Introduction: Nucleic acid amplification tests to detect Mycobacterium tuberculosis in clinical specimens are used increasingly as a laboratory tool. We aimed to investigate the routine using pattern and the effects on therapeutic decision of diagnostic tests for tuberculosis in our hospital. Methods: In this descriptive study, we investigated retrospectively the routine using pattern and the effects on therapeutic decision of diagnostic tests for tuberculosis. Patients with discordant results were clinically evaluated retrospectively by a chest physician. Samples were tested for the presence of M. tuberculosis by a smear technique, M. tuberculosis culture growth technique (Lowenstein-Jensen and/or BACTEC-960), and IS6110 polymerase chain reaction (PCR). Results: Culture positivity was 7.2% (83 of 1159 patients). In total, 198 (62.4%) were tested with PCR, acid-fast bacilli, and culture. On the basis of culture results as a gold standard, sensitivity, specificity, positive predictive value, and negative predictive value of PCR were 46%, 89%, 23%, and 93.5%, respectively. Conclusions: Selection of appropriate patients for further testing and exclusion of low-risk patients from microbiologic testing by experienced clinicians may help to optimize the positive predictive value of PCR.