparison Widget
Trigger User requests comparison
Priority P1

Main Success Scenario:

1. Select two imaging studies for comparison
2. Align anatomically (same view/slice level)
3. Display side-by-side with synchronized scrolling
4. Overlay lesion measurements (baseline vs current)
5. Calculate and display % change per lesion
6. Highlight new lesions

Alternative Flows:

Alt-1: Different Modalities - Studies are different modalities (CT vs MRI) - Display warning about limited comparability - Allow comparison with disclaimer

Observability: - Metric: onc_imaging_comparisons, onc_new_lesions_detected - Log: {"event": "imaging_compare", "study1": "CT-2024-01", "study2": "CT-2024-04", "lesions_compared": 3, "new_lesions": 1}

Acceptance Criteria: 1. [ ] Side-by-side synchronized view 2. [ ] Measurement overlay 3. [ ] Highlights new lesions


UC-ONC-079: Generate Handover/Referral Packet

Purpose: Generate comprehensive patient packet for handover to another physician or referral to external facility.

Property Value
Actor Packet Generator
Trigger Referral requested
Priority P1

Main Success Scenario:

1. Compile patient data:
   - Diagnosis and staging
   - Treatment history (all modalities)
   - Key pathology and imaging reports
   - Genomics summary
   - Current medications
   - Recent labs
2. Generate structured PDF packet
3. Include table of contents
4. Allow selective inclusion/exclusion
5. Apply appropriate redactions if needed

Alternative Flows:

Alt-1: Consent Required - External referral requires consent - Check consent status - Block generation if consent not available - Prompt for consent collection

Observability: - Metric: onc_referral_packets_generated, onc_packet_sections_included - Log: {"event": "referral_packet", "patient_id": "P1234", "destination": "external", "sections": 8, "pages": 15}

Acceptance Criteria: 1. [ ] Comprehensive patient summary 2. [ ] Selective section inclusion 3. [ ] Consent-aware for external referrals


UC-ONC-080: Ingest Patient-Shared Documents

Purpose: Ingest and process documents shared by patients via WhatsApp, email, or file upload.

Property Value
Actor Patient Document Ingestion Service
Trigger Patient document received
Priority P1

Main Success Scenario:

1. Receive document from patient channel:
   - WhatsApp image
   - Email attachment
   - Web/app file upload
2. Validate document (image quality, file type)
3. Queue for OCR processing
4. Extract structured data (lab values, dates, findings)
5. Match to patient record via phone/ABHA ID
6. Present for clinician review before merging
7. Link to patient timeline with "Patient-Uploaded" tag

Alternative Flows:

Alt-1: Poor Image Quality - Image too blurry/dark for OCR - Notify patient to resend with better quality - Store original with "Unprocessed" flag
Alt-2: Patient Not Matched - Cannot match document to existing patient - Queue for manual matching - Alert registration desk

Observability: - Metric: onc_patient_docs_received, onc_patient_docs_processed, onc_patient_docs_merged - Log: {"event": "patient_document", "source": "whatsapp", "patient_id": "P1234", "doc_type": "outside_lab", "status": "merged"}

Acceptance Criteria: 1. [ ] Supports multiple input channels 2. [ ] OCR extraction accurate 3. [ ] Requires clinician review before merge