The AI model meets and, in some cohorts, exceeds the WHO’s minimal TPP requirements, demonstrating its potential for effective integration into screening workflows for TB detection. The high AUROC across different demographic variables indicated the model’s robust discriminative ability, reinforcing the potential of AI-assisted tools in large-scale TB screening programs.

This economic evaluation provides compelling evidence that AI-assisted TB screening is cost-effective for case detection in resource-limited settings. The integration of artificial intelligence into TB screening programs demonstrated significant advantages over conventional methods, offering a promising strategy to address the substantial TB burden in high-prevalence regions.

TN-KET offers a scalable model for sustained TB death rate reduction. This effect needs to be confirmed through confounders adjusted individual level data analysis and is ongoing.

The evening camps helped to screen the working male population who would otherwise be missed during daytime ACF camps. This innovative time adaptive method demonstrates the importance of tailoring healthcare delivery to the schedules and needs of specific populations, and has also yielded higher case finding as compared to conventional ACF screening camps.

Performing VA by trained health staff and assigning ICD-10 by MO has yielded information on the various comorbidities causing deaths in PwTB. Implementing TB-DSR helps the program managers in addressing comorbidities upfront at the time of initiation of TB treatment which is a key strategy in reducing mortality rates.