Using control charts to monitor TB laboratory test performance
Reliable tuberculosis testing depends on more than a validated assay or a functioning instrument. Laboratory teams also need a practical way to see whether performance remains stable from day to day, identify early warning signs, and distinguish random variation from a genuine problem. Control charts provide that visual and analytical framework.
A control chart displays quality control results in sequence, usually against a target value, a central line, and statistically derived warning or control limits. For TB laboratories, it can support monitoring of molecular testing, smear microscopy, culture, drug susceptibility testing, and related processes. Used consistently, the chart becomes part of routine quality management rather than an occasional troubleshooting exercise.
The GLI Quality Tool places this activity within a broader system that includes equipment management, personnel competency, documentation, assessment, and continual improvement. A chart cannot correct a weak process by itself, but it can help the laboratory detect process changes early enough to investigate them before patient results are affected.
Why longitudinal monitoring matters
A single quality control result offers limited information. It may be acceptable, yet a sequence of results can reveal a gradual shift, increasing variability, or a repeated pattern associated with a particular operator, reagent lot, instrument, or environmental condition. Reviewing results over time makes these signals easier to recognize.
For example, a TB molecular assay may continue producing valid runs while internal control values slowly move toward an unacceptable range. A sudden increase in invalid results may also appear first as a cluster on a chart. These patterns can indicate reagent deterioration, maintenance needs, pipetting problems, contamination, temperature excursions, or changes in specimen quality.
Control charts are especially useful where laboratories manage several testing platforms or work across changing conditions. They create a common visual language for discussions during supervisory reviews, internal audits, and management meetings. The aim is not to penalize staff for normal variation, but to understand whether the testing process is predictable and in statistical control.
Selecting data and establishing a baseline
The first step is to define what the laboratory will monitor. Suitable measures may include cycle threshold values for a stable positive control, optical density readings, smear staining control scores, culture contamination rates, invalid or error rates, turnaround time, and external quality assessment results. The selected indicator should be linked to a clear quality objective and recorded in a consistent format.
For quantitative measurements, the laboratory should establish a baseline using enough results to represent routine operation. The mean or median may serve as the central value, while standard deviation or another suitable measure describes expected variation. Control limits should be based on the method, the control material, and the laboratory’s historical performance rather than copied without review from another setting.
Qualitative results require a different approach. A p-chart can track the proportion of invalid tests or contaminated cultures, while a u-chart can be used for counts when the number of opportunities varies. Laboratories with limited statistical support can begin with run charts and simple rules for detecting shifts, then introduce formal control limits as data quality improves.
Before implementation, document the control material, lot number, testing frequency, instrument, analytical method, calculation method, review interval, and responsibilities. If several instruments are used, keep their data separate at first. Combining results can hide platform-specific problems and produce misleading limits.
Reading signals without overreacting
A point outside the control limits is a signal for investigation, not automatic proof that every patient result is unreliable. The team should first check transcription, calculations, control preparation, instrument status, reagent lots, maintenance records, environmental conditions, and operator activity. The investigation should be documented, along with the decision about whether patient testing must be paused or repeated.
Patterns within the limits can be equally important. Several consecutive points on one side of the center line may indicate a shift. A steady upward or downward trend can signal gradual deterioration. Repeated alternating extremes may suggest excessive adjustment or inconsistent technique. Rules should be defined in advance so that staff respond consistently rather than interpreting the chart according to personal expectations.
The laboratory should avoid changing control limits every time a result looks inconvenient. Limits may need revision after a documented, sustained process improvement or a change in method, instrument, or control material. Any revision should include the reason, approval, effective date, and comparison with previous performance.
| Monitoring approach | Best suited to | What it can reveal | Practical limitation |
|---|---|---|---|
| Individual values chart | Quantitative control results from regular runs | Shifts, trends, and outliers | Requires stable, comparable measurements |
| Moving range chart | Small datasets with one result per run | Changes in short-term variability | Less informative when results are infrequent |
| p-chart | Proportions such as invalid tests | Changing rates across different sample volumes | Needs accurate numerator and denominator data |
| u-chart | Counts per testing opportunity | Events such as contamination or errors | Interpretation depends on consistent event definitions |
| Run chart | Laboratories beginning routine monitoring | Persistent shifts and trends | Does not provide formal control limits |
Applying charts to GeneXpert workflows
Molecular TB testing on GeneXpert or similar platforms generates several useful quality indicators. Laboratories may monitor positive and negative control behavior, internal control results, error codes, invalid rates, unsuccessful runs, module performance, and turnaround time. These indicators should be connected to the manufacturer’s instructions and the laboratory’s own quality objectives.
Staff need to understand what each control means and what action is required when it fails. Practical competency training should cover specimen preparation, cartridge handling, data entry, result interpretation, contamination prevention, instrument maintenance, and escalation pathways. The guidance on GeneXpert staff training can support laboratories that are formalizing these responsibilities.
A chart should never replace the instrument’s built-in controls or manufacturer-defined acceptance criteria. Instead, it adds a longer-term view. If invalid results rise on one module while other modules remain stable, the laboratory can prioritize module inspection. If errors cluster after a cartridge lot change, the team can compare lot records and control behavior. If performance differs by shift, direct observation and competency review may be appropriate.
Data should be stratified carefully. Combining results from different assay versions, control materials, instruments, or testing conditions can create false trends. When a change is unavoidable, mark it on the chart and begin a new baseline if the analytical characteristics have changed substantially.
Turning chart findings into quality actions
A useful control chart leads to a documented response. When a signal appears, assign an investigator, record the suspected cause, review affected results, and define corrective action. Root cause analysis may include the five whys, a process map, review of maintenance logs, interviews with staff, and examination of reagent or control storage conditions.
Corrective action should be proportionate to the risk. A minor transcription error may require retraining and a form revision. Repeated invalid molecular tests may require instrument service, environmental assessment, verification of specimen handling, or temporary referral of testing. If patient results may have been affected, the laboratory must follow its policy for result review, notification, and repeat testing.
After action is taken, continue monitoring to determine whether performance has returned to the expected pattern. This follow-up is essential. Closing an incident because a repair was completed does not demonstrate that the process improved. A sustained return toward the center line, reduced variability, or lower error rate provides stronger evidence of effectiveness.
Charts should be reviewed at a defined frequency. Bench staff may review them after each run or at the end of each testing day, while the quality manager may conduct a monthly trend review. Management reviews can examine recurring causes, resource needs, and whether the laboratory’s quality indicators remain aligned with service priorities.
Building the practice into the quality system
Control charting works best when it is integrated with the laboratory’s existing documentation and assessment processes. The procedure should specify data sources, calculations, chart format, review responsibilities, alert rules, investigation steps, and record retention. A controlled electronic spreadsheet may be sufficient for a small laboratory if formulas are protected and changes are traceable.
Personnel should be trained to enter data accurately and recognize defined signals. Competency assessment can include creating a chart, interpreting example patterns, selecting an appropriate response, and completing an investigation record. Supervisors should periodically compare the chart with source documents to verify completeness and accuracy.
The approach should also accommodate resource limitations. Laboratories may begin with paper forms, a wall chart, or a basic spreadsheet. The essential requirements are consistent data, clear decision rules, timely review, and documented action. As capacity grows, automated dashboards or laboratory information system integration can reduce transcription and improve trend analysis.
Recommendations for a sustainable monitoring process include:
- Start with a small number of high-risk, clearly defined indicators.
- Separate data by instrument, method, control lot, or testing site when differences may affect results.
- Establish alert and action rules before routine monitoring begins.
- Record investigations, corrective actions, and evidence that the action was effective.
- Review chart performance during internal audits and management meetings.
A well-maintained control chart gives TB laboratory teams an early view of process behavior. Begin by selecting one stable quality control measure, collect dependable baseline data, and agree on the response to each signal. With regular review and documented follow-up, the laboratory can turn routine test records into practical evidence for safer, more consistent patient testing.