FeedbackNone of the screens investigated reached the target product profile of 90% minimum sensitivity in these high-risk subpopulations of children, although CXRs came the closest. Concerns were noted about the risk of incorporation bias when using a composite reference standard in this group, thus potentially inflating the estimates of accuracy observed. Concerns were also noted about the increased risk of false-positive diagnoses in children following a false-positive screening test compared with the risk in adults because children are more likely to be diagnosed using clinical evaluation rather than bacteriological confirmation, and the process of confirming a clinical diagnosis will weight the results of the screening test or tests. When screening high-risk groups of children, including close contacts and children living with HIV, the balance between the benefits of early case detection arising from true-positive screening results and the possible risk of overtreatment from false-positive screening results is in favour of screening. When screening populations of children with a lower risk of TB or in lower prevalence settings, the trade-off between early case detection and possible overdiagnosis will be different and should be carefully considered. In such situations, while a highly sensitive screening approach is important in order to maximize early case detection, health care workers must remain vigilant to possible false-positive diagnoses and monitor responses to treatment carefully, including evaluating children and adolescents for other potential diagnoses if symptoms or CXR abnormalities persist.
Given the high risk of TB and of mortality if TB is left untreated among children with HIV and among those in close contact with TB patients, there is an urgent need to utilize any and all available screening tests to increase timely diagnosis among these high-risk populations.
Children living with HIV represent an important group that should be considered for regular TB screening and the provision of preventive therapy, given their high risk of TB and of poor outcomes if not diagnosed in a timely manner. An essential minimum screening strategy for this group would be to ensure regular screening for TB symptoms at each visit to a health centre. While data were lacking to evaluate CXR as an initial screen for children living with HIV, CXR is a useful part of a diagnostic evaluation for TB in all children, including those living with HIV, especially younger children in whom bacteriological evaluation is commonly negative (68). Children and adolescents who are close contacts of someone with TB likewise represent an important group for screening for TB disease and for the initiation of preventive therapy, given their high risk for developing TB and rapid disease progression.
For children at high risk of TB, countries should position symptom screening and screening chest radiography within national TB screening and diagnostic algorithms, according to their feasibility, level of health facility, resources and equity. Algorithms exploring the different screening tools are presented in the operational handbook (7).