# Business & Functional Requirements

## 1. Purpose
This document defines the detailed business and functional requirements for the AI-enhanced Cardio Thoracic Surgery patient intake system that enables Doctor's staff to conduct structured patient evaluations and provides cardiothoracic surgeons with comprehensive AI-generated insights within 2-3 minute consultation windows.

## 2. Scope
- **In-Scope:**  
    - Cardio Thoracic Surgery department MVP deployment at single hospital facility
    - AI-guided patient intake workflow with handheld device support
    - Real-time senior consultation alerting (non-diagnostic, non-emergency classification)
    - Doctor AI query interface for symptom analysis and patient insights
    - Multi-language audio transcription to English text capabilities
    - Session resumption after device/network interruptions
    - Audit logging and session tracking with 48-hour retention
    - Patient ID-based daily identification system
    - Daily reporting capabilities for patient encounters
    - Multi-user concurrent access (1-5 Doctor's staff)
    - Administrative management module for staff onboarding and system configuration
    - Doctor training and configuration module for protocol customization and alert pattern definition
- **Out-of-Scope:**  
    - Medical diagnosis recommendations or treatment suggestions
    - Permanent patient data storage on system servers
    - Integration with hospital EMR/EHR systems
    - Offline functionality or poor connectivity operation
    - Multi-hospital or healthcare network deployment
    - Specialties other than Cardio Thoracic Surgery (initial phase)

## 3. Stakeholders
- **Primary:** Doctor's Staff (junior doctors and medical assistants familiar with Cardio Thoracic Surgery terminology who conduct AI-guided patient intake sessions)
- **Secondary:** Cardiothoracic Surgeons (primary consultant doctors who review AI-generated patient summaries and ask follow-up questions)
- **Tertiary:** Team Leads (leadership responsible for onboarding/offboarding users, configuring system parameters, and training the system with department-specific protocols)

## 4. Key Use Cases

### **Use Case UC-001: Standard AI-Guided Patient Intake**
**Description:** Doctor's staff conducts a structured patient intake session using handheld device with AI guidance.  
**Actors:** Doctor's Staff, Patient, AI System  
**Preconditions:** Doctor's staff logged in, device connected to network.  
**Main Flow:**

```mermaid
flowchart TD
    A[Doctor's staff starts intake session] --> B[Session recording activated and<br/>System prompts for patient information]
    B --> C[AI presents initial Cardio Thoracic Surgery questions]
    C --> D[Doctor's staff asks questions to patient]
    D --> E{Patient response method?}
    E -->|Audio input| F[Patient responds via audio]
    E -->|Text input| G[Doctor's staff types English response]
    F --> H[System transcribes regional language to English]
    G --> I[Response captured in English]
    H --> I
    I --> J[AI analyzes response and determines next question]
    J --> K{More questions needed?}
    K -->|Yes| D
    K -->|No| L[System generates intake summary]
    L --> M[Doctor's staff reviews and confirms]
    M --> N[Recording stops, Session completed and saved]
    
    %% Alternate flows
    K --> O[Network interruption detected]
    O --> P[System auto-saves progress]
    P --> Q[Connection restored]
    Q --> R[Session resumed from last checkpoint]
    R --> K
    
    I --> S[Concerning pattern detected]
    S --> T[System alerts: Suggest senior consultation]
    T --> J
    
    classDef user fill:#E1F5FE,stroke:#0277BD,stroke-width:2px
    classDef system fill:#FFF3E0,stroke:#F57C00,stroke-width:2px
    classDef alert fill:#FFEBEE,stroke:#C62828,stroke-width:2px
    
    class A,D,G,M user
    class B,C,F,H,I,J,L,N,P,R system
    class O,S,T alert
```

**Alternate Flows:**
- Network interruption: System auto-saves progress every 30 seconds, resumes from exact checkpoint
- Device power loss: Session data preserved, resumable upon restart
- Audio input failure: Doctor's staff can switch input sources or enter text manually
- Patient continuation: When same daily ID entered by different Doctor's staff, system displays previously captured details, shows last interaction status, and allows continuation of data capture without creating new patient intake
- Patient unable to respond: Doctor's staff types or speaks whatever the patient conveyed ("patient doesn't recall", "NA") and submits normally — Claude continues to the next question

### **Use Case UC-002: Senior Consultation Alert Scenario**
**Description:** AI system identifies concerning symptom patterns and suggests Doctor's staff consult senior/superior staff for evaluation.  
**Actors:** Doctor's Staff, AI System, Senior Doctor's Staff  
**Preconditions:** Patient intake session in progress, AI analysis enabled.  
**Main Flow:**

```mermaid
flowchart TD
    A[AI analyzes patient responses] --> B[Concerning pattern detected]
    B --> C[System generates alert within 30 seconds]
    C --> D[Alert displayed to Doctor's staff]
    D --> E[Doctor's staff reviews alert details]
    E --> F[Doctor's staff decides on action]
    F --> G[Consult senior staff immediately]
    F --> H[Continue intake with notes]
    G --> I[Senior staff reviews case]
    I --> J[Senior provides guidance]
    J --> K[Doctor's staff continues with guidance]
    H --> K
    K --> L[Intake session continues/completes]
    
    classDef user fill:#E1F5FE,stroke:#0277BD,stroke-width:2px
    classDef system fill:#FFF3E0,stroke:#F57C00,stroke-width:2px
    classDef alert fill:#FFEBEE,stroke:#C62828,stroke-width:2px
    classDef senior fill:#E8F5E8,stroke:#2E7D32,stroke-width:2px
    
    class E,F,K user
    class A,C,L system
    class B,D alert
    class G,I,J senior
```

**Alternate Flows:**
- Multiple alerts: System prioritizes most concerning patterns first
- Senior staff unavailable: Doctor's staff documents alert for later review
- False alert: Doctor's staff can provide feedback to improve AI accuracy

### **Use Case UC-003: Doctor Intelligence Query**
**Description:** Cardiothoracic Surgeon requests AI analysis and insights during patient consultation.  
**Actors:** Cardiothoracic Surgeon, AI System  
**Preconditions:** Patient intake completed, doctor reviewing patient summary.  
**Main Flow:**

```mermaid
flowchart TD
    A[Doctor reviews patient summary] --> B[Doctor has specific question]
    B --> C[Doctor enters query to AI system]
    C --> D[AI analyzes patient data]
    D --> E[AI provides relevant insights]
    E --> F[Doctor reviews AI response]
    F --> G{Need more information?}
    G -->|Yes| C
    G -->|No| H[Doctor proceeds with consultation]
    
    classDef doctor fill:#E8F5E8,stroke:#2E7D32,stroke-width:2px
    classDef system fill:#FFF3E0,stroke:#F57C00,stroke-width:2px
    
    class A,B,C,F,H doctor
    class D,E system
```

**Alternate Flows:**
- No relevant data: AI indicates insufficient information for query
- Complex query: AI breaks down response into multiple focused insights

### **Use Case UC-004: Session Resumption**
**Description:** Doctor's staff resumes interrupted intake session from exact stopping point.  
**Actors:** Doctor's Staff, AI System  
**Preconditions:** Previous session interrupted, session data auto-saved.  
**Main Flow:**

```mermaid
flowchart TD
    A[Doctor's staff logs back in] --> B[System detects interrupted session]
    B --> C[System displays resumption option]
    C --> D[Doctor's staff selects resume]
    D --> E[System loads last checkpoint]
    E --> F[Session continues from exact point]
    F --> G[Intake proceeds normally]
    
    classDef user fill:#E1F5FE,stroke:#0277BD,stroke-width:2px
    classDef system fill:#FFF3E0,stroke:#F57C00,stroke-width:2px
    
    class A,D user
    class B,C,E,F,G system
```

**Alternate Flows:**
- Start new session: Doctor's staff can choose to start fresh intake instead
- Data corruption: System alerts to data issue and starts new session

### **Use Case UC-006: Administrative Management Configuration**
**Description:** Team Leads configure system parameters, onboard and offboard Doctor's staff, and set up regional language options for localized patient communication.
**Actors:** Team Lead, System  
**Preconditions:** Team Lead has system access, configuration permissions enabled.
**Main Flow:**

```mermaid
flowchart TD
    A[Team Lead logs into admin module] --> B[System displays configuration dashboard]
    B --> C{Configuration task?}
    C -->|Staff onboarding| D[Add new Doctor's staff member]
    C -->|System parameters| E[Configure system settings]
    C -->|Language setup| F[Set regional language options]
    D --> G[Enter staff details and permissions]
    G --> H[Assign role and access levels]
    H --> I[Generate login credentials]
    I --> J[Staff account created successfully]
    E --> K[Modify system parameters and values]
    K --> L[Validate configuration changes]
    L --> M[Apply settings to system]
    F --> N[Select supported regional languages]
    N --> O[Configure audio transcription settings]
    O --> P[Test language processing capabilities]
    P --> Q[Language setup completed]
    
    classDef admin fill:#F3E5F5,stroke:#7B1FA2,stroke-width:2px
    classDef system fill:#FFF3E0,stroke:#F57C00,stroke-width:2px
    
    class A,D,G,H,K,N,O admin
    class B,I,J,L,M,P,Q system
```

**Alternate Flows:**
- Invalid configuration: System validates settings and provides error messages for correction
- Permission conflicts: Team Lead can override or modify existing staff permissions
- Language testing failure: System provides diagnostic information for troubleshooting

### **Use Case UC-007: Doctor Training & Protocol Configuration**
**Description:** Cardiothoracic Surgeons customize AI questioning protocols, define critical alert patterns, and establish default AI agent sequences based on clinical expertise.
**Actors:** Cardiothoracic Surgeon, AI System  
**Preconditions:** Cardiothoracic Surgeon has configuration access, clinical protocols identified.
**Main Flow:**

```mermaid
flowchart TD
    A[Cardiothoracic Surgeon accesses training module] --> B[System displays protocol configuration interface]
    B --> C{Configuration type?}
    C -->|New protocol| D[Create new questioning protocol]
    C -->|Alert patterns| E[Define critical alert criteria]
    C -->|AI sequences| F[Configure default agent protocols]
    D --> G[Specify protocol name and purpose]
    G --> H[Add question sequences and logic]
    H --> I[Define response analysis criteria]
    I --> J[Test protocol with sample scenarios]
    J --> K[Protocol validation and activation]
    E --> L[Identify concerning symptom combinations]
    L --> M[Set alert trigger thresholds]
    M --> N[Define escalation recommendations]
    N --> O[Test alert pattern recognition]
    O --> P[Alert pattern configured and active]
    F --> Q[Design AI agent question flows]
    Q --> R[Set conditional branching logic]
    R --> S[Configure response analysis rules]
    S --> T[Validate agent sequence effectiveness]
    T --> U[AI sequence deployed to system]
    
    classDef doctor fill:#E8F5E8,stroke:#2E7D32,stroke-width:2px
    classDef system fill:#FFF3E0,stroke:#F57C00,stroke-width:2px
    
    class A,D,G,H,L,M,Q,R doctor
    class B,I,J,K,N,O,P,S,T,U system
```

**Alternate Flows:**
- Protocol conflicts: System identifies overlapping protocols and suggests resolution
- Testing failures: Cardiothoracic Surgeon can modify criteria and retest configurations
- Clinical validation required: System flags protocols requiring peer review before activation

### **Use Case UC-005: Multi-User Concurrent Access**
**Description:** Multiple Doctor's staff (1-5 users) access system simultaneously for different patients.  
**Actors:** Multiple Doctor's Staff, AI System  
**Preconditions:** Multiple staff logged in, sufficient system capacity.  
**Main Flow:**

```mermaid
flowchart TD
    A[Staff 1 starts Patient A intake] --> B[Staff 2 starts Patient B intake]
    B --> C[Staff 3 starts Patient C intake]
    C --> D[System manages concurrent sessions]
    D --> E[Each session operates independently]
    E --> F[System maintains individual progress]
    F --> G[Sessions complete individually]
    
    classDef user fill:#E1F5FE,stroke:#0277BD,stroke-width:2px
    classDef system fill:#FFF3E0,stroke:#F57C00,stroke-width:2px
    
    class A,B,C user
    class D,E,F,G system
```

**Alternate Flows:**
- System capacity reached: When all 5 concurrent user slots are occupied, additional Doctor's staff attempting to log in are placed in a queue and notified of their position. They receive automatic login prompts as slots become available when other staff complete their sessions.
- Resource conflict: During high system load, active intake sessions receive priority for processing power and network bandwidth over new login requests. For example, if the system is processing 4 active transcriptions and a 5th user tries to start a new session, the new session waits until existing transcription tasks complete to ensure quality is maintained for ongoing patient interactions.

## 5. Functional Requirements

1. **FR-001:** System shall guide Doctor's staff through structured Cardio Thoracic Surgery-specific questionnaires adapted based on patient responses.
2. **FR-002:** System can accept audio input from any available input sources (built-in microphone, bluetooth devices, wired connections).
3. **FR-003:** System shall transcribe audio from user's regional language to English text in real-time.
4. **FR-004:** System shall adapt questioning protocols dynamically based on patient responses and Cardio Thoracic Surgery-specific algorithms.
5. **FR-005:** System shall not impose automatic session termination, allowing Doctor's staff to capture all required patient information without time-based interruption.
6. **FR-006:** System shall analyze patient data for concerning symptom patterns using Cardio Thoracic Surgery-specific AI models.
7. **FR-007:** System shall generate alerts suggesting senior consultation without providing diagnosis or emergency classification.
8. **FR-008:** System shall provide doctors with AI query interface for symptom analysis and patient insight requests.
9. **FR-009:** System shall generate comprehensive patient summaries within 30 seconds of session completion, structured for cardiothoracic surgeon review within 1-2 minutes.
10. **FR-010:** System shall support 1-5 concurrent Doctor's staff users with independent session management.
11. **FR-011:** System shall automatically save session progress every 30 seconds of data input or after section changes.
12. **FR-012:** System shall resume interrupted sessions from the exact stopping point with complete data preservation.
13. **FR-013:** System shall maintain detailed audit logs of all sessions, user actions, and system events, capturing the specific role (Agent/Doctor/Doctor's Staff) performing each action with date and time. For example: Agent prompts question (logged with question details and timestamp), Doctor's Staff ignores/captures/dismisses agent's question (logged with staff response type and timestamp).
14. **FR-014:** System shall identify and track patients using daily patient IDs without external system integration.
15. **FR-015:** System shall provide comprehensive administrative interface for onboarding Doctor's staff, configuring system parameters and their values, and setting up regional language options for patient data input.
16. **FR-016:** System shall enable cardiothoracic surgeons to add new questioning protocols, define critical alert patterns that trigger senior consultation suggestions, and establish default AI agent protocol sequences based on clinical expertise.
17. **FR-017:** System shall validate and test new protocols and configurations before activation to ensure clinical safety and effectiveness.
18. **FR-018:** System shall support role-based access control with different permission levels for Doctor's staff, cardiothoracic surgeons, and Team Leads.
19. **FR-019:** System shall enable continuation of patient intake by different Doctor's staff within the same day, displaying previously captured details and allowing additional information capture based on the last interaction without creating a new patient intake session.

## 6. Acceptance Criteria

**FR-001 / UC-001:**  
- AC1: Doctor's staff can navigate questionnaire without training
- AC2: Questions adapt within 2 seconds of patient response entry
- AC3: Cardio Thoracic Surgery-specific protocols cover 95% of common cardiothoracic conditions
- AC4: System interface is intuitive and easy to understand, interpret and navigate with minimal to no training required for Doctor's staff

**FR-002 / UC-001:**  
- AC1: System detects and switches between input sources seamlessly
- AC2: Audio quality indicators show connection status in real-time
- AC3: Input source failure triggers automatic fallback options

**FR-003 / UC-001:**  
- AC1: Regional language transcription accuracy ≥ 90% for medical terminology
- AC2: Transcription completion within 5 seconds of audio input
- AC3: System supports minimum 5 regional languages common in deployment area

**FR-007 / UC-002:**  
- AC1: Alert generation completes within 30 seconds of pattern detection
- AC2: Alerts provide specific rationale without diagnostic language
- AC3: Doctor's staff can acknowledge or dismiss alerts with notes

**FR-009 / UC-003:**  
- AC1: Patient summary generation completes within 30 seconds
- AC2: Summary contains symptom highlights, concerning patterns, and relevant history
- AC3: Summary length suitable for 1-2 minute doctor review while ensuring no critical information is missed or omitted

**FR-011 / UC-004:**  
- AC1: Auto-save triggers every 30 seconds during active data input
- AC2: Auto-save triggers immediately when moving between questionnaire sections
- AC3: Save operations complete without interrupting user workflow

**FR-013 / All UCs:**  
- AC1: Audit logs capture all user actions with timestamps
- AC2: System events logged include session start/stop, alerts, errors
- AC3: Logs retained for exactly 48 hours then automatically purged

**FR-015 / UC-006:**  
- AC1: Administrative interface provides intuitive staff onboarding and offboarding workflow
- AC2: System parameter configuration includes validation and rollback capabilities
- AC3: If regional language support is activated, at least 1 regional language must be configured and supported

**FR-016 / UC-007:**  
- AC1: Protocol creation interface allows cardiothoracic surgeons to define custom question sequences
- AC2: Alert pattern configuration provides clinical pattern matching capabilities
- AC3: AI agent sequences can be tested with sample scenarios before activation

**FR-017 / UC-007:**  
- AC1: All new protocols undergo validation testing before clinical deployment
- AC2: Configuration conflicts detected and resolved before activation
- AC3: Safety checks prevent deployment of potentially harmful protocols

**FR-018 / UC-006:**  
- AC1: Role-based permissions restrict access based on user type (Doctor's Staff/Cardiothoracic Surgeon/Team Lead)
- AC2: Permission changes take effect immediately without system restart
- AC3: Access control audit trail tracks all permission modifications and Team Lead access & configuration changes

**FR-019 / UC-001:**  
- AC1: System displays previously captured patient details when same daily ID is entered by different Doctor's staff
- AC2: Continuation session shows last interaction status and allows additional data capture
- AC3: System treats subsequent intakes as continuation rather than new patient intake, maintaining data continuity
- AC4: System prevents concurrent data intake for the same patient by multiple Doctor's staff, allowing only one active intake session per patient daily ID at any given time

## 7. Non-Functional Requirements

- **Performance:** 
    - Audio transcription response time ≤ 5 seconds
    - Senior consultation alert generation ≤ 30 seconds  
    - Patient summary generation ≤ 30 seconds (CRITICAL)
    - System response time for standard operations ≤ 2 seconds
- **Security:** 
    - Data encryption requirements to be defined during implementation
    - Simple login authentication method (CRITICAL)
    - Session timeout after 30 minutes of inactivity
- **Reliability:** 
    - System uptime 99.9% during hospital operating hours
    - Network requirement: Decent 5G speed or equivalent bandwidth
    - Session data backup frequency: Every 30 seconds or section change (CRITICAL)
    - Audit trail retention period: 48 hours (CRITICAL)

## 8. Data Models & Entities

```mermaid
classDiagram
  class Patient {
    +String dailyId
    +DateTime intakeDate
    +String demographicInfo
    +List~String~ symptoms
    +String medicalHistory
  }

  class IntakeSession {
    +String sessionId
    +String patientDailyId
    +String medicalStaffId
    +DateTime startTime
    +DateTime endTime
    +String status
    +List~QuestionResponse~ responses
    +Boolean hasAlerts
    +Boolean isContinuation
    +String previousSessionId
    +String lastInteractionStatus
  }

  class QuestionResponse {
    +String questionId
    +String questionText
    +String audioInput
    +String transcribedText
    +DateTime timestamp
    +String responseCategory
  }

  class Alert {
    +String alertId
    +String sessionId
    +String alertType
    +String description
    +DateTime generatedAt
    +String severity
    +Boolean acknowledged
  }

  class PatientSummary {
    +String summaryId
    +String sessionId
    +String keySymptoms
    +String significantFindings
    +List~String~ alertHistory
    +DateTime generatedAt
  }

  class AuditLog {
    +String logId
    +String userId
    +String userRole
    +String sessionId
    +String action
    +DateTime timestamp
    +String details
    +String responseType
  }

  class SystemConfiguration {
    +String configId
    +String parameterName
    +String parameterValue
    +String teamLeadId
    +DateTime lastModified
    +String description
  }

  class MedicalStaff {
    +String staffId
    +String name
    +String role
    +String permissions
    +DateTime createdAt
    +String teamLeadId
    +Boolean isActive
  }

  class QuestioningProtocol {
    +String protocolId
    +String protocolName
    +String cardiothoracicSurgeonId
    +List~String~ questionSequence
    +String analysisLogic
    +DateTime createdAt
    +Boolean isActive
  }

  class AlertPattern {
    +String patternId
    +String patternName
    +String cardiothoracicSurgeonId
    +List~String~ symptomCombinations
    +String triggerThresholds
    +String escalationRecommendations
    +DateTime createdAt
    +Boolean isActive
  }

  class AIAgentSequence {
    +String sequenceId
    +String sequenceName
    +String cardiothoracicSurgeonId
    +String questionFlows
    +String conditionalLogic
    +String analysisRules
    +DateTime createdAt
    +Boolean isActive
  }

  class RegionalLanguage {
    +String languageId
    +String languageName
    +String languageCode
    +Boolean transcriptionEnabled
    +String teamLeadId
    +DateTime configuredAt
  }

  Patient "1" --> "1..*" IntakeSession : has
  IntakeSession "1" --> "1..*" QuestionResponse : contains
  IntakeSession "1" --> "0..*" Alert : generates
  IntakeSession "1" --> "1" PatientSummary : produces
  IntakeSession "1" --> "1..*" AuditLog : creates
  MedicalStaff "1" --> "0..*" IntakeSession : conducts
  QuestioningProtocol "1" --> "0..*" IntakeSession : guides
  AlertPattern "1" --> "0..*" Alert : triggers
  AIAgentSequence "1" --> "0..*" QuestionResponse : influences
  SystemConfiguration "1" --> "0..*" RegionalLanguage : configures
```

## 9. Business Rules & Constraints

- Rule 1: No patient data shall be permanently stored on system servers after session completion.
- Rule 2: Senior consultation alerts must not provide diagnosis or emergency classification language.
- Rule 3: All audio transcription must convert regional languages to English text only.
- Rule 4: Session resumption must preserve exact state including partially completed responses.
- Rule 5: Audit logs must be automatically purged after exactly 48 hours.
- Rule 6: Maximum concurrent users limited to 5 Doctor's staff to ensure system performance.
- Rule 7: Patient identification uses daily IDs only - no integration with hospital patient systems.
- Rule 8: Only Team Leads can onboard new Doctor's staff and modify system configurations.
- Rule 9: Only cardiothoracic surgeons can create, modify, or activate questioning protocols and alert patterns.
- Rule 10: All new protocols and configurations must undergo validation testing before clinical deployment.
- Rule 11: Role-based access control prevents unauthorized access to configuration interfaces.
- Rule 12: Regional language configurations can be defined and modified only by the Team Lead.
- Rule 13: Patient intake continuation by different Doctor's staff within the same day shall display previous interaction details and maintain data continuity without creating duplicate patient records.

## 10. Assumptions & Dependencies

- **Assumptions:** 
    - Doctor's staff familiar with Cardio Thoracic Surgery terminology and handheld devices
    - Hospital has reliable 5G or equivalent network infrastructure
    - Patients can communicate in supported regional languages
    - Senior Doctor's staff available for consultation during operating hours
    - Team Leads have necessary permissions to onboard, offboard, and manage system configurations
    - Cardiothoracic Surgeons willing to invest time in customizing AI protocols based on clinical expertise
    - Hospital IT security policies support role-based access control implementation
- **Dependencies:** 
    - Hospital IT infrastructure readiness and network connectivity
    - Cardio Thoracic Surgery department workflow integration approval
    - Regional language AI models availability and accuracy
    - Handheld device procurement and deployment
    - Team Lead training for system configuration and management
    - Cardiothoracic Surgeon availability for protocol development and validation
    - Hospital compliance approval for role-based access control implementation

## 11. Glossary

- **Daily Patient ID:** Unique identifier assigned to patients for single-day tracking without permanent storage
- **Senior Consultation Alert:** System suggestion for Doctor's staff to consult superior staff based on concerning patterns
- **Regional Language:** Local languages spoken by patients that require transcription to English
- **Session Checkpoint:** Automatically saved progress point allowing exact resumption after interruption
- **Cardio Thoracic Surgery-Specific Protocol:** AI questioning algorithms designed specifically for cardiothoracic condition assessment
- **Administrative Management Module:** System interface for Team Leads to onboard/offboard users, manage configurations, and manage regional language settings
- **Doctor Training & Configuration Module:** Interface for cardiothoracic surgeons to customize AI protocols, define alert patterns, and configure AI agent sequences
- **Alert Pattern:** Predefined symptom combinations that trigger senior consultation suggestions
- **AI Agent Sequence:** Structured flow of AI-driven questions and responses designed by cardiothoracic surgeons
- **Role-Based Access Control:** Security system that restricts system functions based on user roles (Doctor's Staff/Cardiothoracic Surgeon/Team Lead)
- **Patient Intake Continuation:** Feature allowing different Doctor's staff to continue patient data capture within the same day, maintaining previous interaction context and data continuity

## 12. Open Questions

- Regional language support priority: Which 5 languages should be implemented first?
- Data encryption standards: Specific healthcare compliance requirements to be determined
- Senior staff notification method: This is not handled within the system - external notification methods to be determined
- AI model training data: Availability and quality of Cardio Thoracic Surgery-specific datasets for local medical practices
- Team Lead training requirements: What level of training do Team Leads need for system configuration and management?
- Protocol validation process: What clinical review process should be required before activating new cardiothoracic surgeon-defined protocols?
- Default configuration settings: What initial system parameters and protocols should be pre-configured for immediate deployment?

## 13. Approvals
## Approval

Approved by: Prabhakar
Role: Product Owner
Date: 09-May-2026

Approved by: [Name]
Role: [Role]
Date: [Date]