Performance of clinical scoring and microscopic combinations for diagnosing pulmonary tuberculosis with the GenXpert criteria

Abstract

Pulmonary tuberculosis is a prominent health issue in Indonesia, which ranks second in the world regarding the number of patients. Rapid tuberculosis detection is crucial for early treatment, a better prognosis, and a reduction in disease transmission; however, the availability of molecular rapid tests is limited. Cross-sectional design and retrospective analyses of pre-pandemic data from 723 patients with suspected pulmonary tuberculosis from 2017 to 2019 were conducted in this study. The study aimed to assess the performance of clinical scoring and microscopic examination in tuberculosis diagnosis at RSUD Sayang Cianjur. The effectiveness of sequential (two-stage) and simultaneous combinations of clinical scoring and sputum smear microscopic were investigated. Performance assessments consisted of 2x2 tables, calculation according to Gordis, and Receiver Operator Characteristic (ROC) analysis, with GenXpert results as the gold standard. The results showed lower performance of the individually performed scoring system, with clinical scoring having a sensitivity of 34.44% and a specificity of 97.15%. Microscopic Acid Fast Bacteria (AFB) had a sensitivity of 70.20% and a specificity of 98.57%. The net sensitivity of the sequential combination was 28.48%, and the specificity was 99.05%. The net sensitivity of the simultaneous combination was 77.15%, and the specificity was 95.72%. The area under the curve from the sequential diagnostic method was 0.728, and the area under the curve of the simultaneous diagnostic was 0.884. The sequential and simultaneous combinations of clinical scoring and the AFB microscopy improved the test performance. The simultaneous combination performed slightly better than the sequential combination.