Skip to content

ROLE_BASED_CHATBOT_GUIDE.md

Latest commit

 

History

History
426 lines (357 loc) · 10.1 KB

ROLE_BASED_CHATBOT_GUIDE.md

File metadata and controls

426 lines (357 loc) · 10.1 KB

🤖 Role-Based Chatbot System

Overview

The SmartCare Hospital Chatbot now features role-based intelligence that provides different features and responses based on whether the user is a Patient or Management/Admin.

Two Distinct Modes

👤 Patient Mode

For: Patients and visitors Focus: Personal assistance, appointments, health information

👨‍💼 Management Mode

For: Hospital administrators and staff Focus: Analytics, operations, queue management, reports

Patient Mode Features

1. 📅 Book Appointments

  • Browse departments
  • View available doctors
  • Check doctor ratings and experience
  • Select appointment slots

Example:

Patient: "Book an appointment"
Bot: Shows departments → Select doctor → Choose time

2. 🔍 Check Appointment Status

  • Track appointment by phone number
  • View appointment details
  • See current status (confirmed, waiting, completed)

Example:

Patient: "Check my status"
Bot: "Please provide your phone number"
Patient: "9876543210"
Bot: Shows all appointments with status

3. ⏰ Estimated Appointment Time

NEW FEATURE - Get precise timing information:

  • Scheduled appointment time
  • Current queue position
  • Estimated actual time
  • Expected wait duration

Example:

Patient: "What's my estimated time?"
Bot: 
"⏰ Your Appointment Estimate:
📅 Date: March 15, 2024
🕐 Scheduled: 10:00 AM
📊 Queue Position: #3
⏱️ Estimated Time: 10:30 AM
⌛ Expected Wait: ~30 minutes"

4. 📋 Precautions & Preparation

NEW FEATURE - Department-specific guidance:

  • What to bring
  • How to prepare
  • Fasting requirements
  • Clothing recommendations
  • Timing advice

Example:

Patient: "Precautions for cardiology"
Bot:
"📋 Precautions for Cardiology:
🩺 Bring previous ECG/Echo reports
💊 List of current medications
🚫 Avoid caffeine 2 hours before
👕 Wear comfortable, loose clothing
📋 Fasting may be required for some tests"

Department-Specific Precautions:

  • Cardiology: ECG prep, medication list, avoid caffeine
  • Neurology: MRI reports, symptom notes, adequate sleep
  • Orthopedic: X-rays, comfortable shoes, walking aids
  • Pediatrics: Vaccination records, growth chart, comfort toys
  • General: ID proof, medical records, insurance

5. 👨‍⚕️ Find Doctors

  • Search by specialty
  • View doctor ratings
  • Check experience years
  • See availability

6. ⏱️ Wait Time Estimates

  • Current wait times by department
  • Crowd level indicators
  • Best time to visit recommendations

7. 🏥 Department Information

  • List all departments
  • Floor locations
  • Available services

8. 📊 Crowd Predictions

  • Best times to visit
  • Peak hour warnings
  • Tomorrow's forecast

Management Mode Features

1. 📊 Queue Statistics

Real-time operational data:

  • Total appointments today
  • Patients waiting
  • Patients in progress
  • Completed consultations
  • Completion rate
  • Department-wise breakdown

Example:

Admin: "Queue statistics"
Bot:
"📊 Live Queue Statistics:
📅 Today's Overview:
• Total Appointments: 45
• ⏳ Waiting: 12
• 🔄 In Progress: 8
• ✅ Completed: 25
🏥 Active Queues by Department:
• Cardiology: 5 patients
• Neurology: 3 patients
• Orthopedic: 4 patients
⚡ Completion Rate: 55.6%"

2. 📈 Today's Summary Report

Comprehensive daily overview:

  • Total appointments
  • Completion statistics
  • No-show rate
  • Active departments
  • Available doctors
  • Average wait time
  • System status indicator

Example:

Admin: "Today's summary"
Bot:
"📈 Today's Summary Report
📅 March 15, 2024
👥 Appointments:
• Total: 45
• Completed: 25
• No-shows: 2
• Show Rate: 95.6%
🏥 Operations:
• Active Departments: 8
• Available Doctors: 15
• Avg Wait Time: ~25 min
💡 Status: 🟢 Normal"

3. 🏥 Department Performance

Performance metrics by department:

  • Total appointments per department
  • Completion rates
  • No-show rates
  • Efficiency analysis

Example:

Admin: "Department performance"
Bot:
"📊 Department Performance Today:
• Cardiology:
  Total: 15 | Completed: 10 | Rate: 66.7%
• Neurology:
  Total: 12 | Completed: 8 | Rate: 66.7%
• Orthopedic:
  Total: 10 | Completed: 5 | Rate: 50.0%"

4. 👨‍⚕️ Doctor Availability

Staff availability tracking:

  • Number of available doctors
  • Number of unavailable doctors
  • List of currently available doctors
  • Department assignments

Example:

Admin: "Doctor availability"
Bot:
"👨‍⚕️ Doctor Availability:
✅ Available: 15 doctors
❌ Unavailable: 3 doctors
Currently Available:
• Dr. John Smith - Cardiology
• Dr. Sarah Johnson - Neurology
• Dr. Mike Brown - Orthopedic"

5. 🚨 High-Risk Patient Alerts

Priority patient identification:

  • Emergency cases
  • High-risk patients
  • Patients requiring immediate attention
  • Appointment times
  • Department assignments

Example:

Admin: "High-risk patients"
Bot:
"🚨 High-Risk Patients Today:
Found 3 high-priority cases:
• John Doe - Cardiology (10:00 AM)
• Jane Smith - Neurology (11:30 AM)
• Bob Wilson - Emergency (09:45 AM)
⚠️ These patients require immediate attention.
💡 Consider prioritizing in queue management."

6. ⚠️ No-Show Predictions

ML-powered predictions:

  • Identify high no-show risk appointments
  • Risk percentage for each patient
  • Proactive reminder suggestions

Example:

Admin: "No-show predictions"
Bot:
"⚠️ No-Show Risk Analysis:
High-risk appointments: 3/45
• Patient A - 10:00 AM (Risk: 65%)
• Patient B - 02:30 PM (Risk: 58%)
• Patient C - 04:00 PM (Risk: 52%)
💡 Consider sending reminder SMS to these patients."

7. 📉 Crowd Forecast

Predictive analytics:

  • Tomorrow's crowd predictions
  • Hour-by-hour forecast
  • Peak time identification
  • Staffing recommendations

Example:

Admin: "Crowd forecast"
Bot:
"📊 Crowd Forecast for Tomorrow:
March 16, 2024
🟢 8:00 AM: Low
🟡 10:00 AM: Medium
🟠 12:00 PM: High
🟡 2:00 PM: Medium
🟢 4:00 PM: Low
💡 Recommendations:
• Schedule more staff during peak hours
• Prepare for high-volume periods
• Consider opening additional counters"

How Role Detection Works

Automatic Detection

The chatbot automatically detects user role based on:

  1. Authentication Status: Checks if user is logged in
  2. User Type: Checks current_user.is_admin()
  3. Context: Uses session and user data

Role Assignment

if current_user.is_authenticated:
    if current_user.is_admin():
        role = "admin"  # Management Mode
    else:
        role = "patient"  # Patient Mode
else:
    role = "patient"  # Default to Patient Mode

Comparison Table

Feature Patient Mode Management Mode
Book Appointments ✅ Yes ❌ No
Check Status ✅ Yes (own) ❌ No
Estimated Time ✅ Yes ❌ No
Precautions ✅ Yes ❌ No
Find Doctors ✅ Yes ✅ Yes
Wait Times ✅ Yes ✅ Yes
Queue Statistics ❌ No ✅ Yes
Today's Summary ❌ No ✅ Yes
Dept Performance ❌ No ✅ Yes
Doctor Availability ❌ No ✅ Yes
High-Risk Alerts ❌ No ✅ Yes
No-Show Predictions ❌ No ✅ Yes
Crowd Forecast ✅ Limited ✅ Detailed

Testing the Role-Based System

Test as Patient

  1. Login as patient: test@patient.com / test123
  2. Open chatbot
  3. Try these commands:
    • "Hello"
    • "Book appointment"
    • "Check my status"
    • "What's my estimated time?"
    • "Precautions for cardiology"
    • "Find a doctor"

Test as Admin

  1. Login as admin: admin@hospital.com / admin123
  2. Open chatbot
  3. Try these commands:
    • "Hello"
    • "Queue statistics"
    • "Today's summary"
    • "High-risk patients"
    • "Department performance"
    • "Doctor availability"
    • "No-show predictions"
    • "Crowd forecast"

Technical Implementation

Files Modified/Created

Created:

  • ✅ app/services/chatbot_handlers.py - Patient and Management handlers

Modified:

  • ✅ app/services/chatbot_service.py - Role-based routing
  • ✅ app/routes/chatbot.py - User role detection

Key Classes

class HospitalChatbot:
    - process_message() - Main entry point
    - _handle_patient_intent() - Route patient requests
    - _handle_management_intent() - Route management requests
    - _detect_intent() - Role-based intent detection

class PatientHandlers:
    - handle_estimated_time() - Appointment timing
    - handle_precautions() - Health advice

class ManagementHandlers:
    - handle_queue_stats() - Queue analytics
    - handle_today_summary() - Daily reports
    - handle_high_risk_patients() - Priority alerts

Benefits

For Patients

✅ Personalized assistance for their specific needs ✅ Estimated timing reduces uncertainty ✅ Precautions help them prepare properly ✅ Easy booking process ✅ Status tracking without calling

For Management

✅ Real-time analytics for decision making ✅ Performance metrics for optimization ✅ Risk alerts for proactive management ✅ Predictive insights for planning ✅ Operational efficiency improvements

Future Enhancements

Phase 2 (Potential)

  1. Patient Mode:

    • Complete booking through chat
    • Payment integration
    • Medical history access
    • Prescription reminders
    • Lab report notifications

  2. Management Mode:

    • Staff scheduling assistance
    • Resource allocation suggestions
    • Financial analytics
    • Patient satisfaction metrics
    • Automated report generation

Summary

The role-based chatbot system provides:

✅ Two distinct modes for different user types ✅ 8 patient features for personal assistance ✅ 7 management features for operations ✅ Automatic role detection based on login ✅ Context-aware responses for each role ✅ Real-time data integration with ML models ✅ Production-ready implementation

This enhancement makes the chatbot more useful and relevant to each user type, improving overall system efficiency and user satisfaction.