Abstract
Background: Tuberculosis (TB) remains a leading cause of death among people living with HIV (PLWH), yet diagnostic methods vary in accuracy, accessibility, and implementation. Understanding how diagnostic modality influences TB detection is essential to optimizing co-infection management. Methods: We conducted a retrospective analysis of institutional data from Bamrasnaradura Infectious Diseases Institute (BIDI), Thailand, covering 2016-2023. TB detection rates were assessed across five diagnostic methods-chest radiography (CXR), smear microscopy, acid-fast bacilli (AFB) staining, culture, and GeneXpert MTB/RIF-relative to annual HIV-related visit volumes. Results: Among 56,599 HIV-related visits, TB detection rates varied substantially by diagnostic method. CXR was the most commonly used tool, detecting TB in up to 99 cases out of 6964 visits (1.42%) in 2016, though declining to 23 cases out of 6947 visits (0.33%) in 2023. GeneXpert was employed more consistently, yielding between 7 cases out of 7577 visits (0.09%) and 13 cases out of 6593 visits (0.20%) annually. Smear microscopy and AFB staining declined markedly, falling below 0.22% after 2020. These patterns reflect a gradual transition toward molecular diagnostics, which offer improved accuracy but remain underutilized in lower-tier settings. To address these gaps, we incorporated trend analyses confirming significant temporal shifts and propose a tiered TB screening framework tailored to resource availability across healthcare levels. Conclusions: TB detection among PLWH is strongly influenced by the diagnostic method used. Unlike HIV diagnosis-which is definitive and standardized-TB diagnosis remains fragmented and resource-dependent. Context-sensitive screening protocols are urgently needed to improve TB case detection and management, particularly in lower-level HIV care facilities.