ABSTRACT
Traditional diagnostic methods, such as culture or smear microscopy, are slow or low in sensitivity. More sensitive modern techniques, such as Xpert MTB/RIF, require a certain infrastructure, are costly, and are not widely accessible. Diagnostic accuracy was determined separately for each cohort, and 95% confidence intervals were computed using Wilson's score method. Sensitivity and specificity of LF-LAM and SILVAMP-LAM for each cohort were compared using the McNemar test. The inclusion of different cohorts with different study designs resulted in heterogeneity and different exclusion rates; however, differences in diagnostic sensitivity between SILVAMP-LAM and LF-LAM persisted in a sensitivity analysis including the “unclassifiable” category and in the analyses by CD4 subgroup. Further, antiretroviral treatment may have influenced patient outcome and thus the reference standard categories “possible TB” and "unclassifiable". Limited budgets, lack of country-specific data, administrative hurdles such as local regulatory approval, lack of coordination between national TB and HIV programs, and small perceived patient population size.
