ISO/IEC 17025 clause 7.5.1: capture original observations as work happens

What clause 7.5.1 requires in practice

Clause 7.5.1 is about the content and timing of the record for each laboratory activity. The record holds the results, the report and enough supporting information to identify what could have affected the measurement result and its uncertainty, and to make it possible to run the activity again in circumstances matching the original as closely as can be managed. Dates and the identity of the people responsible for the activity, and for checking data and results, are included.

The most practical words in the clause concern timing and linkage. Original observations, data and calculations are written down when they are made, and each can be tied to the specific task. That excludes reconstruction from memory at the end of the day and loose notes that never reach the file. The standard leaves the design of the record to the laboratory, so the right amount of detail depends on the method and on how much judgement the analyst applies.

Decide what another competent person needs to follow

Start with the method and its critical decisions. Identify the observations, equipment references, conditions and calculations needed to understand the result and associated uncertainty. Include the date and personnel responsible for the activity and for checking the data and results.

The requirement does not mean every irrelevant room detail must be recorded. It means the evidence must be sufficient for the actual activity. A routine automated measurement and a method involving analyst judgement may need different supporting information. Use technical understanding to design the record rather than copying a generic template.

Why clause 7.5.1 matters to the laboratory

When a result is questioned months later, the original observations are the one neutral witness. If they were copied, tidied or reconstructed, the laboratory cannot show what actually happened, and an assessor, customer or court will treat the answer accordingly. Good contemporaneous records also help day to day: they let a second person review with confidence, make trend investigations quicker and give staff a defensible position when something goes wrong. For laboratories moving towards accreditation, the quality of raw records is one of the clearest signs of whether the system works in practice. It is far harder to fix after the fact than to design in from the start.

How to implement clause 7.5.1

Make the right way of recording the easiest way. Involve the analysts in designing it, since they know which details they actually rely on when something looks wrong.

1. Identify, method by method, the decisions and observations that most affect the result, and require those to be captured in the record.
2. Provide bench-side or field-side recording tools, such as bound books, prepared sheets or tablets, so staff are not tempted to use scraps.
3. Set a rule on instrument-generated data: how files are named, linked to the job and stored before the device overwrites them.
4. Identify the version of every worksheet, formula or software routine used, and record it on the job.
5. Require units and item identity on every entry, and prompt for them in templates.
6. Show who performed and who checked each stage, with individual sign-off.
7. Run periodic reviews of completed jobs against these expectations and share the findings with the team.

Capture evidence at the point of work

Make calculations understandable

Preserve inputs, the applicable calculation and its connection to the result. Where a controlled calculation tool is used, the record should identify the relevant version and data. A reviewer should not have to guess which formula or correction factor was active when a result was produced.

Coordinate with systematic calculation and transfer checks under clause 7.11.6. Recording a number and checking it are distinct activities, and the technical record must identify the responsible personnel. Shared usernames can obscure accountability unless the system otherwise reliably records who performed the work.

Common nonconformities with clause 7.5.1

Weak technical records tend to show up in a small number of repeating ways.

• Original observations are written on scrap paper and transcribed later, and the scrap is discarded.
• Worksheets omit units, item identification or the version of the calculation used.
• Instrument memory overwrites raw data before it is saved to the job file.
• The person who checked the data cannot be identified, or the same shared log-in appears throughout.
• Records are completed in a single batch at the end of a period, with times that do not match the work.

Two illustrative recording decisions

Evidence and common errors

The actual contemporaneous records are the evidence, not a statement that staff are expected to keep records. Review completed jobs for missing units, ambiguous item IDs, undocumented decisions and unexplained gaps. A neat final worksheet can conceal a weak record if the original observations were reconstructed later.

Do not turn “repeat the activity” into a promise that an identical physical result can always be reproduced. The standard qualifies this by what is possible and by conditions as close as possible to the original. The purpose is sufficient technical understanding, not a claim of perfect physical repeatability.