Performance Evaluation (IVDR) is the continuous, documented process by which an IVD manufacturer assesses data to demonstrate a device’s scientific validity, analytical performance, and clinical performance for its intended purpose. Required under Article 56 and Annex XIII of Regulation (EU) 2017/746, it produces the clinical evidence that supports CE marking.
What is Performance Evaluation (IVDR)?
Performance Evaluation (IVDR) is the in vitro diagnostic (IVD) counterpart to clinical evaluation under the Medical Device Regulation (MDR). Where clinical evaluation asks whether a device delivers its intended clinical benefit, performance evaluation asks a narrower set of questions: does the test measure what it claims to measure, how accurately, and does its output actually support the clinical decision the manufacturer describes?
The process is defined in Article 56 and structured by Annex XIII of Regulation (EU) 2017/746, the IVDR. It is not a one-time gate. Manufacturers establish it before placing a device on the market and keep it current across the product’s life through post-market data.
Why Performance Evaluation (IVDR) matters in medical device development
A weak performance evaluation is now a direct path to a Notified Body finding or a refused CE certificate. Under the older IVDD, a Class C equivalent could often reach the market on published literature and in-house data with limited external review. The IVDR closed that gap. A structured performance evaluation plan, formal data across all three pillars, and a Notified Body-assessed performance evaluation report are expected for the great majority of devices.
The stakes go beyond paperwork. An IVD result drives a treatment decision, so an overstated sensitivity claim or a poorly defined intended population can translate into patient harm and field action. Gaps also surface late: rework discovered during conformity assessment delays launch and burns engineering budget that should have gone into the next product.
How Performance Evaluation (IVDR) works
Annex XIII Part A frames performance evaluation as a planned cycle built on three pillars of evidence:
Scientific validity
The association between the analyte (or marker) and the target clinical condition or physiological state. This is usually supported by literature, consensus, and prior data, not by the manufacturer’s own assay.
Analytical performance
How well the device detects or measures the analyte: trueness, precision, sensitivity, specificity, limit of detection, measuring range, and similar characteristics, addressed under Annex I, Section 9, and Annex II, Section 6.
Clinical performance
The device’s ability to produce a result that correlates with the clinical condition in the intended population, expressed through measures such as diagnostic sensitivity and specificity and predictive values.
The work runs in a defined sequence. Manufacturers first write a Performance Evaluation Plan (PEP) that sets the intended purpose, target population, and acceptance criteria. They then gather data across the three pillars, generating performance studies under Articles 57 to 77 where existing evidence is not enough. Finally, they assess and report the data in a Performance Evaluation Report (PER). Risk management under ISO 14971 and the General Safety and Performance Requirements in Annex I tie into the same evidence base. MDCG 2022-2 gives practical guidance on what sufficient clinical evidence looks like.
Common challenges and best practices
The frequent failure is vague language. “Good sensitivity” with no numeric threshold does not satisfy the analytical requirements of Annex I, and an unclear intended purpose misaligns study data with real-world use. Define measurable acceptance criteria in the plan before generating data, not after.
A second trap is selective evidence. Leaving out unfavorable literature or contradictory study results undermines scientific validity and rarely survives Notified Body scrutiny. Strong files document the search strategy and include the data that complicates the claim.
Teams also treat the PER as a deliverable that ends at certification. It does not. Part B of Annex XIII requires Post-Market Performance Follow-up (PMPF), a continuous process that feeds new field data back into the evaluation. For higher-risk Class C and D devices, the report is expected to be updated as needed, typically at least annually. Build the PEP, PER, and PMPF plan as one connected system from the start, and keep version control tight so the documents never drift apart.
How SJML helps with Performance Evaluation (IVDR)
SJML’s QARA team supports IVD manufacturers through the full performance evaluation lifecycle under Regulation (EU) 2017/746. That includes building the Performance Evaluation Plan, running scientific literature reviews, and structuring scientific validity, analytical performance, and clinical performance data into a Performance Evaluation Report ready for Notified Body assessment. The team also plans post-market performance follow-up and post-market surveillance so the clinical evidence stays current after CE marking. As an end-to-end medical device CDMO serving the diagnostics and IVD segment, SJML can align performance evaluation with design, risk management (ISO 14971), and quality system work under one roof.
Frequently asked questions
Both demonstrate clinical evidence, but they answer different questions. Clinical evaluation under MDR focuses on whether a device achieves its intended clinical benefit and acceptable residual risk. Performance evaluation under IVDR focuses on whether an in vitro diagnostic measures what it claims, how accurately, and whether that result supports the stated clinical decision, assessed across scientific validity, analytical performance, and clinical performance.
The three pillars are scientific validity, analytical performance, and clinical performance.
Scientific validity links the analyte to a clinical condition.
Analytical performance covers how well the device detects or measures that analyte, including trueness, precision, and limit of detection.
Clinical performance shows the device produces results that correlate with the condition in the intended population.
All three must be addressed together in the Performance Evaluation Report.
Yes. Article 5(3) of the IVDR requires every IVD, regardless of class, to meet the performance requirements of Article 56 and Annex XIII Part A. The depth scales with risk: a Class A device needs a proportionate evaluation, while Class C and D devices face formal performance studies and Notified Body review of the report. No device is exempt from having a documented performance evaluation.
The Performance Evaluation Report is living documentation. It must be updated whenever new information from post-market performance follow-up, post-market surveillance, or the literature affects the conclusions. For higher-risk Class C and D devices, manufacturers are generally expected to review and update the report at least annually when warranted, keeping the clinical evidence aligned with real-world use.
Related terms
- Clinical Evaluation (MDR)
- Post-Market Performance Follow-up (PMPF)
- In Vitro Diagnostic Regulation (IVDR)
- General Safety and Performance Requirements (GSPR)
- ISO 14971
