TB Retreatment and MDR-TB Treatment.

aThe actual proportion of patients that are able to access either re-treatment or MDR-TB treatment is not reported for Uganda. However, we based our model on optimal access to care and varied this parameter in sensitivity analyses.bThe measures of treatment success in our analysis is different from the original publications by Jones-Lopez [13] and Seung [15]. We simplistically assumed that all surviving patients experienced a successful treatment outcome in our model and applied the full cost of TB re-treatment or MDR-TB treatment to surviving patients.cDepartment of Medical Microbiology, Makerere University Kampala, Uganda.dUgandan National Tuberculosis and Leprosy Program (NTLP) reported in 2010.eInternal analysis conducted at the Infections Diseases Institute in Kampala, Uganda.fRHZE, rifampicin, isoniazid, pyrazinamide, ethambutol.gRHE, rifampicin, isoniazid, ethambutol.hFor HIV-positive and HIV-negative patients combined, the mortality rate in Jones-Lopez [13] was reported as 7, 17, and 14 in the periods of 0–2 months, 2–5 months, and 5–8 months, respectively. Since more detailed information on the timing of deaths was not available, we assumed that these patients would have incurred half of the relevant treatment costs.iIn Seung [15], death occurred after a median 66 days in treatment. Among surviving patients, the median duration of treatment was reported at 252 days, thus, we assumed that non-surviving patients would consume 26.2% (66/252 = 26.2%) of resources relative to surviving patients.