RESUMO
OBJECTIVE: To compare the cost-utility of microscopic observation drug-susceptibility assay (MODS) and Xpert® MTB/RIF implementation for tuberculosis (TB) diagnosis in rural northern Mozambique. METHODS: Stochastic transmission compartmental TB model from the healthcare provider perspective with parameter input from direct measurements, systematic literature reviews and expert opinion. MODS and Xpert® MTB/RIF were evaluated as replacement test of smear microscopy (SM) or as an add-on test after a negative SM. Costs were calculated in 2013 USD, effects in disability-adjusted life years (DALY). Willingness to pay threshold (WPT) was established at once the per capita Gross National Income of Mozambique. RESULTS: MODS as an add-on test to negative SM produced an incremental cost-effectiveness ratio (ICER) of 5647.89USD/DALY averted. MODS as a substitute for SM yielded an ICER of 5374.58USD/DALY averted. Xpert® MTB/RIF as an add-on test to negative SM yielded ICER of 345.71USD/DALY averted. Xpert® MTB/RIF as a substitute for SM obtained an ICER of 122.13USD/DALY averted. TB prevalence and risk of infection were the main factors impacting MODS and Xpert® MTB/RIF ICER in the one-way sensitivity analysis. In the probabilistic sensitivity analysis, Xpert® MTB/RIF was most likely to have an ICER below the WPT, whereas MODS was not. CONCLUSION: Our cost-utility analysis favours the implementation of Xpert® MTB/RIF as a replacement of SM for all TB suspects in this rural high TB/HIV prevalence African setting.