Analytical Performance Evaluation of In Vitro Diagnostic (IVD) Medical Devices: Standards, Testing I

Analytical performance evaluation is a mandatory scientific process that demonstrates whether an in vitro diagnostic (IVD) medical device can reliably and accurately measure its target analyte under controlled conditions. This article provides a comprehensive overview of the analytical performance characteristics, applicable standards such as ISO and CLSI guidelines, sample requirements, testing methodologies, and regulatory expectations under frameworks including EU IVDR 2017/746. It serves as a technical reference for IVD manufacturers, regulatory affairs professionals, and quality managers preparing performance evaluation reports or pre-market submissions.

Technical Overview

Analytical performance evaluation refers to the systematic assessment of an in vitro diagnostic medical device's ability to detect or measure a target analyte with acceptable accuracy, precision, and reliability. Unlike clinical performance evaluation, which examines how well a device supports medical decision-making in a defined intended population, analytical performance evaluation focuses on the intrinsic measurement characteristics of the device itself. It constitutes a core component of the performance evaluation report required by major regulatory frameworks worldwide.

Under the European In Vitro Diagnostic Medical Device Regulation (EU) 2017/746 (IVDR), manufacturers must establish and maintain a performance evaluation process covering scientific validity, analytical performance, and clinical performance. Similar requirements exist in the United States under FDA regulations, in China under NMPA medical device regulations and associated YY/T standards, and in international markets guided by ISO standards. The evaluation must be planned, executed, documented, and justified through a documented performance evaluation plan and report.

Key analytical performance characteristics generally include precision, trueness and accuracy, linearity, measuring range, detection capability (limit of blank, limit of detection, limit of quantitation), analytical specificity and interference, and calibration traceability. Internationally recognized standards organizations such as ISO and CLSI (Clinical and Laboratory Standards Institute) publish harmonized methodologies that are widely accepted by regulators and accredited third-party laboratories.

• ISO 13485 – Quality management systems for medical device organizations
• ISO 14971 – Application of risk management to medical devices
• ISO 17511 – Metrological traceability of calibrator and control material values
• ISO 15197 – In vitro diagnostic test systems for blood glucose monitoring
• ISO 20916 – Clinical performance studies using specimens from human subjects
• CLSI EP05, EP06, EP07, EP09, EP15, EP17, EP28 – Harmonized evaluation protocols
• EU IVDR 2017/746 Annex II and III – Technical documentation and performance evaluation

Testing Samples

The selection and preparation of testing samples are critical to the validity of an analytical performance study. Samples must be representative of the intended use population, matrix, and analyte concentration range claimed by the manufacturer.

Clinical Human Specimens

Native human specimens are the preferred sample type for most analytical performance studies. These include serum, plasma (with defined anticoagulants), whole blood, urine, cerebrospinal fluid, stool, nasopharyngeal or oropharyngeal swabs, and other body fluids. Specimens should cover the clinically relevant concentration range, including low concentrations near medical decision levels, and should reflect the intended use population in terms of demographics and disease states where applicable.

Reference Materials and Calibrators

Certified reference materials (CRMs), international standards, and national reference materials with assigned values are used for trueness verification and calibration traceability assessment. Where higher-order reference materials are unavailable, the manufacturer must justify the chosen traceability chain in accordance with ISO 17511.

Contrived and Spiked Samples

When native specimens at extreme concentrations are difficult to obtain, samples may be prepared by spiking pooled human matrix with the target analyte or by diluting high-concentration samples with the appropriate matrix. For molecular IVD devices, contrived samples may be prepared using synthetic RNA/DNA, armored particles, or viral transport medium. The preparation process must be documented and justified.

Quality Control and Interference Panel Materials

Commercial quality control materials with assigned target values and acceptable ranges are used to monitor study validity. Interference panels containing endogenous substances (hemoglobin, bilirubin, lipids, rheumatoid factor, human anti-mouse antibodies) and exogenous substances (common drugs and their metabolites) at physiologically or therapeutically relevant concentrations are prepared according to CLSI EP07.

Testing Items

Analytical performance evaluation typically encompasses the following testing items, with the specific scope determined by the device type, technology platform, and intended use claims.

• Precision: repeatability, within-laboratory (intermediate) precision, between-run, between-day, between-operator, and between-instrument variation
• Trueness and accuracy: bias assessment against reference methods, reference materials, or comparative methods
• Method comparison and bias estimation using patient specimens
• Linearity, dilution linearity, and validating the manufacturer's linear range
• Measuring range, analytical measuring range, and clinically reportable range
• Detection capability: limit of blank (LoB), limit of detection (LoD), and limit of quantitation (LoQ)
• Analytical specificity: interfering substances, cross-reactivity, and endogenous antibodies
• Carryover and sample-to-sample contamination
• Hook effect (prozone effect) at high analyte concentrations
• Calibration traceability and commutability of calibrators
• Specimen stability under collection, transport, and storage conditions
• Reagent stability, on-board stability, and opened-vial stability
• Reference interval verification or establishment
• Software-related analytical functions, such as automatic result calculation and quality control rules

Testing Methods

Evaluation studies should follow internationally harmonized protocols and be conducted under a quality management system with pre-defined acceptance criteria, statistical analysis plans, and documented raw data. Testing may be performed by the manufacturer or delegated to accredited third-party laboratories with appropriate competence.

Precision Study (CLSI EP05)

A precision study is typically conducted over at least 20 operating days using at least two concentration levels covering the measuring range, with two runs per day and duplicate measurements per run. Repeatability, within-laboratory precision, and between-component variation are estimated using nested analysis of variance (ANOVA). Results are expressed as standard deviation (SD) and coefficient of variation (CV), and compared against the manufacturer's claimed precision specifications.

Trueness and Accuracy Assessment (CLSI EP09, EP15)

Trueness is assessed by comparing device results against reference material assigned values, reference measurement procedures, or an established comparative method. The EP09 method comparison protocol typically requires at least 100 patient specimens spanning the measuring range, analyzed with both the test device and the comparative method. Bias is evaluated using difference plots (Bland-Altman), regression analysis (Deming or Passing-Bablok), and estimated bias at medical decision levels. The EP15 protocol provides a shorter verification approach using control materials.

Linearity Evaluation (CLSI EP06)

Linearity is evaluated using a series of at least four to five concentration levels prepared by mixing high and low concentration samples in known proportions. Deviation from linearity is assessed by polynomial regression analysis, comparing observed values against the best-fit linear model. The manufacturer's claimed linear range is verified if the nonlinearity bias at each level remains within predefined acceptance limits.

Detection Capability Study (CLSI EP17)

The detection capability study determines the limit of blank (LoB), limit of detection (LoD), and limit of quantitation (LoQ). LoB is established from repeated measurements of blank samples, LoD from low-level samples near the detection threshold using a distribution-based statistical approach, and LoQ is defined as the lowest concentration at which the total analytical error meets specified accuracy targets. Acceptance criteria must align with clinical requirements.

Interference and Cross-Reactivity Testing (CLSI EP07)

Interference is evaluated by adding potential interferents at clinically relevant concentrations into test and control samples and measuring the analytical bias. Dose-response testing is performed when a potential interference is identified. Cross-reactivity studies assess the device response to structurally related compounds, similar pathogens, or phylogenetically related organisms for molecular and immunoassay devices.

Carryover and Hook Effect Assessment

Carryover is assessed by sequencing high-concentration and low-concentration samples in a defined alternating pattern and statistically comparing the low sample results against the blank baseline. Hook effect testing involves analyzing samples with analyte concentrations far exceeding the upper limit of the measuring range to confirm that the device produces no falsely low results.

Stability Studies

Real-time stability, accelerated stability, transport stability, in-use stability, and freeze-thaw stability are evaluated using protocols aligned with applicable standards. Statistical trending analysis is applied to confirm that performance remains within specification throughout the claimed shelf life.

Applications

Analytical performance evaluation data support a wide range of regulatory, quality, and commercial purposes throughout the product lifecycle.

• Regulatory submissions: technical documentation for EU IVDR conformity assessment, FDA pre-market submissions (510(k), De Novo, PMA), and NMPA registration dossiers
• Performance evaluation reports: compilation of scientific validity, analytical performance, and clinical performance sections required under EU IVDR Annex XIII
• Design verification and validation: demonstrating that design outputs meet design input requirements under ISO 13485
• Self-declaration of conformity and CE technical file maintenance for lower-risk IVD devices
• Post-market surveillance and periodic safety update reports, including re-evaluation following design or manufacturing changes
• Laboratory verification: hospitals and clinical laboratories verifying manufacturer claims under quality management systems aligned with ISO 15189
• Support for companion diagnostic claims where analytical reliability underpins clinical utility
• Evidence packages for point-of-care testing devices, molecular assays, immunoassays, clinical chemistry systems, and self-testing IVD products

FAQ

What is the difference between analytical performance and clinical performance?

Analytical performance describes how accurately and reliably a device measures the target analyte, including characteristics such as precision, trueness, and detection capability. Clinical performance describes how well the device results support the intended medical purpose in the intended use population, typically expressed through clinical sensitivity and clinical specificity. Both are required components of a complete performance evaluation under EU IVDR, with clinical performance studies governed by ISO 20916 when human subjects are involved.

Which standards should be followed for analytical performance evaluation?

Manufacturers commonly follow CLSI evaluation protocols (EP05 for precision, EP06 for linearity, EP07 for interference, EP09 for method comparison, EP15 for user verification, EP17 for detection capability, and EP28 for reference intervals) together with ISO 17511 for traceability, ISO 13485 for quality management, and ISO 14971 for risk management. Regional requirements, such as EU IVDR 2017/746 or national medical device regulations, define the documentation and reporting obligations.

Can analytical performance studies be outsourced?

Yes. Manufacturers may delegate study execution to accredited third-party laboratories or clinical sites, provided the manufacturer retains overall responsibility for the performance evaluation, documented agreements define data ownership and access, and the test sites operate under adequate quality systems with qualified personnel and calibrated equipment.

How many samples are required for a method comparison study?

According to the CLSI EP09 approach, a minimum of 100 patient specimens covering the full measuring range and clinically relevant concentrations is generally recommended. Additional samples may be needed for rare analytes, extreme concentration ranges, or when method differences require more robust statistical power.

What acceptance criteria should be used?

Acceptance criteria must be pre-defined in the performance evaluation plan and justified based on clinical requirements, intended use, medical decision levels, and biological variation. Criteria should not be established retrospectively after data collection, as this undermines the scientific validity of the study.

How often must analytical performance be re-evaluated?

Performance evaluation must be treated as a continuous, lifecycle process. It should be updated when new evidence becomes available, when significant design or manufacturing changes occur, when post-market surveillance or user feedback identifies performance concerns, or at the intervals specified in the manufacturer's performance evaluation plan.

What role does metrological traceability play?

Metrological traceability, addressed by ISO 17511, ensures that calibrator values are linked to higher-order references through an unbroken chain of comparisons. Demonstrating traceability is essential for achieving comparable results across laboratories and measurement systems, and it is a documented requirement for regulatory compliance in most major jurisdictions.