Settings

Configure platform preferences and review security architecture.

Data Protection

Security architecture and privacy principles

Encrypted Communication Architecture

All data transmission uses encrypted channels. Production deployment would enforce TLS 1.3+ for all connections.

Role-Based Access Architecture

Access to patient data is governed by role-based permissions, ensuring users only see data relevant to their clinical role.

Audit Logging

All access to patient data and risk assessments is logged for accountability and compliance review.

Minimum Necessary Data Principle

The system collects and processes only the data necessary for risk assessment — no excessive data retention.

Isolated Patient Dataset

This platform environment is configured with isolated local patient records. No external patient information is processed.

Production deployment would require appropriate regulatory, security, privacy, and clinical validation.

This environment does not hold HIPAA, GDPR, or other regulatory certifications.

Local Storage Only

Notification Preferences

Configure when and how you receive alerts

Urgent review alerts

Receive immediate notifications for urgent-risk patients

High-risk alerts

Get notified when a patient enters high-risk status

Daily summary

Receive a daily summary of all monitored patients

Anomaly detection alerts

Notifications when new anomalies are detected

Risk Engine Configuration

Signal weights used in risk score calculation

Heart Rate
25%
Oxygen Saturation (SpO₂)
30%
Temperature
15%
Blood Pressure
15%
Respiratory Rate
10%
Activity Level
5%

These weights represent parameters configured for the clinical diagnostic support engine. A production deployment requires specific site clinical calibration and validation.

AI Architecture

Local deterministic risk engine — zero API keys required

Risk scoring runs entirely in the browser using deterministic TypeScript logic.

The AI explanation interface is modular — a real LLM service can replace the local logic without UI changes.

No external API calls, no cloud dependencies, no data leaves the device.

Upchar AI is a prototype for clinical decision support. Risk scores are not medical diagnoses and should be reviewed by qualified healthcare professionals.