Skip to main content

Performance evaluation for IVDs

Regulatory basis

Performance evaluation requirements for IVDs are in Schedule 1A (Essential Requirements for IVDs) of Part III, UK MDR 2002. They are the IVD equivalent of clinical evaluation for general medical devices.


What is performance evaluation?

Performance evaluation for IVDs is the systematic examination and documentation of the data used to assess and demonstrate that an IVD performs as intended by the manufacturer. It is the IVD equivalent of clinical evaluation.

The output is the Performance Evaluation Report (PER), which must be maintained throughout the device's lifecycle.


Three pillars of performance evaluation

1. Scientific validity

Scientific validity establishes that the analyte (what the test measures) is associated with — and therefore clinically relevant to — the clinical condition or physiological state the test claims to address.

For well-established analytes (e.g., glucose for diabetes, troponin for myocardial infarction), scientific validity is demonstrated by reference to published literature and established medical consensus. For novel biomarkers, a more substantial evidence base is required.

2. Analytical performance

Analytical performance characterises how accurately, precisely, and reliably the device measures the analyte. Key parameters:

ParameterDefinition
Sensitivity (analytical)Lowest concentration of analyte the device can reliably detect (limit of detection, LoD)
Specificity (analytical)Ability to measure only the target analyte without interference from similar substances
Accuracy / truenessCloseness of measured value to the true value; assessed against a reference method or certified reference material
PrecisionReproducibility of results — repeatability (within-run) and reproducibility (between-run, between-day, between-site)
LinearityRange over which the device produces proportional results
Measuring rangeConcentration range within which the device produces valid results
InterferenceEffect of potentially interfering substances (haemolysis, lipaemia, icterus; drugs; other analytes)
Matrix effectsPerformance variation between specimen types (e.g., serum vs plasma vs capillary blood)
Calibration stabilityStability of calibration over time; frequency of recalibration required
Reagent stabilityOpen-vial and closed-vial stability of reagents under defined storage conditions

3. Clinical performance

Clinical performance characterises the relationship between the device's output (test result) and the clinical condition in the intended patient population.

ParameterDefinition
Diagnostic sensitivityProportion of patients with the condition who test positive (true positive rate)
Diagnostic specificityProportion of patients without the condition who test negative (true negative rate)
Positive predictive value (PPV)Proportion of positive results that are true positives (depends on prevalence)
Negative predictive value (NPV)Proportion of negative results that are true negatives
Likelihood ratiosHow much a test result changes the probability of disease
Reference intervalsThe range of results expected in a healthy reference population
Cut-off determinationThe threshold between positive and negative results; method of determination; performance at cut-off

Clinical performance data is typically generated from prospective clinical studies, retrospective analysis of clinical samples, or literature data on equivalent devices.


Performance evaluation report (PER)

The PER documents the entire performance evaluation and must include:

  1. Device description and intended purpose
  2. Scientific validity summary
  3. Analytical performance data (test reports, validation studies)
  4. Clinical performance data (clinical studies, literature)
  5. Conclusions on overall performance
  6. Benefit-risk assessment in the context of the clinical condition
  7. Comparison with the state of the art (equivalent or competitor devices)
  8. References and data sources
  9. Date of PER and next scheduled update

Post-market performance follow-up (PMPF)

As with PMCF for general devices, IVD manufacturers must conduct Post-Market Performance Follow-up (PMPF) — ongoing collection of post-market performance data to:

  • Confirm the continued analytical and clinical performance of the device
  • Identify previously unknown analytical interferences
  • Detect emerging issues in real-world use
  • Update the PER with current data

PMPF data sources include:

  • Real-world analytical performance monitoring (quality control data, proficiency testing results)
  • External quality assurance (EQA) scheme participation
  • Published literature on the device or equivalent devices
  • Clinical audit data
  • Post-market clinical studies


Official references

ReferenceDescription
UK MDR 2002, Part III, Schedule 1AIVD Essential Requirements — analytical and clinical performance
MHRA: IVD medical devices guidanceMHRA's IVD guidance
EU IVDR 2017/746, Annex XIIIEU performance evaluation requirements (detailed reference — applies in NI)
ISO 17511:2020Metrological traceability of values assigned to calibrators
CLSI EP seriesClinical and Laboratory Standards Institute — analytical performance standards