RESUMEN
BACKGROUND: Tuberculosis continues to be a leading cause of infectious disease mortality, and effective screening and diagnosis remains crucial. Despite progress made, diagnostic gaps remain due to poor access to diagnostic tools and testing, particularly in rural and remote areas. As such, the development of target product profiles is essential in guiding the development of new diagnostic tools, however target product profiles often lack evidence-based information and do not consider trade-offs between test accuracy and accessibility. METHODS: A simulation-based model, in the form of a decision tree, was used to map out the baseline patient tuberculosis diagnostic pathway for individuals in Kenya, South Africa, and India. The model was then used to adapt this pathway to evaluate the trade-offs between increased access to testing and varying accuracy of new tuberculosis diagnostic tools within the health-care contexts of Kenya, South Africa, and India. The model aims to support target product profile development by quantifying the impact of new diagnostics on the standard of care. The model considered three diagnostic attributes, namely sample type (sputum vs non-sputum), site of testing (point of care, near point of care, and health setting) and turnaround time. FINDINGS: Our results indicate that per sample type, novel point-of-care tests would be the most accessible and even with lower sensitivities can achieve comparable or better case detection than the current standard of care in each country. Non-sputum diagnostics also have lower sensitivity requirements. Overall, target product profile parameters with reduced sensitivities from 70% for non-sputum and 78% for sputum tests could be accepted. INTERPRETATION: Diagnostics which bring tuberculosis tests and test results closer to the patient could reduce overall diagnostic loss despite potential reductions in sensitivity. This work provides a novel framework for guiding the future development of diagnostics, with an approach towards balancing accessibility and test performance. FUNDING: The Bill and Melinda Gates Foundation (INV-045721).