Incremental Cost Effectiveness of Bedaquiline for the Treatment of Rifampicin-Resistant Tuberculosis in South Africa: Model-Based Analysis

Background Nearly 20,000 people were diagnosed with multi-drug and rifampicin-resistant tuberculosis (MDR/RR-TB) in South Africa in 2015, yet only one-half of the patients who start treatment are expected to have a successful outcome. There is increasing evidence of the effectiveness and safety of new drug regimens containing bedaquiline for MDR/RR-TB; however, whether they are affordable for high-burden, limited-resource settings is uncertain. Objective Our objective was to determine the incremental cost effectiveness of a bedaquiline-based regimen for MDR/RR-TB treatment in South Africa compared with the standard kanamycin-based regimen. Methods We established a Markov model for ambulatory treatment of MDR/RR-TB in a high-HIV prevalence setting, parameterized using clinical outcomes from the South African National TB Programme (SA NTP) before (2012-2014) and after (2015-2016) bedaquiline roll-out. The effectiveness of treatment was evaluated in disability-adjusted life-years (DALYs). Ingredient costs from the provider's perspective were collected in 2016 South African Rand and converted to $US, including bedaquiline at $US675.23 per 6-month treatment course. Culture conversion rates were derived from the phase IIb trial of bedaquiline, and disability adjustments were adapted from published literature. Costs and effectiveness were discounted at 3%. Results For non-bedaquiline regimens, the total expected cost over the 10-year time horizon for a patient with MDR/RR-TB was $US4439 with disability-adjusted survival of 5.1 years. Replacing capreomycin with bedaquiline in patients who failed MDR/RR-TB treatment and required treatment for extensively drug-resistant (XDR-TB) resulted in cost savings ($US4356; 1.8% less) and similar effectiveness (0.02 DALYs averted). As a result, the standard regimen (no bedaquiline) was dominated. Replacing kanamycin with bedaquiline to provide all patients with MDR/RR-TB access to bedaquiline cost $US4647 (4.3% more) and averted 0.17 DALYs compared with the no bedaquiline regimen. The incremental cost-effectiveness ratio was $US1242/DALY averted. Conclusion Markov modelling indicates providing bedaquiline for all patients with MDR/RR-TB could increase the 24-month treatment success rate in South Africa from 56.3% using the current regimen to 60.6%, at a cost $US2.6 million over a 10-year horizon, less than 1% of the estimated $US425 million SA NTP annual budget.