Fig 1.
Xpert results at time of diagnosis, and eventual Xpert and sputum culture results, among empirically treated patients.
No other bacteriologic testing such as sputum smear examination was used in diagnosing these individuals.
Table 1.
Characteristics of TB cases and of TB-negative controls, as classified by clinicians after evaluation for presumptive pulmonary TB.
Fig 2.
Xpert results and treatment decisions, by sputum culture result.
Red dashed outlines highlight the patients for whom culture results were discordant with their previous diagnosis, including 4 with positive culture who had not been treated (false-negative Xpert) and 27 with negative cultures who had been treated either based on Xpert or empirically. For some patients with positive Xpert, the decision to treat was made before the Xpert result became available; these patients are classified as “Positive Xpert” here, but they are also included in Fig 1 which describes all empirically treated cases. Controls were included in this figure only if their matched case also had Xpert and culture results.
Table 2.
Sensitivity and specificity of Xpert and clinical judgment, relative to sputum culture.
Fig 3.
Key characteristics of patients with culture-concordant and culture-discordant clinical TB diagnoses.
Error bars represent 80% binomial confidence intervals. Patients with negative evaluations (controls) but subsequent positive cultures are excluded, as are individuals with no culture result and the matched controls of cases with no culture result. Numerical values are also listed in S6 Table.