Verifying AFB smear microscopy through blinded rechecking

Accurate acid-fast bacilli (AFB) smear microscopy remains important for tuberculosis (TB) services, particularly where rapid molecular testing is unavailable, delayed, or used alongside smear results for clinical decisions. A well-designed blinded rechecking system provides an independent measure of slide-reading performance without turning routine quality assurance into a punitive exercise.

Blinded rechecking means that a second reader examines a selected set of stained smears without knowing the original result. The process can identify false-positive and false-negative readings, errors in grading, staining problems, and weaknesses in reporting. When linked to clear corrective action, it becomes a practical part of a laboratory quality management system rather than a retrospective inspection.

Define the purpose and scope

Before collecting slides, the laboratory should document what the rechecking programme is intended to verify. It may assess the accuracy of AFB detection, the consistency of semiquantitative grading, the reliability of staining, or the performance of individual readers. These aims should be agreed with the laboratory manager, TB programme and supervising reference service.

The scope should cover the relevant specimen types, staining method, reporting categories and staff groups. A laboratory using Ziehl–Neelsen staining may need a different review process from one using auramine fluorescence microscopy. The protocol should state who selects slides, who performs the blinded review, how disagreements are resolved, and how results are recorded.

Blinded rechecking should complement, not replace, internal quality control, external quality assessment and direct observation of workplace practice. The GLI roadmap can help laboratories establish the supporting systems in phase 1 quality foundations, including responsibilities, documentation, equipment control and basic process management.

Select a representative set of smears

The credibility of the exercise depends on how slides are selected. If staff can predict which smears will be reviewed, they may handle those slides differently from routine work. The person selecting slides should therefore use an agreed random method and remove identifying information that could reveal the original result to the second reader.

A sample should reflect the laboratory’s normal workload, including negative, scanty or low-positive, and clearly positive smears. If only positive slides are reviewed, false-negative errors remain invisible. If only convenient slides are chosen, the findings may describe storage quality rather than routine diagnostic performance. Selection can be systematic, such as every tenth eligible slide, or generated using a random-number method.

The number of slides should be justified by workload, risk, historical performance and the precision required for the monitoring period. A small rural service may review a manageable number each month and combine results over several months, while a high-volume metropolitan laboratory may review a larger sample at shorter intervals. The sampling plan, exclusions and reasons for missing slides must be retained.

Slides should be stored securely and remain traceable to the original worksheet or laboratory information system. In Australia, this matters across public hospital networks, private pathology providers and regional services where specimens may travel by scheduled courier from areas outside Sydney, Melbourne, Brisbane or Perth. Delays, temperature exposure and broken slides can affect the review and should be recorded rather than silently excluded.

Preserve blindness and standardise the second reading

The second reader should receive slides labelled with a study code, not the original patient identifier or result. The accompanying form should contain only the information necessary to perform the examination, such as the staining method and the area to be examined. The reader should record the result independently before any comparison takes place.

A standard reading method is essential. The procedure should specify microscope checks, acceptable staining quality, the number of fields examined, how scanty bacilli are reported, and the grading scale used by the laboratory. Readers should use the same definitions for negative, scanty, 1+, 2+ and 3+ results, or for the equivalent local reporting categories.

The reviewer must be competent in AFB microscopy and should not routinely be the person who produced the original result, where staffing allows. In a small service this separation may be difficult, so an external reader or regional reference laboratory can provide an independent check. Any limitations should be documented instead of presenting the arrangement as fully independent.

The rechecking workflow should avoid contamination and protect staff. Slides must be handled according to local biosafety procedures, with suitable disinfectants, sharps controls and waste disposal. A practical cleaning and decontamination job aid can support consistent work-surface procedures before and after microscopy activities.

Compare results using meaningful categories

After the second reading is completed, the original and recheck results can be compared. The most serious discrepancies are false-positive and false-negative classifications because they may influence isolation, treatment, contact investigation or decisions about further testing. A difference between negative and scanty should be considered carefully, particularly where the laboratory’s reporting policy treats scanty results in a special way.

Quantification errors also matter. A slide originally reported as 3+ but re-read as 1+ indicates a different problem from a negative slide reported as 2+. The review form should therefore capture the exact original grade, the exact recheck grade, the reader, date, stain batch if relevant and any comments on debris, precipitate, poor staining or inadequate smear preparation.

Results should be summarised with simple measures that staff can understand. These may include the proportion of slides with agreement, the number of major classification errors, the number of grading discrepancies and trends by reader or month. Percentage agreement alone can be misleading when most slides are negative, so laboratories should interpret it alongside the distribution of results and the number of discordant cases.

A disagreement should be confirmed through a defined resolution process. A senior microscopist or reference laboratory can perform an adjudication reading, preferably without seeing the earlier interpretations. If the slide is damaged, poorly stained or unreadable, it should be classified as an invalid review rather than automatically assigning blame to the original reader.

Investigate errors and apply corrective action

A discrepant result is a signal for investigation, not proof of negligence. The laboratory should look for possible causes across the whole testing pathway: specimen quality, smear thickness, fixation, staining time, reagent preparation, microscope condition, reader fatigue, workload, transcription and result entry. A recurring false-negative pattern may reflect insufficient fields examined, weak staining or excessive workload rather than a single knowledge gap.

Corrective action should match the cause. Staff may need a short competency refresher, supervised reading of a teaching set, changes to stain quality control, microscope maintenance or clearer reporting prompts. If the issue involves surface contamination or poor bench practice, responsibilities and cleaning frequency should be reviewed alongside the relevant work instructions.

Patient-impact assessment is essential when a confirmed error could have affected clinical management. The laboratory should notify the appropriate clinical or public health contact according to its incident procedure, review related specimens where justified and document the decision. Australian laboratories should align this process with their accreditation arrangements, including NATA expectations under ISO 15189 and applicable state or territory health service requirements.

Corrective and preventive actions need owners and due dates. The laboratory should check whether the intervention worked through a later blinded recheck, competency assessment or targeted quality control review. Staff working in remote Queensland, the Northern Territory or regional Western Australia may need telehealth support, shared digital teaching sets or scheduled visits because immediate access to a senior microscopist is limited.

Use findings for continual improvement

Blinded rechecking is most useful when results are reviewed as a trend. A single discordant slide may arise from an unusual specimen or a damaged preparation, while repeated errors in one reporting category indicate a system issue. Monthly or quarterly review meetings can examine patterns without naming individuals unnecessarily and can link findings to training, equipment and workload planning.

The laboratory should report key indicators to its quality committee and TB programme. Useful records include the sampling frame, number of slides reviewed, agreement by result category, confirmed major errors, invalid slides, investigations, actions and follow-up outcomes. Retaining these records supports internal audits and demonstrates that quality findings lead to improvement.

Local operational realities should shape the programme. Australian laboratories often balance TB microscopy with molecular platforms such as Xpert MTB/RIF or other nucleic acid tests, while maintaining microscopy capacity for specific clinical and public health needs. Staff may also work across several benches, use rostered weekend coverage and send specimens between metropolitan hubs and smaller hospitals. A rechecking plan must fit those workflows rather than rely on an idealised full-time microscopy team.

The final procedure should be controlled, accessible and reviewed at a defined interval. Train each reader in the current method, verify competency before independent work, and update the protocol when staining methods, reporting categories, instruments or legislation change. With consistent sampling, genuine blindness, fair interpretation and timely follow-up, rechecking becomes a dependable safeguard for TB diagnosis.

Implement a documented blinded rechecking programme using the GLI Quality Tool’s checklists and practical resources. Assign responsibility, select a representative sample, record every discrepancy and review the results with the people who can correct the underlying process. Consistent action will strengthen AFB smear microscopy and give clinicians and public health teams greater confidence in the results.