Section 5
Clinical Evaluation and Performance Studies
Understanding the evidence requirements for demonstrating safety and performance of medical devices and IVDs under MDR and IVDR.
The Foundation of Device Safety
Clinical evaluation is the systematic and planned process to continuously generate, collect, analyse, and assess clinical data pertaining to a medical device. Under the MDR, clinical evaluation is not a one-time event but a continuous process throughout the device lifecycle. It forms the foundation upon which manufacturers demonstrate that their devices are safe and achieve their intended performance.
For healthcare workers, understanding clinical evaluation requirements is essential for several reasons: it informs procurement decisions, supports clinical risk assessment, and helps identify devices where evidence may be insufficient. The enhanced clinical evidence requirements under MDR represent one of the most significant changes from the previous directive regime.
Similarly, the IVDR introduces performance evaluation and performance studies as the equivalent processes for in vitro diagnostic devices, reflecting the unique nature of diagnostic testing where accuracy, sensitivity, specificity, and clinical utility must be demonstrated.
Clinical Evaluation Under MDR
Article 61 of the MDR establishes the requirement for clinical evaluation. Every medical device must undergo clinical evaluation to confirm conformity with the relevant general safety and performance requirements set out in Annex I. The clinical evaluation must be thorough and objective, taking into account any relevant harmonised standards and the state of the art.
MDR Clinical Evaluation Requirements (Article 61)
- 1Based on clinical data providing sufficient clinical evidence to demonstrate conformity with general safety and performance requirements
- 2Must be planned, continuously updated with data from post-market surveillance and post-market clinical follow-up (PMCF)
- 3Clinical data must be appropriate considering device characteristics and intended purpose
- 4Must address residual risks and side effects, including benefit-risk analysis
Sources of Clinical Data
The MDR defines clinical data as information concerning safety or performance generated from the use of a device. This data can come from multiple sources, each with different evidentiary weight and applicability.
Clinical Investigations
Systematic investigations in human subjects undertaken to assess safety and/or performance of a device. This is the gold standard for clinical evidence, particularly for novel devices.
When required: Class III devices, implantables (unless justified exception), novel technologies, new intended purposes.
Clinical Data from Equivalent Devices
Data generated from devices demonstrated to be equivalent. Under MDR, equivalence requirements are significantly stricter than under MDD.
Limitations: Requires contract with equivalent device manufacturer for Class III/implantables.
Clinical Data from Literature
Published scientific literature concerning the safety, performance, design characteristics, and intended purpose of the device or equivalent devices.
Requirements: Systematic literature review following defined protocol, critical appraisal of data quality.
Post-Market Clinical Data
Data from post-market surveillance, PMCF studies, registries, and real-world evidence gathered after market placement.
Purpose: Continuous update of clinical evaluation, identification of emerging risks, long-term performance confirmation.
The Clinical Evaluation Report (CER)
The Clinical Evaluation Report is the documented output of the clinical evaluation process. Under MDR Annex XIV Part A, the CER must follow a defined procedure and be updated throughout the device lifecycle. For healthcare workers, the CER is a key document for assessing the evidence base supporting a device.
CER Required Contents (Annex XIV Part A)
Device Description and Specifications
Complete description including intended purpose, indications, contraindications, target patient population, and intended users.
Clinical Evaluation Plan
Documented methodology for the clinical evaluation including scope, literature search strategy, and appraisal criteria.
Clinical Data Analysis
Systematic analysis of all relevant clinical data, including data from investigations, literature, and post-market sources.
Equivalence Demonstration (if applicable)
Technical, biological, and clinical equivalence demonstration with justification and contractual access evidence.
Safety and Performance Conclusions
Conclusions on conformity with GSPRs, identification of residual risks, and benefit-risk determination.
PMCF Plan
Plan for ongoing collection of clinical data post-market, including rationale for any exemption from PMCF.
CER Update Frequency
| Device Class | Update Frequency |
|---|---|
| Class III and Implantables | At least annually |
| Class IIa and IIb | At least every 2-5 years (depending on risk) |
| Class I | When necessary based on PMS data |
The Equivalence Challenge Under MDR
One of the most significant changes under MDR is the enhanced requirements for demonstrating equivalence. Under the MDD, manufacturers could rely on literature data from similar devices without formal demonstration of equivalence or access to the equivalent device's technical documentation. This is no longer possible under MDR.
Three Pillars of Equivalence (MDR Annex XIV Part A, Section 3)
Technical Equivalence
- Same design and specifications
- Similar conditions of use
- Similar relevant critical performance specifications
- Same deployment methods
Biological Equivalence
- Same materials in contact with same human tissues/fluids
- Same duration and frequency of contact
- Similar release characteristics for substances
Clinical Equivalence
- Same clinical condition or purpose
- Same site in the body
- Similar population (including age, anatomy, physiology, disease stage)
- Same kind of user
- Similar relevant critical performance in similar conditions of use
Critical Requirement: Contractual Access
For Class III devices and implantable devices, where the manufacturer claims equivalence to a device not manufactured by them, MDR Article 61(5) requires:
- A contract with the manufacturer of the equivalent device granting access to technical documentation
- The original clinical evaluation clearly demonstrates conformity with relevant GSPRs
- Evidence that the original manufacturer consented to full access to the technical documentation on an ongoing basis
This effectively means that for high-risk devices, manufacturers cannot claim equivalence to competitor devices without formal agreement, significantly limiting the equivalence pathway.
Clinical Investigations Under MDR
Clinical investigations are systematic studies in human subjects undertaken to assess the safety or performance of a device. Chapter VI of the MDR (Articles 62-82) establishes comprehensive requirements for conducting clinical investigations in the EU.
When Clinical Investigations Are Required
- MandatoryClass III devices (unless justified by existing clinical data from equivalent devices with contractual access)
- MandatoryImplantable devices (unless justified exception per Article 61(6))
- LikelyDevices with novel technology where existing data is insufficient
- LikelyDevices with new intended purpose not covered by existing evidence
- Case-by-caseClass IIa/IIb devices where literature and PMS data are insufficient
| Investigation Type | Purpose | Regulatory Pathway |
|---|---|---|
| Pre-market CI | Generate clinical data for CE marking | Article 62-81 requirements apply; Ethics Committee and Competent Authority approval required |
| PMCF Study | Ongoing data collection post-market | Follows same principles but may have simplified requirements for CE-marked devices |
| First-in-Human | Initial use of novel device in humans | Strictest oversight; must demonstrate adequate non-clinical evidence first |
| Pivotal Study | Definitive safety/performance demonstration | Typically randomised controlled trial with statistical endpoints |
Post-Market Clinical Follow-Up (PMCF)
PMCF is a continuous process that updates the clinical evaluation and is part of post-market surveillance. The MDR significantly strengthens PMCF requirements compared to the MDD, making it a mandatory and ongoing obligation for most devices.
PMCF Objectives (Annex XIV Part B)
- 1Confirm safety and performance throughout expected lifetime of the device
- 2Identify previously unknown side effects and monitor identified side effects and contraindications
- 3Identify and analyse emergent risks on the basis of factual evidence
- 4Ensure continued acceptability of the benefit-risk ratio
- 5Identify possible systematic misuse or off-label use to feed into risk management
PMCF Methods
PMCF Studies
Prospective clinical studies designed to collect specific safety and performance data post-market.
Registry Studies
Participation in or establishment of device registries collecting standardised outcome data.
User Surveys
Structured surveys of healthcare professionals and patients on device performance and safety.
Literature Review
Ongoing systematic review of published literature relating to the device and similar devices.
Complaint Analysis
Systematic analysis of complaints and adverse event reports for clinical trends.
Real-World Data
Analysis of healthcare databases, insurance claims, and other real-world data sources.
Performance Evaluation Under IVDR
For in vitro diagnostic devices, the IVDR establishes performance evaluation as the equivalent of clinical evaluation under the MDR. Performance evaluation is the assessment and analysis of data to establish or verify the scientific validity, analytical performance, and clinical performance of an IVD.
Scientific Validity
The association of an analyte with a clinical condition or physiological/pathological state.
Example: Is HbA1c a valid marker for diabetes management? Is troponin a valid marker for myocardial infarction?
Analytical Performance
The ability of a device to correctly detect or measure a particular analyte.
Parameters: Accuracy, precision, sensitivity, specificity, linearity, measuring range, limit of detection/quantification.
Clinical Performance
The ability to yield results correlated with a particular clinical condition or physiological/pathological state.
Parameters: Diagnostic sensitivity, diagnostic specificity, positive/negative predictive values, likelihood ratios.
Performance Evaluation Report (PER)
Similar to the CER for medical devices, the IVDR requires a Performance Evaluation Report that documents the performance evaluation process and conclusions. The PER must address:
- •Scientific validity of the intended purpose and target analyte
- •Analytical performance characteristics with supporting data
- •Clinical performance evidence from performance studies
- •Benefit-risk analysis specific to the diagnostic context
- •Post-market performance follow-up (PMPF) plan
Performance Studies Under IVDR
Performance studies are investigations undertaken to establish or confirm the analytical or clinical performance of a device. The IVDR distinguishes between studies that use only leftover specimens (no direct patient contact) and those involving additional specimen collection.
| Study Type | Description | Regulatory Requirements |
|---|---|---|
| Leftover Specimen Studies | Studies using residual specimens from routine clinical care, with no additional intervention | May be exempt from some requirements; informed consent/ethics requirements vary by Member State |
| Prospective Specimen Collection | Studies requiring additional specimen collection from subjects for the study purpose | Full IVDR performance study requirements; Ethics Committee and informed consent required |
| Companion Diagnostic Studies | Studies to validate companion diagnostics alongside drug/therapy development | Coordination with medicinal product clinical trials; complex regulatory pathway |
| PMPF Studies | Post-market studies to confirm ongoing performance | Part of post-market surveillance; may use routine laboratory data or prospective collection |
Class D IVD Requirements
For Class D IVDs (highest risk, including blood screening and transfusion safety), the IVDR requires that clinical performance studies are conducted in the EU unless the manufacturer can demonstrate that conducting the study outside the EU is justified and that the study results are transferable to the EU population. This has significant implications for diagnostic manufacturers serving the EU market.
Role of Healthcare Facilities in Clinical Evidence
Healthcare facilities play a crucial role in the generation and evaluation of clinical evidence. Understanding this role helps healthcare workers contribute effectively to device safety and supports informed procurement decisions.
Participating in Clinical Investigations
Healthcare facilities may be invited to participate in pre-market clinical investigations or PMCF studies. Participation involves:
- Review and approval through local ethics committee/IRB
- Informed consent processes for study subjects
- Compliance with study protocol and GCP requirements
- Adverse event reporting to sponsor and ethics committee
- Data collection and documentation requirements
Contributing to Registries
Device registries collect standardised outcome data that supports post-market surveillance. Healthcare facilities may be required or encouraged to participate in:
- National implant registries (orthopaedic, cardiac, etc.)
- Disease-specific registries with device data collection
- Manufacturer-sponsored registries for specific devices
- Multi-centre outcome databases
Vigilance Reporting
Healthcare facilities contribute to clinical evidence through systematic reporting of incidents and adverse events. This data feeds into manufacturers' post-market surveillance and may identify safety signals.
Procurement Clinical Assessment
Healthcare facilities should assess clinical evidence as part of device procurement. Key questions include:
- What clinical evidence supports this device's claims?
- Was the evidence generated in relevant patient populations?
- Are there published clinical studies or registry data available?
- What are the manufacturer's PMCF plans for this device?
- Are there any field safety corrective actions or recalls for this device?
Case Study: Hospital Participation in PMCF Registry
Situation: A cardiac surgery department in a Belgian hospital was approached by a heart valve manufacturer to participate in a European PMCF registry for a transcatheter aortic valve implantation (TAVI) device. The registry aimed to collect 5-year follow-up data on approximately 10,000 patients across the EU.
Clinical Evidence Context:
- The TAVI device was Class III, requiring annual CER updates
- Pre-market clinical investigation data covered 2-year follow-up
- Long-term durability data (5+ years) was limited
- Registry participation would contribute to mandatory PMCF requirements
Hospital Considerations:
- Data collection burden on clinical staff estimated at 30 minutes per patient follow-up
- Ethics committee approval required for registry participation
- Patient consent needed for data sharing with manufacturer
- IT infrastructure needed to support secure data transfer to registry
- Hospital would receive benchmarking data on outcomes versus other centres
Outcome: The hospital joined the registry, implementing standardised data collection into their routine follow-up pathway. After three years, registry data identified a higher-than-expected paravalvular leak rate in certain patient anatomies, leading to updated patient selection guidance and implantation technique training across participating centres.
Case Study: Laboratory Performance Evaluation for LDT
Situation: A clinical chemistry laboratory in a Danish university hospital developed an in-house assay (Laboratory Developed Test, LDT) for a rare metabolic disorder. Under the IVDD, such LDTs were largely exempt from regulatory requirements. The IVDR changed this landscape significantly.
IVDR Requirements Analysis:
- The LDT fell within IVDR scope as an IVD used within a health institution
- Article 5(5) exemption conditions had to be evaluated
- If exemption did not apply, full IVDR compliance including performance evaluation was required
Performance Evaluation Approach:
- Scientific Validity: Literature review confirming biomarker association with the metabolic disorder
- Analytical Performance: Validation studies including precision, accuracy against reference method, linearity, and interference testing
- Clinical Performance: Retrospective analysis of 150 patient samples with confirmed diagnosis status to determine sensitivity and specificity
Documentation: The laboratory created a Performance Evaluation Report following IVDR Annex XIII requirements, including a PMPF plan for ongoing monitoring of assay performance through external quality assessment participation and periodic revalidation.
Key Learning: The laboratory found that formalising their performance evaluation process, while initially time-consuming, improved documentation quality and provided evidence of test reliability for clinical staff and accreditation bodies.
Case Study: AI Radiology Software Clinical Evidence Assessment
Situation: A Spanish hospital radiology department was evaluating an AI-powered chest X-ray triage system for emergency department use. The software was CE-marked under MDR as Class IIa (decision support for non-critical conditions). The hospital's clinical governance team was tasked with assessing the clinical evidence.
Clinical Evidence Review:
- Manufacturer's CER Summary: Based on retrospective analysis of 50,000 images from three datasets
- Performance Claims: 95% sensitivity for pneumonia detection, 92% for pneumothorax
- Study Population: Mixed adult population, dataset from US and Asian hospitals
- Limitations Identified: No prospective clinical investigation data; limited paediatric data; no data from European population
Hospital Assessment Concerns:
- Could retrospective performance be replicated in prospective clinical use?
- Were the training datasets representative of the hospital's patient population?
- How would the software perform with the hospital's specific imaging equipment?
- What was the manufacturer's PMCF plan for performance monitoring?
Resolution: The hospital negotiated a pilot implementation with prospective performance monitoring. The manufacturer agreed to support a 6-month evaluation comparing AI performance against radiologist ground truth in the hospital's specific setting. This approach allowed the hospital to generate local clinical evidence while contributing to the manufacturer's PMCF programme.
Clinical Evidence Requirements Summary
| Device Class | Clinical Investigation | CER Update | PMCF Requirements |
|---|---|---|---|
| Class III | Generally required unless equivalence with contractual access | Annual | Mandatory PMCF with defined plan; typically includes PMCF studies |
| Implantables | Generally required unless exemption under Article 61(6) | Annual | Mandatory PMCF; often requires participation in implant registries |
| Class IIb | May be required depending on device novelty and available data | Every 2-5 years | Required PMCF plan; methods proportionate to risk |
| Class IIa | Generally not required if sufficient literature/PMS data available | Every 2-5 years | PMCF plan required; may rely on PMS data analysis |
| Class I | Generally not required | When needed based on PMS | PMCF may be justified as not needed for well-established devices |
Key Takeaways for Healthcare Workers
Clinical evaluation is continuous under MDR—manufacturers must update CERs regularly based on PMCF and PMS data.
Equivalence claims require contractual access to technical documentation for Class III and implantables—limiting this pathway significantly.
IVDR performance evaluation covers scientific validity, analytical performance, and clinical performance—all three must be demonstrated.
Healthcare facilities contribute to clinical evidence through registry participation, vigilance reporting, and clinical investigation support.
Procurement decisions should consider the quality of clinical evidence—ask manufacturers about CER contents and PMCF plans.
AI/software devices should have clinical evidence from relevant populations—consider local validation for novel technologies.