npj digital medicine | Perspective
CORE-MD Clinical Risk Score for AI Medical Devices
The CORE-MD consortium proposes a novel risk scoring system to guide regulatory evaluation of AI-based medical device software, balancing innovation with patient safety.
Executive Impact Summary
This framework streamlines regulatory pathways, ensuring that AI medical devices are rigorously evaluated while facilitating timely market access for beneficial innovations.
0%
Experts supported risk-benefit approach
0%
Experts supported scoring system
0%
Experts on low-risk market access
Deep Analysis & Enterprise Applications
Select a topic to dive deeper, then explore the specific findings from the research, rebuilt as interactive, enterprise-focused modules.
Introduction to CORE-MD
Evaluation Process Overview
Risk Score Criteria
Regulatory Impact
Practical Application
Context and Objectives of CORE-MD
The European CORE-MD consortium (Coordinating Research and Evidence for Medical Devices) aims to provide methodological principles for the clinical evaluation of AI MDSW throughout its full life cycle. This initiative addresses the need for clear guidelines to balance the potential benefits of AI in healthcare with the risks of misuse and negative societal effects. The project emphasizes a risk-benefit approach for pre- and post-release phases, incorporating both regulatory and end-user perspectives. This article presents the final recommendations from this extensive collaborative effort.
Enterprise Process Flow
Plan and design: audit and impact assessment
→
Data and Input: collect and process data
→
AI model: build and use
→
AI model: verify and validate
→
Deploy and integrate
→
Pilot evaluation
→
Comparative evaluation
→
Long term operation and monitoring
CORE-MD AI Risk Score Components
| Criterion and explanation | Level | Score |
|---|---|---|
| Valid clinical association score | Strong association with easy human oversight and full transparency | 1 |
| Moderate association with difficult human oversight and incomplete transparency | 2 | |
| Weak association without the possibility for human oversight and absent transparency | 3 | |
| Valid technical performance score | Strong with broad external validation | 1 |
| Moderate with narrow external validation | 2 | |
| Weak with only internal validation | 3 | |
| Clinical performance score: Context of use (Non-serious/Serious/Critical) | Non-serious | 1 |
| Serious | 2 | |
| Critical | 3 | |
| Clinical performance score: Medical function (Inform/Drive/Diagnose or treat) | Inform | 1 |
| Drive | 2 | |
| Diagnose or treat | 3 |
The CORE-MD AI Risk Score is composed of three parts: Valid Clinical Association Score (VCAS), Valid Technical Performance Score (VTPS), and Clinical Performance Score (CPS). These scores determine the required level of clinical evidence.
Total Risk Score for Pre-Market Evaluation
≤ 7
Total score for proportionate lower-level pre-market clinical evaluation
A total score of 7 or less (CPS + VTPS + VCAS) suggests a lower risk, allowing for less extensive pre-market clinical evaluation, balanced with more post-market evidence. Higher scores necessitate more rigorous clinical investigations before approval.
Case Study: Low-Risk AI Diagnostic Tool
Consider an AI tool designed to inform a non-serious condition, validated with broad external data, and offering full transparency and easy human oversight. This tool would likely achieve a low CORE-MD risk score.
Outcome: Faster Market Access & Reduced Burden Such a device, if it meets the low-risk criteria, could gain market access with a streamlined clinical evaluation process, significantly reducing time-to-market and regulatory burden, while ensuring ongoing post-market surveillance.
Calculate Your AI ROI Potential
Understand the potential efficiency gains and cost savings by integrating AI into your enterprise workflows, using our advanced ROI calculator.
Estimated Annual Savings
$0
Total Hours Reclaimed Annually
0
Your AI Implementation Roadmap
Our phased implementation approach ensures a smooth transition and continuous optimization of AI solutions within your organization.
Phase 1: Strategic Alignment
Define clear objectives, assess current infrastructure, and conduct a thorough impact assessment to align AI strategy with business goals.
Phase 2: Data & Model Development
Gather, clean, and prepare data. Develop or select appropriate AI models, ensuring robust training and validation with representative datasets.
Phase 3: Integration & Pilot
Seamlessly integrate AI tools into existing workflows, conduct pilot evaluations, and gather user feedback for iterative improvements.
Phase 4: Scaling & Monitoring
Deploy AI solutions at scale, establish continuous performance monitoring, and implement mechanisms for managing model drift and updates.
Ready to Transform Your Enterprise with AI?
Book a free consultation with our AI strategists to discuss your specific needs and how the CORE-MD principles can guide your AI adoption.
Ready to Get Started?
Book Your Free Consultation.
Let's Discuss Your AI Strategy!
Lets Discuss Your Needs
Select a Date
Sun
Mon
Tue
Wed
Thu
Fri
Sat