Background: Even when tuberculosis (TB) treatment is free, hidden costs incurred by patients and their households (TBaffected
households) may worsen poverty and health. Extreme TB-associated costs have been termed ‘‘catastrophic’’ but are
poorly defined. We studied TB-affected households’ hidden costs and their association with adverse TB outcome to create a
clinically relevant definition of catastrophic costs.
Methods and Findings: From 26 October 2002 to 30 November 2009, TB patients (n = 876, 11% with multi-drug-resistant
[MDR] TB) and healthy controls (n = 487) were recruited to a prospective cohort study in shantytowns in Lima, Peru. Patients
were interviewed prior to and every 2–4 wk throughout treatment, recording direct (household expenses) and indirect (lost
income) TB-related costs. Costs were expressed as a proportion of the household’s annual income. In poorer households,
costs were lower but constituted a higher proportion of the household’s annual income: 27% (95% CI = 20%–43%) in the
least-poor houses versus 48% (95% CI = 36%–50%) in the poorest. Adverse TB outcome was defined as death, treatment
abandonment or treatment failure during therapy, or recurrence within 2 y. 23% (166/725) of patients with a defined
treatment outcome had an adverse outcome. Total costs $20% of household annual income was defined as catastrophic
because this threshold was most strongly associated with adverse TB outcome. Catastrophic costs were incurred by 345
households (39%). Having MDR TB was associated with a higher likelihood of incurring catastrophic costs (54% [95%
CI = 43%–61%] versus 38% [95% CI = 34%–41%], p,0.003). Adverse outcome was independently associated with MDR TB
(odds ratio [OR] = 8.4 [95% CI = 4.7–15], p,0.001), previous TB (OR = 2.1 [95% CI = 1.3–3.5], p = 0.005), days too unwell to
work pre-treatment (OR = 1.01 [95% CI = 1.00–1.01], p = 0.02), and catastrophic costs (OR = 1.7 [95% CI = 1.1–2.6], p = 0.01).
The adjusted population attributable fraction of adverse outcomes explained by catastrophic costs was 18% (95%
CI = 6.9%–28%), similar to that of MDR TB (20% [95% CI = 14%–25%]). Sensitivity analyses demonstrated that existing
catastrophic costs thresholds ($10% or $15% of household annual income) were not associated with adverse outcome in
our setting. Study limitations included not measuring certain ‘‘dis-saving’’ variables (including selling household items) and
gathering only 6 mo of costs-specific follow-up data for MDR TB patients.
Conclusions: Despite free TB care, having TB disease was expensive for impoverished TB patients in Peru. Incurring higher
relative costs was associated with adverse TB outcome. The population attributable fraction indicated that catastrophic
costs and MDR TB were associated with similar proportions of adverse outcomes. Thus TB is a socioeconomic as well as
infectious problem, and TB control interventions should address both the economic and clinical aspects of this disease.