S1 Data - Higher loss of livelihood and impoverishment in households affected by tuberculosis compared to non-tuberculosis affected households in Zimbabwe: A cross-sectional study

We used principal component analysis on five capital asset indicators to create a binary outcome variable indicating loss of livelihood. Log-binomial regression was used to determine associations between loss of livelihood and type of household. TB-affected households were more likely to report episodes of TB and household members requiring care than non-TB households.

The proportions of impoverished households were 81% (non-TB), 88% (DS-TB) and 94% (DR-TB) by the time of interview. Overall, 56% (152/270) of households sold assets: 44% (40/90) non-TB, 58% (52/90) DS-TB and 67% (60/90) DR-TB. Children’s education was affected in 33% (55/168) of TB-affected compared to 14% (12/88) non-TB households.

Overall, 133 (50%) households experienced loss of livelihood, with TB-affected households almost twice as likely to experience loss of livelihood; adjusted prevalence ratio (aPR = 1.78 [95%CI:1.09–2.89]). The effect of TB on livelihood was most pronounced in poorest households (aPR = 2.61, [95%CI:1.47–4.61]). TB-affected households experienced greater socioeconomic losses compared to non-TB households.

Multisectoral social protection is crucial to mitigate impacts of TB and other shocks, especially targeting poorest households.