Patrick P J Phillips, Samuel G Schumacher, Cara Cassino, Salome Charalambous, Richard E Chaisson, Gavin J Churchyard, Francesca Conradie, Christopher Cousins, Geraint Davies, Susan E Dorman, Hanif Esmail, Katherine Fielding, Carole D Mitnick, David McNeeley, Payam Nahid, Karen Sanders, Jodie A Schildkraut, Derek J Sloan, Gustavo E Velásquez, Michael Hoelscher, Andrew A Vernon, Christian Lienhardt
Tuberculosis remains a major global health challenge, despite recent advances in drug and regimen development. Pragmatic randomised trials are needed to evaluate the effectiveness, safety, and tolerability of new tuberculosis treatments under routine care conditions to appropriately inform practice guidelines and facilitate uptake. In this Review, we propose guiding principles for pragmatic tuberculosis treatment trials. Using the PRECIS-2 framework, we address eligibility, recruitment, setting, organisation of care, adherence support, outcomes, follow-up, primary analysis, data collection, and monitoring. Pragmatic trials should maximise generalisability through broad inclusion criteria, integration within routine health systems, and flexible delivery approaches while maintaining appropriate standards for participant safety and data reliability. We discuss considerations for individually randomised and cluster-randomised designs, outcome definitions, risk-proportionate monitoring, and streamlined data collection. Adoption of these principles will improve the generation of real-world evidence and facilitate global implementation of effective tuberculosis treatments.