About Prompt
- Prompt Type – Dynamic
- Prompt Platform – ChatGPT, Grok, Deepseek, Gemini, Copilot, Midjourney, Meta AI and more
- Niche – Voice-Based Data Entry
- Language – English
- Category – Speech-to-Text EHR Input
- Prompt Title – AI Prompt for Automating Voice-based Patient Data Collection
Prompt Details
This prompt is designed for dynamic generation and usage across various AI platforms for speech-to-text EHR input, automating patient data collection via voice. It prioritizes accuracy, completeness, and adherence to HIPAA regulations.
**Prompt Template:**
“`
## Voice-based Patient Data Collection for EHR Input
**Patient Identifier:** {{patient_id}}
**Date and Time of Recording:** {{date_time}}
**Physician/Clinician Name:** {{clinician_name}}
**Medical Specialty:** {{medical_specialty}}
**Encounter Type:** {{encounter_type}} (e.g., Inpatient, Outpatient, Telehealth)
**Instructions for the AI:**
1. **Transcribe the following voice recording accurately, correcting for common speech disfluencies (e.g., “um,” “ah,” repetitions).** The recording contains patient responses related to their medical history, current symptoms, and lifestyle.
2. **Identify and extract key medical information** according to the specified data fields below. If a data field is not mentioned in the recording, mark it as “Not Mentioned.” Ensure numerical values are extracted correctly with appropriate units.
3. **Structure the extracted information** using the provided JSON format. Ensure data types match the specified format.
4. **Prioritize accuracy and completeness.** If information is ambiguous or uncertain, flag it with a confidence score (e.g., “high,” “medium,” “low”) within the JSON structure. If necessary, include the verbatim transcribed text snippet related to the uncertain information.
5. **Adhere to HIPAA regulations.** Remove or redact any Protected Health Information (PHI) not explicitly requested in the data fields below, including names of relatives, addresses, and specific dates (except date of birth). Replace such information with “[PHI Removed].”
**Data Fields (JSON Format):**
“`json
{
“patient_id”: “{{patient_id}}”,
“date_time”: “{{date_time}}”,
“chief_complaint”: { “value”: “”, “confidence”: “”},
“history_of_present_illness”: { “value”: “”, “confidence”: “”},
“past_medical_history”: { “value”: “”, “confidence”: “”},
“medications”: { “value”: [], “confidence”: “”}, // List of medications
“allergies”: { “value”: [], “confidence”: “”}, // List of allergies
“family_history”: { “value”: “”, “confidence”: “”},
“social_history”: {
“smoking_status”: { “value”: “”, “confidence”: “”},
“alcohol_use”: { “value”: “”, “confidence”: “”},
“exercise_frequency”: { “value”: “”, “confidence”: “”}
},
“review_of_systems”: { “value”: “”, “confidence”: “”},
“vital_signs”: {
“heart_rate”: { “value”: “”, “unit”: “bpm”, “confidence”: “”},
“blood_pressure”: { “value”: “”, “unit”: “mmHg”, “confidence”: “”},
“respiratory_rate”: { “value”: “”, “unit”: “breaths/min”, “confidence”: “”},
“temperature”: { “value”: “”, “unit”: “°F/°C”, “confidence”: “”},
“oxygen_saturation”: { “value”: “”, “unit”: “%”, “confidence”: “”}
},
“physical_exam_findings”: { “value”: “”, “confidence”: “”},
“assessment_and_plan”: { “value”: “”, “confidence”: “”},
“verbatim_transcript”: “”, // Include the full transcript here for reference and auditing
“flagged_uncertainties”: [] // Array of objects: { “field”: “”, “text_snippet”: “”, “confidence”: “”}
}
“`
**Voice Recording:**
*Begins here*
… (Patient’s voice data)…
*Ends here*
**Additional Notes:**
* This prompt is adaptable. You can modify the requested data fields within the JSON structure to match specific clinical needs.
* Consider providing examples within the prompt if dealing with highly specialized medical terminology.
* For improved performance, train the AI model on a dataset of medical conversations and transcribed EHR data.
* Regularly evaluate the AI’s output for accuracy and identify areas for improvement.
This dynamic prompt provides a robust framework for automating voice-based patient data collection for EHR input, ensuring high quality, structured data while maintaining patient privacy. It’s essential to validate and refine this prompt with real-world data and user feedback for optimal performance within your specific clinical setting.
“`
This detailed prompt enables consistent, accurate, and structured data extraction from voice recordings, streamlining EHR data entry and improving clinical workflow. The dynamic nature of the prompt allows for customization based on specific clinical requirements and continuous improvement based on performance evaluation. Remember to prioritize data security and HIPAA compliance throughout the implementation process.