Background: Maximising the efficiency of national tuberculosis programmes is key to improving service coverage, outcomes, and progress towards End TB targets. We aimed to determine the overall efficiency of tuberculosis spending and investigate associated factors in 121 low-income and middle-income countries between 2010 and 2019. Methods: In this data envelopment and stochastic frontier analysis, we used data from the WHO Global TB report series on tuberculosis spending as the input and treatment coverage as the output to estimate tuberculosis spending efficiency. We investigated associations between 25 independent variables and overall efficiency. Findings: We estimated global tuberculosis spending efficiency to be between 73·8% (95% CI 71·2–76·3) and 87·7% (84·9–90·6) in 2019, depending on the analytical method used. This estimate suggests that existing global tuberculosis treatment coverage could be increased by between 12·3% (95% CI 9·4–15·1) and 26·2% (23·7–28·8) for the same amount of spending. Efficiency has improved over the study period, mainly since 2015, but a substantial difference of 70·7–72·1 percentage points between the most and least efficient countries still exists. We found a consistent significant association between efficiency and current health expenditure as a share of gross domestic product, out-of-pocket spending on health, and some Sustainable Development Goal (SDG) indicators such as universal health coverage. Interpretation: To improve efficiency, treatment coverage will need to be increased, particularly in the least efficient contexts where this might require additional spending. However, progress towards global End TB targets is slow even in the most efficient countries. Variables associated with TB spending efficiency suggest efficiency is complimented by commitments to improving health-care access that is free at the point of use and wider progress towards the SDGs. These findings support calls for additional investment in tuberculosis care. Funding: None.
The Lancet Global Health, Volume 10, May 2022,