The fortnight before an inspection has a recognisable character in most laboratories. Binders are reorganised, overdue reviews are signed in a rush, temperature logs acquire a suspicious uniformity of handwriting, and somebody stays late reconciling a competency file that has been incomplete for a year. Everyone knows this is not how it is supposed to work, and it happens anyway.
The reason it happens is a misconception about what is being assessed. Inspectors are not primarily checking whether the laboratory produces correct results today. They are checking whether the system reliably produces correct results, which is a different question and can only be answered from records accumulated over time. A laboratory that assembles evidence in the final fortnight is demonstrating exactly the opposite of what it hopes to show.
Reframing readiness as continuous documentation habit rather than an event changes both the work and the anxiety. It is more effort spread thinly and much less effort concentrated, and it removes the specific dread of being asked for something that does not exist.
Key takeaways
- Inspectors assess whether a system works over time, which only records can demonstrate.
- Most findings concern documentation gaps rather than technical incompetence.
- Competency assessment must be evidenced, not asserted, and must recur at defined intervals.
- A finding is not a failure; how the laboratory responds to it carries more weight.
- Records created in real time look and read differently from records created retrospectively.
What Inspectors Are Actually Checking
An inspection tests three things: that the laboratory has decided how it will work, that it works that way, and that it can prove it. Almost every finding traces back to a gap between those three.
Assessors follow a small number of standard techniques, and knowing them removes most of the mystery.
The commonest is document-to-practice tracing. An assessor picks a procedure, reads it, then watches or asks staff to describe how the task is done. Any divergence is a finding, and it does not much matter which side is better: if practice has improved beyond the procedure, the procedure is out of date, and if it has drifted below, the procedure is not being followed. Both are the same defect.
The second technique is the vertical audit, sometimes called a sample trace. The assessor picks one specimen and follows it end to end: request, collection, transport, receipt, processing, analysis, quality control valid at that moment, calibration state of the instrument, the operator’s competency record, result validation, report, and any amendment afterwards. This single exercise touches nearly every system in the laboratory, and it is unforgiving of gaps, because a missing link anywhere breaks the chain visibly.
The third is examining how the laboratory handles its own problems, through incident records, non-conformity logs, corrective actions, complaints and internal audit findings. Counter-intuitively, an empty log is worse than a full one: every laboratory has problems, and one reporting none is failing to detect or record them.
The fourth is verifying that the laboratory checks itself. Internal audits, management review, quality indicators and external quality assessment all demonstrate self-monitoring, and where they exist and are acted upon, an assessor can lean on the laboratory’s own findings rather than duplicating them.
Documentation That Must Be Current
Document control is the single largest source of findings, and the underlying requirement is simple: everyone must be using the current version, and nobody should be able to use a superseded one by accident.
Procedures need an owner, a version, an approval, a review date and a distribution record. The review date is the item most often missed. Procedures typically require review at a defined interval whether or not anything has changed, and a review that concludes no change is needed is a valid review provided it is recorded. A procedure last reviewed years ago is a finding regardless of how well it describes practice.
Superseded versions cause the most avoidable trouble. Old copies in drawers, on shared drives, pinned above benches or saved locally by individuals are all findable, and an assessor who finds staff working from a version that was withdrawn has found both a document control failure and a practice failure at once. Removing obsolete copies systematically when a new version is issued, and keeping the single archived master where it is clearly marked as historical, avoids the whole category.
Several other document families are commonly overlooked. Method validation and verification records must exist for every test in use, including tests running long before the current staff arrived. Risk assessments need review after changes, safety data sheets must match the chemicals actually held, and supplier approvals and reagent specifications should be traceable. Where reference intervals are reported, their basis must be documented, whether established locally or adopted and verified.
Records of authorisation matter as much as the documents. Who may authorise results, approve procedures, release a method after a fault, and who deputises in their absence should all be written down, because delegation existing only as informal understanding cannot be evidenced.
Competency Records and Training Files
Competency is the area where laboratories most often believe they are compliant and most often are not, because the requirement is for demonstrated and recorded competence rather than for training attendance.
The distinction runs through everything. A record showing someone attended a session on an analyser proves attendance. A record showing someone was observed performing the procedure correctly, interpreted results appropriately, and handled a defined problem scenario proves competence. Assessors look for the second.
A defensible file covers qualifications and their verification, a job description matching the work actually done, induction, a training record for each procedure performed, initial assessment before independent working, and periodic reassessment thereafter.
Reassessment intervals are frequently weak. Many schemes expect more frequent assessment during an initial period and less once someone is established, and expect reassessment to be triggered by events regardless of schedule: a significant procedural change, a new instrument, a period away from the work, or an incident involving the individual.
The methods used to assess competence should vary, because different methods reveal different weaknesses.
| Assessment method | What it demonstrates | What it cannot show |
|---|---|---|
| Direct observation of the task | Correct technique and sequence in practice | Judgement when something goes wrong |
| Review of the person’s recorded work | Consistency and accuracy over time | Real-time decision making |
| Blind or unknown sample testing | Analytical accuracy independent of expectation | Handling of the pre-analytical stage |
| Problem-solving scenarios | Troubleshooting and escalation judgement | Routine manual dexterity |
| Written or verbal questioning | Understanding of principles and limits | Whether understanding translates into practice |
Using a single method throughout is a common finding, because a laboratory that only ever observes tasks has no evidence anyone can handle a failure, and one that only ever uses written questions has no evidence anyone can perform the work.
Two supporting elements complete the picture. Training should be linked to a specific version of the procedure, so a substantive revision prompts retraining rather than silently invalidating the record. And gaps must be visible: where someone is not yet competent, the record should state what they may do under supervision and what they may not.
Equipment Calibration and Maintenance Evidence
Equipment records fail in predictable ways, and all of them are avoidable with a schedule and a habit.
Every significant piece of equipment needs an identity, meaning a unique identifier that appears on the item and in the records, so that a certificate can be tied to a specific instrument rather than to a model. Two identical centrifuges without individual identifiers make every record ambiguous.
Calibration evidence needs to be current and traceable. Traceability means the calibration links through an unbroken chain of comparisons to a recognised national or international standard, which in practice means the calibration certificate names the reference used and its own traceability. A certificate showing an instrument was checked against something unspecified establishes very little.
The commonly overlooked items are not the analysers, which are usually covered by service contracts and are hard to forget. They are the supporting devices: pipettes, thermometers, timers, balances, centrifuge speeds, water baths, incubators, refrigerators and freezers. Pipettes in particular are a recurring finding, since they drift with use and are frequently outside their calibration interval simply because nobody owns the schedule.
Monitoring records for temperature-controlled storage need more than a reading. They need defined acceptable limits, evidence that readings were reviewed rather than merely recorded, and documented action whenever a limit was breached. A log with an excursion sitting unremarked among the readings is a worse finding than no log, because it shows monitoring exists and is ignored.
Maintenance records should distinguish planned from unplanned work and record what was done, by whom, and what verification followed. Verification after repair is the step most often missing: something must demonstrate the instrument is fit to resume patient work, usually recalibration and acceptable quality control, recorded alongside the repair.
Finally, equipment out of service should be labelled unambiguously and removed from use, and records for retired equipment should be kept for the required period rather than discarded with the instrument.
Handling Findings and Corrective Actions
Findings are normal. Assessors expect to raise some, and a report with none from a laboratory of any size invites more scrutiny rather than less. What distinguishes a well-run laboratory is the response.
The first discipline is answering the actual finding. Assessors write findings against a specific requirement, and the response should address that requirement directly rather than the broader topic. A finding that competency reassessment intervals were not defined is answered by defining them and evidencing the definition, not by an essay on the training programme.
The second is separating correction from corrective action. Correction fixes the immediate instance: the overdue calibration is performed, the missing signature is obtained. Corrective action addresses why it happened so it does not recur: a schedule with an owner, an alert, a change to the review process. A response containing only correction is routinely rejected, because it guarantees the same finding next time.
Root cause analysis is where responses go thin. The honest question is what allowed the gap to persist unnoticed, and the answer is usually structural: no owner for the task, no scheduled prompt, no check that would have caught it. Answers stopping at human error are rarely accepted, because human error is a description rather than a cause.
The third discipline is evidence. A response asserting that a procedure has been updated and staff retrained needs the updated procedure, the training records and a date. Assessors verify closure, and a response without attachments generates a follow-up rather than a closure.
The fourth is timeliness. Responses are due within a defined and usually short period. Where a corrective action genuinely takes longer, such as one requiring capital purchase, state the interim control managing the risk meanwhile and a realistic completion date. An honestly described interim measure is acceptable; a missed deadline with no communication is not.
Finally, treat findings as intelligence. A cluster of findings across different areas that all reduce to the same underlying weakness, such as no ownership of recurring scheduled tasks, is telling the laboratory something more useful than any individual item. Reviewing findings collectively at management review is where that pattern becomes visible.
Staff Interviews and What to Expect
Assessors talk to staff at every level, and this is the part people dread most and need to dread least, because the questions are more predictable than the anxiety suggests.
The purpose is to establish whether the system exists in practice or only on paper. An assessor is testing whether staff know where to find current procedures, understand what they are doing and why, know what to do when something goes wrong, and know the limits of their own authority.
Typical questions are unglamorous. Where is the current procedure for this task. What do you do if a control fails. What would you do with a critical result. How do you know this pipette is in calibration. Where do you record this. When were you last assessed on it.
The best preparation is knowing where things are rather than memorising content. Nobody is expected to recite a procedure. Being able to locate the current version quickly is a stronger answer than reciting an outdated one from memory, and it demonstrates precisely what the assessor is checking.
Three habits make interviews go well. Answer the question asked rather than a broader one, since volunteering unrelated detail opens new lines of enquiry. Say plainly when you do not know something and say where you would find out, which is a genuinely good answer rather than a failure. And do not speculate about how other departments work, because an inaccurate second-hand account creates a discrepancy the assessor will then pursue.
For managers, one further point matters. If a member of staff cannot answer a question about a procedure they perform, the finding lands on the training and competency system rather than on the individual. Preparing staff by rehearsing scripted answers is both transparent and counterproductive; ensuring they have genuinely been trained and assessed is the actual preparation.
Habits That Make Inspections Routine
Everything above reduces to a handful of ordinary habits, none individually demanding.
Record at the time. A log completed as the work happens looks different from one completed later, and assessors are experienced at spotting the difference. Uniform ink, uniform handwriting, perfectly regular intervals and readings that never vary are all patterns that invite scrutiny, and a genuine record with an occasional annotation explaining a gap is far more credible than an implausibly tidy one.
Give every recurring task an owner and a date. Most findings exist because a scheduled activity had no name attached to it. A simple register of recurring obligations, each with a responsible person, a frequency and a next-due date, prevents more findings than any other single measure.
Run internal audits as if they were external, spread through the year. Vertical audits following one sample end to end are particularly effective, because they replicate the assessor’s most revealing technique and surface the gaps between departments that departmental audits miss. Close those findings out: internal items left open until an assessor arrives convert a manageable problem into an external one. Treat document review dates as real deadlines, spread so dozens do not fall due together.
Make the quality management system usable. Systems that are hard to search, demand unnecessary approvals for small corrections, or bury current procedures among archived ones cause non-compliance, because staff route around obstacles.
Finally, walk the laboratory occasionally with an assessor’s eye. Look at what is pinned above benches, what is in the drawers, whether labels are legible, whether the out-of-service instrument in the corner is labelled, and whether the fridge log has been reviewed. Most findings are visible to anyone genuinely looking, and the fortnight of panic exists mainly because nobody has looked in a year.
Frequently asked questions
How much notice does a laboratory usually get?
It varies by scheme and purpose. Planned accreditation assessments are usually scheduled well in advance with the scope and timetable agreed beforehand. Regulatory inspections may give short notice or none, particularly when they follow a complaint, an incident or a previous adverse finding. Many schemes have moved towards unannounced elements precisely because advance notice encourages the fortnight of preparation described above. Assuming a visit could occur at any time is both safer and, in practice, less work.
Is a finding the same as failing?
No, and treating it that way distorts the response. Findings are graded, and most are minor observations about documentation or consistency that close on a written response with evidence. Serious findings, meaning those suggesting results may be unreliable or patients at risk, are a different matter and can suspend a scope of testing. Accreditation is not usually withdrawn over a set of ordinary findings, and the response to them signals far more about a laboratory than their existence.
Who should escort the assessor?
Someone senior enough to answer system-level questions and authorise access, usually the quality or laboratory manager, accompanied where relevant by the person responsible for the area being examined. The escort’s job is to facilitate and to record what is asked and seen, not to answer for staff. Answering on behalf of someone asked a direct question prevents the assessor from establishing what that person knows, and implies they may not know it.
What if an assessor asks for a record that does not exist?
Say so plainly and immediately. Creating the record during the visit is a far more serious matter than the missing record itself, and back-dated documentation is usually detectable and always damaging to the laboratory’s credibility on everything else. A missing record produces a finding that can be corrected; a fabricated one raises questions about the integrity of every other record. Where the information exists elsewhere in another form, offer that and explain the situation.
How should a small laboratory approach this with limited staff?
By concentrating on the highest-yield items rather than building a full-scale quality department. A single register of recurring obligations with owners and due dates prevents the largest category of findings. Competency records on a simple repeatable template cover the second largest, and equipment records with unique identifiers and a calibration schedule cover much of the third. Small laboratories are assessed against the same standards but not the same scale of implementation, and a modest system genuinely followed is far stronger than an elaborate one that is not.
Inspection readiness is not a project with a deadline. It is a set of small habits that produce, as a by-product, the evidence an assessor needs. A laboratory that records at the time, gives every recurring task an owner, audits itself honestly and closes its own findings has already done the work, and the visit becomes a description of what is there rather than a search for what is missing. The fortnight of panic is a symptom, and the cure sits entirely upstream of it.
