ISO/IEC 17025:2017 is the international standard for testing and calibration laboratories. One of its most critical requirements is the evaluation and reporting of measurement uncertainty. Proper uncertainty evaluation demonstrates technical competence and ensures reliable measurement results.
Key Principle
ISO 17025 requires laboratories to have and apply procedures for estimating uncertainty of measurement. For calibration laboratories, uncertainty evaluation is mandatory for all calibrations. For testing laboratories, uncertainty must be estimated where relevant and possible.
ISO 17025 Measurement Uncertainty Requirements
The following table summarizes the key ISO 17025 requirements related to measurement uncertainty:
| Clause | Title | Category | Requirement |
|---|---|---|---|
| 7.6.1 | General Requirements | General | The laboratory shall identify the contributions to uncertainty. When evaluating measurement uncertainty, all contributions using appropriate methods of analysis shall be taken into account. |
| 7.6.2 | Measurement Uncertainty for Calibrations | Calibration | A calibration laboratory shall evaluate the measurement uncertainty for all calibrations. For calibration laboratories, all statements of conformity shall include the measurement uncertainty. |
| 7.6.3 | Measurement Uncertainty for Testing | Testing | A testing laboratory shall evaluate measurement uncertainty. Where the test method precludes rigorous evaluation of measurement uncertainty, an estimation shall be made based on understanding of the theoretical principles or practical experience. |
| 7.8.3.1 | Reporting Calibration Results | Reporting | For calibration certificates, the measurement uncertainty shall be reported. The certificate shall include where applicable: a statement of the measurement uncertainty. |
| 7.8.3.2 | Reporting Testing Results | Reporting | Where applicable, a statement on the estimated uncertainty of measurement shall be reported. The laboratory shall document the basis upon which the uncertainties have been evaluated. |
Calibration vs Testing Laboratories: Uncertainty Requirements
Calibration Laboratories
Mandatory Requirement
Uncertainty evaluation is required for all calibrations
Reporting
- Uncertainty must be included in calibration certificates
- All statements of conformity must include uncertainty
- Must follow GUM methodology (JCGM 100:2008)
Typical Sources
- Reference standard uncertainty
- Resolution of device under test
- Environmental conditions
- Repeatability and reproducibility
Testing Laboratories
Conditional Requirement
Uncertainty evaluation is required where relevant and possible
Reporting
- Report where it influences conformity decisions
- Report where client requests it
- Report where specification requires it
Typical Sources
- Sampling variability
- Sample preparation
- Method precision
- Instrument calibration
Key Difference
Calibration laboratories must evaluate and report uncertainty for all calibrations.Testing laboratories must evaluate uncertainty, but reporting is required only when it affects conformity decisions, when requested by the client, or when required by the specification.
ISO 17025 Uncertainty Implementation Guide
Develop Uncertainty Procedure
Document a procedure for uncertainty evaluation that covers all measurement processes.
Details:
The procedure should include: methodology (GUM), sources to consider, calculation methods, reporting format, review frequency, and responsibilities.
Create Uncertainty Budgets
Develop detailed uncertainty budgets for each measurement process.
Details:
Include all significant sources: Type A (repeatability) and Type B (calibration, resolution, environment). Use worst-case or realistic estimates based on data.
Validate Uncertainty Estimates
Verify uncertainty estimates through interlaboratory comparisons or proficiency testing.
Details:
Participate in proficiency testing schemes. Compare results with other labs. Use control charts to monitor measurement performance.
Train Personnel
Ensure all relevant staff understand uncertainty concepts and procedures.
Details:
Training should cover: uncertainty basics, GUM methodology, identifying sources, calculations, and reporting requirements.
Integrate with Quality System
Include uncertainty in method validation, equipment management, and quality control.
Details:
Link uncertainty to method validation data. Include in equipment calibration requirements. Use in acceptance criteria for quality control samples.
Review and Update
Regularly review and update uncertainty estimates.
Details:
Review when: new equipment is installed, methods change, proficiency test results indicate issues, or at least annually.
Reporting Requirements for Uncertainty
Calibration Certificates (Mandatory)
Required Information
- Measurement result with uncertainty
- Coverage factor (k)
- Confidence level (typically 95%)
- Brief explanation if needed for interpretation
Example Statement:
Measurement result: 100.05 g
Expanded uncertainty: U = 0.02 g
Coverage factor: k = 2
Confidence level: Approximately 95%Test Reports (Conditional)
When to Report
- When uncertainty affects conformity to specification
- When client requests uncertainty information
- When the test method specifies uncertainty reporting
- When it's necessary for interpretation of results
Example Statement:
Test result: 25.4 mg/kg
Expanded measurement uncertainty: U = 1.2 mg/kg
The reported uncertainty is an expanded uncertainty calculated using a coverage factor of k=2 which gives a confidence level of approximately 95%.Common ISO 17025 Uncertainty Challenges & Solutions
Identifying all uncertainty sources
Solution:
Use cause-and-effect diagrams (fishbone diagrams). Consult equipment manuals, calibration certificates, and method validation data. Consider environmental factors, operator effects, and sample variability.
Quantifying Type B uncertainties
Solution:
Use calibration certificate data (divide expanded uncertainty by k). For resolution, use rectangular distribution (a/√3). For environmental effects, monitor conditions and estimate variability.
Complex measurement models
Solution:
Break down complex measurements into simpler components. Use sensitivity coefficients (partial derivatives) or Monte Carlo simulation for non-linear models.
Maintaining uncertainty estimates
Solution:
Implement a schedule for regular review. Update when equipment changes occur. Use quality control data to verify ongoing validity.
Training staff
Solution:
Develop in-house training programs. Use practical examples from your laboratory's measurements. Consider external training courses on uncertainty.
Audit preparation
Solution:
Maintain complete documentation: procedures, uncertainty budgets, calculations, validation data. Be prepared to demonstrate how uncertainty was evaluated.
ISO 17025 Compliant Uncertainty Calculator
Meet ISO 17025 uncertainty requirements with our GUM-compliant calculator. Designed specifically for calibration and testing laboratories.
ISO 17025 Uncertainty FAQs
What does ISO 17025 require for measurement uncertainty?
ISO/IEC 17025:2017 requires laboratories to evaluate measurement uncertainty for all calibrations (clause 7.6.2) and to evaluate measurement uncertainty for testing (clause 7.6.3). The uncertainty must be reported with calibration results and, where applicable, with testing results.
Do all tests require uncertainty evaluation under ISO 17025?
Yes, ISO 17025 requires uncertainty evaluation for all tests. However, the standard recognizes that some test methods may preclude rigorous evaluation. In such cases, laboratories must make a reasonable estimation based on understanding of theoretical principles or practical experience.
What is the difference between Type A and Type B uncertainty in ISO 17025?
ISO 17025 follows GUM methodology. Type A uncertainty is evaluated by statistical analysis of measurement series, while Type B uncertainty is evaluated by means other than statistical analysis of series (e.g., from calibration certificates, manufacturer specifications). Both must be included in the uncertainty budget.
How should uncertainty be reported in calibration certificates?
Calibration certificates must include the measurement uncertainty. The statement should include the expanded uncertainty (U), the coverage factor (k), and the confidence level (typically 95%). Example: U = 0.5 mg, k = 2 (95% confidence level).
What are common sources of uncertainty in ISO 17025 laboratories?
Common sources include: calibration of reference standards, environmental conditions (temperature, humidity), instrument resolution, operator technique, measurement repeatability, sample inhomogeneity, and method limitations. All significant sources must be identified and quantified.
How often should uncertainty be re-evaluated?
Uncertainty should be re-evaluated whenever there are changes that could affect it: new equipment, changed environmental conditions, modified methods, or based on a predetermined schedule (e.g., annually). Ongoing monitoring through quality control also helps verify uncertainty estimates.