Practical guide
Unsatisfactory PT result: what the laboratory should do
An unsatisfactory result needs investigation and a documented review. Repeating a measurement until the score improves does not demonstrate that the problem is resolved. The aim is to identify the cause and protect the reliability of laboratory results.
The workflow at a glance
Record → assess risk → investigate → correct → verify effectiveness → close and monitor
Six practical steps
1. Open a record and appoint an owner
Record scheme, batch, sample, method, submitted result, unit, score, acceptance criterion and report date. Retain the original report and raw data. Set a risk-based internal deadline: there is no single deadline valid for every PT.
2. Decide immediately how to protect results
Could the defect affect customer samples? Identify the period, tests and reports potentially affected, starting from the last reliable evidence. If validity is doubtful, consider holding affected work or result release; notify customers and recall or amend reports where necessary. Also justify any conclusion of no impact.
3. Check data, units and assessment first
Compare the submitted value with the worksheet: sample code, decimals, conversions, dilutions and dry or wet basis. Read the instructions and report criteria: z, z′, zₚ and Eₙ are not interchangeable. A warning or repeated trend warrants assessment against defined internal criteria even without an unsatisfactory result.
4. Find an evidence-supported cause
Involve the analyst. Review receipt, storage, preparation, equipment, calibration, internal controls, operator and calculations. For each hypothesis, record what confirms or rejects it. “Human error” alone is insufficient: ask why the system failed to prevent or detect it. If the material or assessment is questioned, send evidence and reasons to the provider.
5. Distinguish correction from corrective action
Correction fixes the identified value or defect. Corrective action prevents its cause recurring, for example by validating the spreadsheet, controlling revisions or checking operator competence. Give each action an owner, deadline and expected evidence. Preserve the amendment trail; never silently overwrite the original result.
6. Close only after verifying effectiveness
Before verification, define material, method, acceptance criterion and who authorises closure. Use evidence relevant to the cause: independent recalculation, controls, fresh preparations and, where required or appropriate, external verification. A better score alone does not resolve doubts about earlier reports. If the cause remains unknown, keep the case open or document why residual risk is acceptable.
When the result is questionable or doubtful
A “questionable” or “doubtful” result is a warning signal. Read the proficiency testing report to understand the assessment: the organiser’s criteria apply.
1. Check the submitted value
Check the sample code, unit, transcription and calculations. If you find an error, record it and ask the organiser how to submit a correction.
2. Review the other results too
Check previous proficiency testing results and internal controls. Does the laboratory tend to report values increasingly above or below the reference? Has the same warning occurred before?
3. Decide what to check
If this is an isolated case and you find no other problems, document your checks and decide how to monitor subsequent results. If the warning recurs, worsens or accompanies unacceptable internal controls, investigate the cause using the six steps above.
4. Assess customer results
If their validity is in doubt, promptly assess which tests and reports may be affected and what action is needed. Do not wait for the next PT to address an identified problem.
5. Record the decision
Record what you checked, what you found, who is following up and when the next check will take place. Also justify any decision not to take further action. Follow laboratory procedures and applicable requirements.
External verification and applicable requirements
Check your accreditation rules. Where required, two consecutive unsatisfactory outcomes trigger notification of self-suspension of the test and its related grouping, unless a justified written explanation applies. This is not a universal rule for every PT.
After an adverse result, check whether you must promptly request residual material or a new comparison from the provider. Review the PT plan: repeating the same material does not automatically constitute a new round. A CST may support verification if technically suitable and acceptable under applicable rules; buying a sample does not close the nonconformity.
Example 1 · Airborne asbestos, SEM/EDS
The problem
The laboratory obtains an unsatisfactory or questionable result. Before repeating the analysis, check how you identified and counted the fibres and calculated the result.
Are you counting the required fibres?
- Did you count the fibre types specified by the organiser, following the required dimensions and counting rules?
- Did you find agglomerates, meaning groups of overlapping or intertwined fibres, that made counting difficult? How did you handle them under the method’s rules?
- Could some fibres be materials other than asbestos? Check their morphology and EDS data; if identification remains uncertain, carry out the further checks required by the method.
Did you calculate the examined area correctly?
Check the area of each field examined, the number of fields and the total area used in the calculation. Also check image scale and magnification.
For example: I counted 40 fibres over 0.87 mm² but reported 40 fibres/mm², as though I had examined 1 mm². The correct calculation is 40 ÷ 0.87 = 45.98 fibres/mm². Reporting 40 fibres/mm² therefore understated the result. These numbers are illustrative only.
Does the examined area represent the whole filter?
Did you examine fields spread across different parts of the deposit or only one area? Are fibres distributed similarly in the inner and outer parts? Do not select only the easiest fields to read or those giving the expected result.
Where the method, scheme instructions and instrument allow, it is preferable to place the whole filter in the SEM so that different parts of the deposit can be examined. This does not mean counting the entire filter surface.
If a portion must be taken, a sector extending from the centre to the edge, such as a quadrant, includes both inner and outer parts. The choice must follow the method and organiser’s instructions: the sector alone does not guarantee that the examined portion is representative. Fibre distribution and field selection also matter.
What to do after the checks
Record the problems found and checks performed. If you find a counting, identification or calculation error, correct it and verify that it does not recur. Assess whether previously issued reports may also be affected. If you suspect a filter problem, send the organiser images, data and a description of the difficulties.
Example 2 · Asbestos in soil, SEM/EDS
The problem
The PT result is unsatisfactory or questionable, or you obtain very different values from portions of the same sample. Check every step, from the received sample to the final result.
Did you prepare the sample correctly?
- Did you take an aliquot, meaning a small portion, that truly represents the sample received?
- Did you thoroughly homogenise the laboratory sample, following the method and organiser’s instructions?
- If grinding is required, did you grind the sample as specified by the method? Check duration, procedure and final particle size, avoiding unprescribed treatments that could alter the fibres.
- Did you correctly take the aliquot of the dispersion to deposit on the filter? Was the dispersion well mixed at the time, or had some material already settled?
- Did you correctly record masses, volumes and dilutions?
Did you examine and measure the fibres correctly?
- Did you calculate the examined area correctly, checking field area and number of fields?
- Did you distribute observations so that they represent the deposit on the filter?
- Did you count and measure enough asbestos fibres under the method and scheme criteria? Also check the criteria for ending the examination, especially when few fibres are found.
- Are the identified fibres those required by the organiser? Are there agglomerates that are difficult to count or fibres that may not be asbestos?
- For chrysotile, did you measure the diameter in the compact part of the bundles, rather than using the width of open or frayed ends? Apply the method’s rules for measuring and handling bundles.
- Did you check calculations and units, including reporting on a dry-soil basis where required?
What to do after the checks
If you find an error, correct the affected step and verify that analysts follow the procedure correctly. Assess whether the same problem could affect previously analysed soils and reports already issued.
Then verify that the correction works using checks suited to the identified cause and external verification where required. Do not select only the result closest to the reference value or change treatment solely for the PT sample to obtain a better result.
Further reading
Guides help structure the investigation; accreditation policies apply within their own scope. Always check the applicable revision. This page is an original practical synthesis, not an official translation of the referenced documents.
- Eurachem · Investigating poor performance: second edition and translations, including the updated Italian version.
- Eurachem · Selection, Use and Interpretation of PT Schemes, 3rd edition, 2021: use of results and Appendix B for investigations.
- UKAS · TPS 47, edition 5: PT participation, planning and corrective actions. A policy for its own accreditation system.
- EA · EA-4/18 G: choosing participation level and frequency according to risk.
- ISO/IEC 17025:2017 · Reference for validity of results, nonconforming work and corrective action (7.7, 7.10, 8.7); full text requires a licence.