3.1.2 Implementation considerations for all tools

The GDG considered that all three approaches – symptom screening, CXR and mWRD – may have roles in screening for TB disease in the general population. The ranking of the tools according to accuracy proposed by the GDG is: first, CXR; second, mWRDs; third, screening for any TB symptom (higher sensitivity and lower specificity); and fourth, screening for any cough or cough lasting 2 weeks or longer (lower sensitivity and higher specificity). However, ease of implementation is highest by far for symptom screening, and less so for CXR and mWRDs with the requirement of additional equipment and resources. The GDG noted that factors related to patient selection, flow and timing may affect the measures of accuracy observed for CXR to detect TB. The inferences made about using mWRDs for screening in the general population relate primarily to pooled studies in high-risk populations and, therefore, there is uncertainty about whether the findings are directly applicable to a general population with a comparable burden of TB. See Web Annex C, Table 1, for more details of the recommendation.

The GDG noted that different symptom screening approaches have varying trade-offs in sensitivity and specificity. The feasibility of implementing symptom screening makes it a much more accessible option programmatically. Symptom screening is a standard part of a clinical workup and can be repeated as often as it is needed.

In contrast, additional resources are needed to undertake chest radiography and mWRDs. Radiography involves exposure to some ionizing radiation, which may increase the long-term risks of cancer. Innovations in radiography in recent years have substantially reduced radiation exposure levels. CXR is largely considered safe, with a radiation dose of 0.1 mSv, which corresponds to 1/30 of the average annual radiation dose from the environment (3 mSv) and 1/10 of the annual accepted dose of ionizing radiation for the general public (1 mSv). Pregnant women are especially vulnerable to ionizing radiation from radiography, and children have a longer life expectancy and, therefore, more time to develop radiation-induced health effects. However, for a pregnant woman and her fetus and for children, CXR has been deemed to not pose a significant risk, provided that good practices are observed, as the primary beam is targeted away from the pelvis (68).

When used for screening, mWRDs have different accuracy than when they are used for diagnosis (69), and different predictive values are associated with a positive test and a negative test due to differences in prevalence of TB in the populations being tested. Therefore, results should be interpreted appropriately, and those who screen positive for TB using an mWRD should receive a thorough clinical evaluation that may include further tests and procedures – such as CXR, repeat mWRD on additional sputum samples and other examinations – to establish a diagnosis of TB definitively prior to initiating treatment. These tests can be used on sputum only when a person can expectorate. Scaling up mWRDs for diagnosis should be prioritized (if full access has not yet been achieved) prior to scaling up mWRDs for screening. Using an mWRD as a screening tool requires significant resources, including increased capacity in and expansion of diagnostic and sample transportation networks. Depending on the feasibility and resources available, countries may choose to prioritize TB screening using mWRDs among certain subpopulations with a higher risk of TB.

Countries should position symptom screening, CXR and mWRDs within national TB screening and diagnostic algorithms, according to the goals and objectives of the screening, the populations being screened, the feasibility, resources available, and equity. A range of possible screening algorithms is presented in the operational handbook, including modelled performance of the expected yield.

Book navigation