Laboratory Analyser Maintenance, Calibration and Troubleshooting: The Reliability Work Every BMS Does
Most of the working day in an NHS pathology laboratory depends on one quiet assumption: that the analysers are working correctly. Keeping that assumption true — through planned maintenance, calibration, watching quality control, reading error flags and knowing when to escalate — is the day-to-day reliability work that every biomedical scientist (BMS) and support staff member shares. This article sets out the standards, the methodology and the practical judgement that turn "the machine is broken" into a controlled, documented, patient-safe response.
Why Equipment Reliability Is a Patient-Safety Issue
An analyser does not have to stop working to cause harm. A drifting electrode, a partly blocked probe or an ageing lamp can keep producing results that look plausible but are wrong — and a wrong result acted on clinically can be more dangerous than no result at all. This is why equipment management is a core competence rather than an engineering afterthought.
In the UK, the framework comes from three directions that must be read together:
- ISO 15189:2022 — Medical laboratories — Requirements for quality and competence, the standard against which UKAS (the United Kingdom Accreditation Service) assesses pathology services. Its Clause 6.4 (Equipment) and Clause 6.5 (Equipment calibration and metrological traceability) set the requirements for selecting, maintaining, calibrating and recording equipment performance.
- The manufacturer's Instructions for Use (IFU) — the legally significant document that defines how a specific analyser must be operated, maintained and calibrated. Deviating from the IFU without validation can invalidate both the result and the accreditation status of the test.
- *The MHRA's Managing Medical Devices guidance — the Medicines and Healthcare products Regulatory Agency guidance for health and social care organisations, which frames planned preventive maintenance (PPM), repair, adverse-incident reporting and device lifecycle management across the Trust.
Planned Preventive Maintenance (PPM): Doing the Boring Things on Time
Planned preventive maintenance is scheduled, routine maintenance carried out at defined intervals to prevent failures before they happen, rather than waiting for an analyser to break down (reactive or "breakdown" maintenance). The MHRA's
Managing Medical Devices guidance is clear that the frequency and type of PPM should follow the manufacturer's instructions, and ISO 15189:2022 expects the programme to be documented and evidenced.PPM is usually split into tasks the laboratory performs and tasks the manufacturer's engineer performs under a service agreement (covered by ISO 15189:2022 Clause 6.7, Service agreements). A typical tiered schedule looks like this:
| Frequency | Typical tasks | Usually performed by | |-----------|---------------|----------------------| | Daily | Probe/needle wash, waste and reagent checks, startup QC, surface cleaning, water/fluidics check | BMS / support staff | | Weekly | Deep clean of sample/reagent pathways, deproteinisation, cuvette/flow-cell checks | BMS | | Monthly | Filter changes, lamp checks, pump/seal inspection, more extensive decontamination | BMS / senior BMS | | Quarterly / Annual | Optics alignment, full preventive service, firmware updates, replacement of wear parts | Manufacturer engineer (service contract) |
Key principles every trainee should absorb:
1. Follow the IFU, not folklore. Local "the way we've always done it" practices must match the current IFU. Manufacturers update maintenance frequencies, and inadequate or skipped maintenance is a recognised cause of gradual result bias — a slow problem that often surfaces through EQA returns rather than an obvious breakdown. 2. Record at the point of doing. A contemporaneous, honest maintenance log is the evidence UKAS assessors will ask for. Back-filled logs are a common, and serious, accreditation finding. 3. Maintenance is not optional under workload pressure. Skipping a clean because the laboratory is busy is exactly when drift and contamination creep in. PPM protects throughput; it does not compete with it. 4. Don't sign for work you didn't do. A signature or electronic record is a professional attestation, governed by Health and Care Professions Council (HCPC) standards of conduct for registered scientists.
Calibration and Metrological Traceability
Calibration establishes the relationship between what the analyser measures and the true, reference value of the analyte. Verification then confirms the analyser is performing within its stated specification — the two are related but distinct, and examiners often test the difference.
ISO 15189:2022 Clause 6.5 requires calibration to be performed at defined intervals, or before use where needed, using calibrators traceable to a recognised reference. Metrological traceability is the requirement that a result can be linked, through an unbroken and documented chain of calibrations each contributing to measurement uncertainty, back to the International System of Units (SI) or another higher-order reference — typically via certified reference materials and reference measurement procedures. UKAS sets out its expectations in its published policy on metrological traceability (TPS 41).
In practical day-to-day terms a BMS should be able to:
- Calibrate when triggered, not just on a timer — common triggers include a new reagent lot, major maintenance or part replacement, QC drift, or a lapsed IFU calibration interval.
- Confirm and record calibrator and reagent lot details and expiry, because traceability depends on knowing exactly which materials produced the result.
- Verify calibration with QC, never accepting a calibration as valid until quality control confirms the analyser is back in specification.
- Keep the records. Calibration date, calibrator lot, acceptance criteria, the result and the responsible person all form part of the traceability chain ISO 15189:2022 demands.
Internal Quality Control and the Westgard Context
Internal quality control (IQC) is how the laboratory detects, in near real time, whether an analyser is producing reliable results. At least two levels of control material — commonly one near the clinical decision level and one in an abnormal range — are run, and the results are plotted on a Levey-Jennings chart relative to the established mean and standard deviation (SD).
Westgard rules are a set of multirule criteria used to decide whether an analytical run is acceptable. They are designed to detect random and systematic error while keeping false rejections low. The commonly used rules are:
| Rule | What it flags | Type of error | Usual action | |------|---------------|---------------|--------------| | 1₂ₛ | One control beyond ±2 SD | Warning only | Inspect; trigger to examine other rules | | 1₃ₛ | One control beyond ±3 SD | Random | Reject the run | | 2₂ₛ | Two consecutive beyond ±2 SD same side | Systematic | Reject the run | | R₄ₛ | Range between two controls exceeds 4 SD | Random | Reject the run | | 4₁ₛ | Four consecutive beyond ±1 SD same side | Systematic (look-back) | Reject / investigate | | 10ₓ | Ten consecutive on the same side of the mean | Systematic drift | Reject / investigate |
The 1₂ₛ rule is a
warning that prompts you to inspect the other rules; it is not on its own a reason to reject results. The rejection rules (1₃ₛ, 2₂ₛ, R₄ₛ and the systematic look-back rules) tell you the run cannot be trusted. Crucially, the pattern hints at the cause: random error (1₃ₛ, R₄ₛ) often points to a one-off problem such as a bubble or clot, while systematic error (2₂ₛ, 4₁ₛ, 10ₓ) points to drift or a shift — a deteriorating reagent, ageing calibration or a maintenance issue building over time.When IQC fails, the single most important rule is simple: do not release patient results from an out-of-control run. The correct sequence is covered below.
Interpreting Analyser Error Flags
Modern analysers generate two broad categories of message, and confusing them is a common trainee error.
- Instrument/system errors describe a fault in the analyser itself: clot detection, insufficient sample, probe crash, pressure/fluidics fault, lamp or optics error, communication failure with the laboratory information management system (LIMS). These usually require an operator action or maintenance.
- Result/sample flags describe something about the specimen or the result: haemolysis, icterus, lipaemia (HIL indices), abnormal reaction kinetics, results above or below the analytical measuring range, delta-check failures, or instrument-suggested reflex testing.
1. Read it, don't clear it. Record the exact flag or error code before dismissing it; the IFU and manufacturer's error-code list are the authoritative reference for what it means. 2. Separate analytical from physiological. A "result above measuring range" flag may be entirely real (and need dilution), whereas a clot flag is an analytical problem to resolve before reporting. 3. Check whether the flag affects one sample or many. A single haemolysis flag is a sample issue; the same flag across a whole rack suggests a pre-analytical or instrument problem. 4. Never authorise around a flag you don't understand. Suppressing a flag to "get the result out" is precisely how erroneous results reach clinicians.
Flags are the analyser telling you what it noticed — information to interpret, not obstacles to clear.
Basic Fault-Finding: A Structured Approach
When something goes wrong, resist the urge to swap parts at random. A structured method is faster, safer and produces the documented evidence an investigation will later need.
1. Stabilise and contain. Stop releasing affected results; hold questionable reports rather than authorising under doubt. 2. Gather the facts. Note the exact error code or flag, the time, what changed (new lot, new shift, recent maintenance), and whether IQC was in or out. 3. Work from simple to complex. Most faults are mundane: empty reagent or waste, an air bubble, a clot, an expired calibrator, a loose connection, a sample-integrity problem. Eliminate these before suspecting optics or electronics. 4. Consult the IFU troubleshooting guide. Manufacturer manuals list error codes with prescribed corrective actions. Follow them; do not improvise repairs on a regulated medical device. 5. Reproduce and confirm. After any corrective action, prove the fix with fresh QC before resuming patient work — a clear QC after maintenance closes the loop on calibration and verification. 6. Document everything. The fault, the investigation, the action and the verification all belong in the maintenance or non-conformity record.
Distinguish
correction (getting this analyser running again) from corrective action (stopping the fault recurring). Recurrent faults should be raised as non-conformities and investigated using root cause analysis, not simply reset each time.Knowing When and How to Escalate
Sound judgement about escalation is one of the clearest markers of a competent BMS. Escalate — promptly and without embarrassment — when:
- IQC cannot be brought back in control after reasonable, IFU-directed troubleshooting.
- A fault is beyond your training, authorisation or the IFU's operator-level instructions (engineer territory).
- Patient results may have been released from an out-of-specification analyser — this becomes a non-conformity and a potential patient-safety event.
- A backup analyser or send-away arrangement is needed to maintain a clinical service.
- You suspect a device-related safety problem. Suspected adverse incidents involving medical devices should be reported to the MHRA through the Yellow Card scheme, in line with the
Frequently Asked Questions
What is the difference between calibration and verification?
Calibration establishes the relationship between the analyser's signal and the true value of the analyte, using traceable calibrators. Verification confirms the analyser is performing within its specification — usually by running quality control after calibration. Under ISO 15189:2022 a calibration is not considered demonstrated until in-control QC verifies it; the two steps work together.
Can I release patient results if QC has failed?
No. The fundamental rule of internal quality control is that patient results must not be released from a run that is out of control. Withhold the affected results, investigate and correct the cause, re-run QC to confirm the analyser is back in specification, then process patient samples. Any results already released from an out-of-control run must be assessed for clinical impact and treated as a potential non-conformity.
What is metrological traceability and why does it matter?
Metrological traceability is the documented, unbroken chain of calibrations — each contributing to measurement uncertainty — linking your result back to a recognised higher-order reference such as the SI units. It matters because it is what makes a result from your laboratory comparable to a result anywhere else, and it is an explicit requirement of ISO 15189:2022 Clause 6.5. In practice it depends on using certified calibrators and reference materials and recording their lot and traceability details.
How do I know how often to maintain or calibrate an analyser?
The manufacturer's Instructions for Use (IFU) define the minimum maintenance and calibration frequencies, and the MHRA's
Managing Medical Devices guidance reinforces that planned preventive maintenance should follow those instructions. Local schedules may add extra checks but must never fall below the IFU requirements. Calibration is also triggered by events — new reagent lots, major maintenance, part replacement or QC drift — not by the timer alone.What should I do when an analyser shows an error flag I don't recognise?
Record the exact error code or flag first, then consult the analyser's IFU and manufacturer error-code list, which are the authoritative source for its meaning and the prescribed corrective action. Do not simply clear or suppress the flag to release results. If the IFU's operator-level actions do not resolve it, escalate to a senior BMS and, where appropriate, manufacturer technical support.
When should an equipment problem be reported to the MHRA?
Suspected adverse incidents involving a medical device — including analysers that malfunction in a way that could affect patient safety, or that produce biased or unreliable results — should be reported to the MHRA via the Yellow Card scheme, alongside internal non-conformity and patient-safety reporting. This is part of the device-vigilance system described in the
Managing Medical Devices* guidance, and your laboratory's medical-device safety officer or quality manager will usually coordinate the report.Further training
Equipment reliability sits at the heart of laboratory quality. Build on this topic with the pillar guide and related articles:
- NHS Laboratory Training — the central hub for biomedical-science laboratory training across all specialties.
- Quality Control in the NHS Lab: EQA, IQA and IQC Explained — the wider quality framework that QC and EQA sit within.
- Method Validation and Verification under ISO 15189:2022 — how new methods and analysers are proven fit for use.
- Root Cause Analysis and CAPA in Pathology — investigating recurrent faults and non-conformities properly.
- UKAS and ISO 15189 Accreditation: A Guide for Biomedical Scientists — the accreditation context behind all equipment and quality requirements.