How to validate a new TB diagnostic test in your laboratory
Introducing a new tuberculosis diagnostic test is a clinical, technical, and quality management decision. A kit may perform well in a manufacturer’s study yet produce different results in a local laboratory because of specimen quality, disease prevalence, staff experience, environmental conditions, equipment, or the characteristics of the patients being tested.
Laboratory validation provides documented evidence that a method is suitable for its intended use. It should show whether the test detects or identifies tuberculosis reliably, produces consistent results, fits the laboratory workflow, and supports safe, timely decisions about patient care and public health action.
The process applies to molecular assays, line-probe assays, culture-based methods, microscopy, antigen tests, and other technologies. The exact experiments will differ, but the underlying principles remain consistent: define the claim, compare performance with an appropriate reference, investigate weaknesses, and authorize use only when the evidence meets pre-established criteria.
Define the intended use and scope
Begin by writing a clear statement of intended use. Specify the organism or target, specimen types, patient population, testing purpose, result categories, and users of the result. For example, a test may be intended to detect Mycobacterium tuberculosis complex in sputum from people with presumptive pulmonary TB, or to identify resistance-associated targets in specimens that have already tested positive.
The intended use determines the validation design. A method for initial diagnosis requires evidence about clinical sensitivity and specificity across relevant patient groups. A test used for drug-resistance detection requires suitable resistant and susceptible strains or specimens, with enough representation of important resistance mechanisms. A method introduced only for smear-positive specimens should not automatically be treated as validated for smear-negative specimens or extrapulmonary samples.
Review the manufacturer’s instructions, regulatory status, published evidence, instrument requirements, and limitations. Compare these claims with local needs. Establish which results will be reported, how invalid or indeterminate outcomes will be handled, and whether confirmatory testing is required. The laboratory director should approve the scope before specimens are tested.
Prepare the quality and safety foundation
Validation is strongest when it sits within a functioning quality management system. Assign responsibilities for the laboratory director, quality manager, technical lead, biosafety officer, testers, data reviewers, and approving authority. Control the protocol, worksheets, raw data, calculations, deviations, and final report as quality records. Laboratories building this foundation can use this quality management system guide to organize core processes around personnel, equipment, documents, assessment, and continual improvement.
Confirm that the laboratory has suitable space, workflow separation, power, temperature control, storage, waste management, and information systems. Equipment should be installed, calibrated, maintained, and checked before validation begins. Reagents must be received, stored, tracked by lot, and monitored for expiry. Staff should demonstrate competency in specimen handling, testing, result interpretation, contamination prevention, and documentation.
Risk assessment should cover infectious specimens, aerosol-generating steps, spills, sharps, chemical hazards, and waste. Put controls in place before the first validation run, including personal protective equipment, engineering controls, procedures for accidents, and access to post-exposure support. A practical laboratory safety checklist can help identify overlooked hazards in laboratories working with limited resources.
Build a protocol with measurable acceptance criteria
A validation protocol should describe the method, materials, operators, instruments, testing sequence, controls, data analysis, deviations, and approval process. Define acceptance criteria before examining the results. This prevents the laboratory from adjusting expectations after seeing favorable or unfavorable findings.
Select specimens or panels that reflect the test’s intended use. Include confirmed positive and negative materials, weakly positive specimens near the expected detection limit, and challenging samples when relevant. Consider smear status, bacterial load, HIV status, age groups, disease sites, common contaminants, and resistance patterns. If fresh clinical specimens are limited, use well-characterized archived specimens or quality-assured external panels, while clearly documenting their limitations.
Plan enough replicates to evaluate repeatability and reproducibility. Repeatability examines variation under the same conditions, such as one operator, instrument, and day. Reproducibility examines changes across operators, days, reagent lots, instruments, or sites. For quantitative or semi-quantitative methods, assess agreement across the measurement range, including clinically important decision points.
| Validation element | Evidence to collect | Typical decision |
|---|---|---|
| Accuracy or agreement | Results compared with a suitable reference method or well-characterized panel | Agreement meets the predefined target |
| Analytical sensitivity | Detection of low-concentration or weakly positive material | Detection limit is acceptable for the intended use |
| Analytical specificity | Negative specimens and organisms or substances that could cause cross-reaction | No clinically important false-positive pattern |
| Precision | Repeated results across runs, operators, days, lots, or instruments | Variation remains within defined limits |
| Invalid and error rate | Frequency and causes of failed, indeterminate, or rejected results | Rate is manageable and corrective actions are defined |
| Workflow performance | Turnaround time, hands-on time, resources, and reporting steps | Method fits routine service capacity |
Generate and analyze reliable evidence
Use a controlled sequence that minimizes bias and contamination. Randomize or blind specimens where practical, separate pre-analytical and post-amplification areas when required, and include positive, negative, internal, and environmental controls according to the method. Record every result as generated, including invalid outcomes, repeats, instrument errors, and protocol deviations.
Compare the candidate method with a reference method appropriate to the question. Culture, a validated molecular assay, sequencing, drug-susceptibility testing, or a composite clinical reference may be suitable in different circumstances. No reference is perfect, so explain discordant results rather than simply excluding them. Investigate unexpected positives and negatives through repeat testing, an alternative method, review of specimen history, and examination of possible contamination or inhibition.
Calculate performance measures that match the test. Sensitivity and specificity are useful for binary outcomes, while positive and negative predictive values depend strongly on disease prevalence. Report confidence intervals when possible. For categorical results, agreement statistics can add value. For quantitative tests, examine bias, imprecision, linearity, and clinically relevant differences. Analyze invalid rates and turnaround time separately; a method with excellent analytical accuracy may still be unsuitable if too many patient specimens fail.
Review the findings and authorize implementation
The validation report should summarize the intended use, protocol, specimen sources, operators, equipment, reagent lots, controls, results, statistical analysis, deviations, limitations, corrective actions, and final decision. State clearly whether the method is accepted, accepted with restrictions, or rejected. Restrictions might include approved specimen types, required confirmatory testing, limited operating hours, or use by specifically trained personnel.
A discrepancy investigation should lead to documented action. Possible responses include retraining, instrument maintenance, revised specimen rejection criteria, changes to extraction procedures, additional quality controls, or a narrower reporting claim. If evidence does not meet the acceptance criteria, do not conceal the problem by repeating tests until a preferred outcome appears. Pause the decision, identify the cause, and revise the protocol only with formal approval.
Before routine release, establish standard operating procedures, result-reporting language, referral arrangements, stock controls, maintenance schedules, and an internal quality control plan. Use available quality checklists to verify that documents, records, equipment, personnel, assessment, and continual improvement activities are ready for service.
Monitor performance after implementation
Validation is a point-in-time demonstration, while verification and ongoing quality assurance show whether performance remains stable. Run internal controls at the required frequency, review control trends, investigate shifts, and participate in an appropriate external quality assessment or proficiency testing program. Compare local results with expected patterns and monitor changes in positivity, invalid rates, contamination, turnaround time, and referral volume.
Reassess the method after major changes, such as a new instrument, software update, reagent lot, specimen type, testing location, supplier, or reporting algorithm. A small verification exercise may be sufficient for a minor change, while a broader validation may be needed when the intended use or technology changes. Document the decision and retain evidence for audits, accreditation, program review, and patient-safety investigations.
Use routine data to improve the service. Repeated invalid results may indicate specimen collection problems, inhibition, equipment instability, or insufficient training. Delayed results may reflect batching, transport, staffing, or reporting barriers rather than analytical performance. Review these patterns with clinical and public health partners so that laboratory improvements translate into faster diagnosis, appropriate treatment, and stronger TB surveillance.
Practical recommendations before launch
- Approve the intended use, specimen types, patient population, and reporting rules before collecting validation data.
- Set measurable acceptance criteria in advance for accuracy, precision, detection capability, invalid rates, and turnaround time.
- Include local specimens and realistic operating conditions instead of relying only on manufacturer-generated evidence.
- Investigate every discordant, invalid, and unexpected result, with documented corrective action and final review.
- Link authorization to staff competency, biosafety controls, controlled procedures, internal quality control, and ongoing external assessment.
A well-designed validation protects patients, laboratory staff, and public health programs from unreliable results. It also gives the laboratory a defensible basis for adopting useful technology in its own setting, including settings where staffing, infrastructure, specimen transport, and supply chains are constrained.
Use the GLI Quality Tool resources to turn the protocol into controlled practice: define the evidence, complete the required checks, document the decision, and review performance after implementation. Start the validation with a signed scope and risk assessment, then move the new test into service only when the evidence and quality controls support that decision.