Incident Response

Table of Contents

• Incident vs. Bug Classification

• Incident Types and Severity

  • Core Features

• Response Team Structure

• Incident Response Process

  • Emergency Contacts

  • Escalation Path

• Response Time Commitments

• Post-Incident Actions

• Communication Channels

• Preventive Measures

At times, users may experience issues with the RadPair platform that could impact their medical reporting workflow. Some issues may severely impact clinical operations. These are classified as Incidents.

RadPair has a streamlined process to address incidents. This document explains our incident response process to ensure quick resolution and minimal impact on clinical operations.

Incident vs. Bug Classification

Not all issues are incidents. Here's how RadPair differentiates incidents from bugs:

Issue Type	Description	Severity	Response Time
Incident	Severely impacts clinical operations and radiology workflow. See Types below.	High	Immediate
Bug	Affects individual features or specific reports without impacting overall operations.	Medium, Low	Standard Support SLA

Incident Types and Severity

RadPair classifies incidents based on impact level and affected functionality:

Incident	< 1% Users	1% - 5% Users	5% - 20% Users	> 20% Users
Core Speech-to-Text Broken	Major	Major	Critical	Critical
Report Generation Failed	Major	Major	Critical	Critical
Performance Severely Degraded	Minor	Major	Critical	Critical
Non-core Feature Broken	Minor	Minor	Major	Major

Core Features

The following features are considered core to RadPair's functionality:

• Speech-to-Text transcription

• Report generation and templating

• Critical AI features

• Core system functionality

These features receive priority attention during incidents as they directly impact clinical operations.

Response Team Structure

RadPair maintains:

• 24/7 on-call operations team

• 24/7 on-call engineering team

• Senior leadership availibility

Our monitoring solution covers end to end system health and proactively alerts engineers of system degradation or failures.

Incident Response Process

Our incident response follows a structured process:

This diagram illustrates our incident response workflow from detection to resolution.

Emergency Contacts

For urgent issues, contact the following:

Role	How to Reach Them	Expected Response Time
On-Call Ops/Eng	[phone]	Immediate
Radpair Support Email	[email protected]	15 minutes

Escalation Path

If an issue is not resolved in a timely manner, follow this escalation path:

1. Level 1: On-Call Operations (Goal: Responds immediately)

   • Initial assessment and basic troubleshooting

2. Level 2: On-Call Engineering (Goal: Response within 15 minutes)

   • Deeper technical investigation and service recovery

3. Level 3: Senior leadsership (Goal: Response within 1 hour)

   • Critical business decisions and external communication

Response Time Commitments

Response Action	Critical	Major	Minor
Acknowledge	Within 5 minutes	Within 15 minutes	Within 4 hours
Begin Investigation	Immediate	Within 15 minutes	Within 8 hours
Status Updates	Every 30 minutes	Every hour	Daily
Target Resolution	Within 4 hours	Within 8 hours	Within 48 hours
Post-Resolution Brief	Within 1 hour	Within 4 hours	Within 24 hours

Post-Incident Actions

After resolving an incident, RadPair:

1. Conducts immediate root cause analysis

2. Prepares detailed incident report including:

   • Incident timeline

   • Impact assessment

   • Root cause identification

   • Resolution steps taken

   • Preventive measures implemented

3. Implements preventive measures

4. Updates monitoring systems if needed

5. Communicates findings to affected users

Communication Channels

During active incidents:

• Phone support for critical and major incidents (24/7): [phone]

• Email notifications to affected users

• Direct communication through [email protected]

Preventive Measures

RadPair continuously:

• Monitors system performance metrics

• Conducts regular system health checks

• Performs proactive maintenance

• Updates AI models and speech recognition systems

• Implements security patches and updates