Purpose Statement

This article provides comprehensive training guidance for EMS personnel to effectively utilize the AI voice-to-text feature in First Due's ePCR system. The AI feature is designed to dramatically reduce documentation time by 60-70% while maintaining accuracy and completeness of patient care reports. By speaking naturally during patient care, the AI automatically populates ePCR fields, allowing personnel to focus more on patient care and less on manual data entry.

Background Information

Feature Overview

The AI voice-to-text feature uses advanced speech recognition and natural language processing to convert spoken patient care narratives into structured ePCR documentation. The system recognizes medical terminology, automatically categorizes information into appropriate fields, and maintains real-time documentation throughout patient encounters.

Common Use Cases

• Field documentation during active patient care
• Reducing post-call documentation time
• Ensuring comprehensive patient care records
• Minimizing transcription errors
• Maintaining hands-free operation during critical procedures

Prerequisites

• First Due ePCR module with AI feature enabled
• Device with microphone capability
• Stable internet connection for cloud processing
• Basic understanding of ePCR structure and requirements

Step-by-Step Guide

1. Preparing for AI-Assisted Documentation

1. Activate AI Feature: Navigate to Settings > ePCR > AI Voice-to-Text and ensure feature is enabled
2. Test Audio Setup: Verify microphone is working and positioned for clear speech capture
3. Review Patient Information Fields: Familiarize yourself with ePCR sections that AI will populate
4. Position Device: Place device where you can monitor AI transcription during patient care

2. Initiating AI Documentation

1. Start Recording: Tap the AI microphone icon when beginning patient contact
2. Begin Speaking Immediately: Start verbalizing patient information as soon as contact is made
3. Speak Continuously: Maintain narrative flow throughout the entire patient encounter
4. Monitor Real-Time Transcription: Glance periodically at screen to ensure AI is capturing information

3. Patient Demographics and Scene Information

1. State Basic Demographics: "68-year-old male patient responding to [full address]"
2. Include Patient Identifiers: "Patient lives at [home address if different], driver's license [number if obtained]"
3. Mention Service Area: "Patient resides within EMS service area" or specify if outside area
4. Note Scene Context: Include any relevant scene safety or environmental factors

4. Chief Complaint and Medical History Documentation

1. Present Chief Complaint: "Patient presents with [specific complaint] with [duration/onset]"
2. Detail Medical History: "Patient has history of [conditions], including [specific diagnoses]"
3. List Current Medications: "Currently takes [medication name] [dosage] [unit] [route]"
4. Document Allergies: "Allergic to [specific allergens], no known drug allergies" or specify known allergies
5. Identify History Source: "History obtained from patient/family/caregiver"

5. Physical Assessment Documentation

1. Use Systematic Approach: Follow head-to-toe or system-by-system assessment pattern
2. State Positive Findings: "Neurological assessment shows [specific findings]"
3. Include Negative Findings: "Head examination shows no abnormalities"
4. Document Injury Details: "Abdominal exam reveals [type of injury] with [current status]"
5. Note Assessment Progression: Reference time when significant changes occur

6. Vital Signs and Measurements

1. Time-Stamp Each Set: "First set of vitals at [time]" or "Second assessment at [time]"
2. State All Parameters: Include blood pressure, heart rate, respiratory rate, oxygen saturation
3. Describe Pulse Characteristics: "Pulse rhythm regularly irregular" or other specific qualities
4. Include Glasgow Coma Scale: "GCS eye [number], verbal [number], motor [number], total [number]"
5. Note Measurement Methods: Specify automatic vs. manual when relevant

7. Protocols and Procedures

1. Identify Protocols Used: "Following medical abdominal pain protocol for adult patient"
2. Document Procedure Details: "Performed [procedure name] at [time]"
3. Include Attempt Information: "One attempt successful" or specify number of attempts
4. State Authorization: "Authorization per standing order protocol" or specify medical direction
5. Note Crew Member: Identify who performed each procedure

8. Medication Administration

1. Complete Drug Information: "Administered [amount] [medication] [concentration] [route]"
2. Include Administration Time: "At [specific time]" for each medication given
3. State Authorization Source: "Authorization per protocol" or "per medical direction"
4. Document Patient Response: Note any immediate effects or reactions
5. Record Complications: Document any adverse reactions or difficulties

9. Response and Transport Information

1. Timeline Documentation: "PSAP call received [date/time], dispatch notified [time]"
2. Scene Arrival: "First EMS on scene, initial responder arrived [time]"
3. Patient Count: "Single patient at scene" or specify multiple patients
4. Unit Information: "EMS vehicle unit [number], [acuity level] dispatch priority"
5. Transport Destination: "Patient transported to [facility type/name]"

10. Final Documentation Review

1. Complete AI Transcription: Allow AI to finish processing all spoken information
2. Review Each Section: Check that information appears in correct ePCR fields
3. Fill Information Gaps: Manually add any details the AI may have missed
4. Verify Accuracy: Confirm all times, dosages, and critical information are correct
5. Complete Signatures: Add required electronic signatures and authorizations

Best Practices

Do's

• Speak naturally using standard medical terminology
• Include time references for all significant events and assessments
• State both positive and negative findings during physical assessment
• Use complete medication information including concentration and route
• Maintain continuous narrative throughout the patient encounter
• Review AI-generated content before finalizing the report

Don'ts

• Don't speak too quickly - allow AI time to process complex medical terms
• Don't rely solely on AI - always review and verify accuracy
• Don't skip systematic assessment - follow logical head-to-toe progression
• Don't forget authorization sources - specify protocols or medical direction
• Don't assume perfection - AI requires human oversight and correction

Optimization Tips

• Use consistent terminology that matches your department's standard language
• Speak clearly and distinctly especially when stating numbers and measurements
• Provide context for complex situations to help AI categorize information correctly
• Practice regularly to develop comfortable speaking patterns during patient care
• Leverage trending language like "improved from previous" or "remains stable"

Troubleshooting & FAQs

Common Issues and Solutions

Issue: AI not capturing spoken information

• Solution: Check microphone settings, reduce background noise, speak more slowly
• Prevention: Test audio setup before each shift, use headset in noisy environments

Issue: Information appearing in wrong ePCR fields

• Solution: Manually move information to correct sections, use more specific keywords
• Prevention: Use section-specific language and provide clear context

Issue: Incomplete medication documentation

• Solution: Repeat complete drug information, verify all required fields are populated
• Prevention: Use standardized medication reporting format consistently

Issue: Time stamps not matching actual events

• Solution: Manually correct times in ePCR, be more specific with time references
• Prevention: State exact times clearly when significant events occur

Frequently Asked Questions

Q: Will the AI work with my accent or speaking style? A: The AI learns from corrections and use patterns. Continue practicing and making corrections - the system adapts to individual speech patterns over time.

Q: Can I use AI in noisy environments like ambulances? A: Yes, but consider using a noise-canceling headset. The AI is designed to filter background noise, but clear audio significantly improves accuracy.

Q: What happens if I need to correct something I said? A: Simply say "correction" and restate the information, or edit the text manually after AI processing. The system allows for real-time and post-processing corrections.

Q: How secure is patient information processed by the AI? A: All AI processing follows HIPAA requirements and department security protocols. Patient data is encrypted during processing and not stored beyond the active session.

Q: Can I turn off AI for certain calls? A: Yes, the AI feature can be disabled on a per-call basis. However, departments typically see the most benefit from consistent use across all patient encounters.

Q: How do I handle complex multi-patient scenes? A: Document each patient separately, clearly identifying which patient you're discussing. The AI can handle multiple patient scenarios when properly structured.

Use Case Examples

Example 1: Cardiac Emergency

Scenario: 65-year-old male with chest pain AI Documentation: "Responding to 123 Main Street for 65-year-old male patient with chief complaint of crushing chest pain radiating to left arm, onset 30 minutes ago. Patient has history of hypertension and diabetes, takes lisinopril 10 milligrams daily and metformin 500 milligrams twice daily. No known allergies. Vital signs at 14:25 show blood pressure 180 over 100, heart rate 110 irregular, respiratory rate 22, oxygen saturation 94% on room air. Following cardiac protocol, established IV access, administered aspirin 324 milligrams chewed, nitroglycerin 0.4 milligrams sublingual per protocol."

Example 2: Trauma Call

Scenario: Motor vehicle collision with injuries AI Documentation: "Responding to motor vehicle collision at Highway 50 and Oak Street. 28-year-old female driver, restrained, airbag deployed. Chief complaint neck and back pain. Patient has no significant medical history, takes no medications, no known allergies. Physical assessment shows tenderness to cervical spine, no obvious deformity, neurological assessment intact all extremities. Applied cervical collar, full spinal immobilization per trauma protocol. Vital signs stable throughout transport."

Advanced Features & Integrations

NEMSIS Compliance

The AI system automatically formats documentation to meet NEMSIS (National EMS Information System) reporting requirements, ensuring compliance with state and federal EMS data standards.

CAD Integration

When integrated with Computer-Aided Dispatch systems, the AI can automatically populate incident numbers, dispatch times, and response information, reducing manual data entry.

Quality Assurance Integration

AI-generated reports integrate with First Due's quality assurance features, allowing supervisors to review documentation efficiency and accuracy metrics.

Performance Metrics

Success Indicators

• Documentation Time Reduction: Target 60-70% decrease in post-call documentation time
• Accuracy Rate: Maintain 95% or higher accuracy in AI-generated content
• Field Completion: Achieve 100% completion of required ePCR fields
• User Adoption: Track percentage of calls using AI documentation

Monitoring Tools

• Real-time accuracy feedback during AI transcription
• Post-call completeness scoring
• Department-wide usage analytics
• Quality assurance review integration