Effects of conditional cash transfers and pre-test and post-test tuberculosis counselling on patient outcomes and loss to follow-up across the continuum of care in South Africa: a randomised controlled trial

Abstract

Economic and behavioural factors lead to poor outcomes in patients with tuberculosis. We investigated the effects of a package of interventions consisting of pre-test and post-test tuberculosis counselling with conditional cash transfers on patient outcomes in adults undergoing investigation for pulmonary tuberculosis. This pragmatic, open-label, individual randomised controlled trial was done in nine clinics in Johannesburg, South Africa. Participants (aged ≥18 years) undergoing investigation for tuberculosis were randomly assigned (1:1) to the intervention group or control group (standard of care) via permuted block randomisation, stratified by clinic; group assignment was concealed using opaque envelopes. The intervention group received pre-test and post-test tuberculosis counselling, and for participants diagnosed with rifampicin-susceptible tuberculosis, a digital payment (R150; approximately US) at treatment initiation and each monthly treatment visit. Payments were contingent on timely attendance: 14 days from initial sputum sample collection and within 7 days on either side of their scheduled monthly appointment. The primary endpoint was successful patient outcome (patients who were cured or completed treatment) or unsuccessful patient outcome (pretreatment loss-to-follow-up, on-treatment loss-to-follow-up, development of rifampicin-resistant tuberculosis while on treatment, treatment failure [ie, smear or culture positive at 5 months or later after commencing treatment], or death). The primary outcome was analysed in the modified intention-to-treat population, defined as all randomly assigned participants with rifampicin-susceptible tuberculosis confirmed before the commencement of tuberculosis treatment. Weighted outcome prevalence, relative risks (RRs), and risk differences were calculated using a multivariable Poisson model with robust standard errors. This trial is registered with the Pan African Clinical Trials Registry (PACTR202410708311054) and is completed.