Quality Assurance

External quality assessment shows how your laboratory performs against its peers on identical material. You are expected to treat a survey specimen exactly as you would a patient specimen: routine method, routine staff, routine run, no extra repeats that would never happen on a real request, and no comparing of results with another laboratory before the submission deadline. When the report comes back you review every analyte, decide whether an unacceptable result reflects a genuine method problem, investigate it, document what you found and record the corrective action taken.

Testing done at the bedside still belongs to the laboratory. This area covers the oversight that keeps a ward glucose meter, a blood gas analyser or a urine strip reader producing results a physician can act on. You train and periodically re-assess the nurses and physicians who operate the device, confirm that quality control is run and accepted at the required frequency, watch lot changes and expiry dates, and compare device results against the central laboratory on split specimens. Deficiencies are documented and followed up exactly as they would be for bench work.

Laboratories improve only when problems are written down. This area asks you to recognize a deviation from the expected process — a missed critical value call-back, an expired reagent found in use, a fridge log signed for a shift nobody worked — and to enter it in the non-conformance system rather than correct it privately. Near misses count, even when no patient was affected. You describe what happened factually and without blame, and you help identify the underlying cause so that the corrective action addresses the process rather than the individual.

Records are the evidence that the work was controlled. Temperature logs, maintenance records, reagent logs and control charts are completed at the time the activity happens, initialled, and kept legible for the full retention period. An error is ruled through once and initialled so the original entry remains readable — never erased, never covered. Control charts carry the mean and standard deviation appropriate to the lot and instrument actually in use, points are plotted promptly, and out-of-range results are annotated with what was done about them. An unrecorded action is treated as an action that never happened.

Before any patient result leaves the analyser you decide whether the run was in control. That means reading quality control against defined limits and multirules rather than judging whether a number looks close enough, knowing the difference between a warning that prompts a look and a rule violation that rejects the run, and confirming that the calibration behind the run is current, in date and matched to the reagent lot. It also means recognizing when calibration must be repeated — after major maintenance, a lamp or electrode change, a new reagent lot, or a drift the control data has started to show.

Errors come in two shapes and they are handled differently. Random error scatters results on either side of the mean and widens imprecision; an air bubble, a partly blocked probe or an unstable bath temperature typically causes it, and it usually appears as an isolated outlier or a broadening spread. Systematic error moves results in one direction — a shift appearing abruptly after a calibration or lot change, a trend developing as a reagent or lamp deteriorates. Reading control charts for these patterns lets you find and fix the cause before patient results are affected.

Nothing new goes into patient use on the manufacturer's word alone. For a new reagent lot you run the old and new lots in parallel with controls, and ideally with patient specimens spanning the reportable range, to confirm the results agree before the changeover. For an unmodified, approved assay on an additional analyser you verify precision, accuracy against a comparative method or reference material, and the reportable range, and you confirm the reference interval suits your patient population. All of it is documented and signed off before the first patient result is reported.

Running out of a reagent partway through a shift is a quality failure, not merely an inconvenience. This area covers knowing your usage rate, setting minimum stock levels and reorder points, inspecting deliveries on arrival, and storing each item at the temperature its insert specifies. Stock is rotated so that the earliest expiry is used first, and expired material is removed from the bench rather than left on the shelf beside the current lot. Lot numbers are recorded against the runs they were used for, so that a manufacturer's recall can be traced to the patient results involved.

Laboratory data has to stay accurate, private and retrievable for years. This area covers entering and verifying results correctly, amending a released result so that the original value, the new value, the reason and the person responsible all remain visible, and never overwriting a record silently. It also covers access — you view the results your work requires and no others — together with backup, secure storage, defined retention schedules, and a tested downtime procedure so that results generated while the information system is unavailable are still recorded, traceable and later reconciled with the permanent record.

Every laboratory procedure exists in exactly one current, approved version, and that is the one the bench works from. This area covers using documents issued through the controlled system rather than photocopies taped inside cupboard doors, checking that the version in your hand is the current revision, and withdrawing superseded copies as soon as a revision is issued. Changes go through review and authorization before release, staff are informed and trained on them, and obsolete versions are either destroyed or archived and clearly marked so they cannot be picked up and followed by mistake.