Quantification and correlates of tuberculosis stigma along the tuberculosis testing and treatment cascades in South Africa: a cross-sectional study.

backgroundSouth Africa has one of the world's worst tuberculosis (TB) (520 per 100 000 population) and TB-human immunodeficiency virus (HIV) epidemics (~ 56% TB/HIV co-infected). While individual- and system-level factors influencing progression along the TB cascade have been identified, the impact of stigma is underexplored and underappreciated. We conducted an exploratory study to 1) describe differences in perceived community-level TB stigma among community members, TB presumptives, and TB patients, and 2) identify factors associated with TB stigma levels among these groups.

methodsA cross sectional study was conducted in November 2017 at public health care facilities in Buffalo City Metro (BCM) and Zululand health districts, South Africa. Community members, TB presumptives, and TB patients were recruited. Data were collected on sociodemographic characteristics, TB knowledge, health and clinical history, social support, and both HIV and TB stigma. A validated scale assessing perceived community TB stigma was used. Univariate and multivariate linear regression models were used to describe differences in perceived community TB stigma by participant type and to identify factors associated with TB stigma.

resultsWe enrolled 397 participants. On a scale of zero to 24, the mean stigma score for TB presumptives (14.7 ± 4.4) was statistically higher than community members (13.6 ± 4.8) and TB patients (13.3 ± 5.1). Community members from Zululand (β = 5.73; 95% CI 2.19, 9.72) had higher TB stigma compared to those from BCM. Previously having TB (β = - 2.19; 95% CI - 4.37, 0.0064) was associated with reduced TB stigma among community members. Understanding the relationship between HIV and TB disease (β = 2.48; 95% CI 0.020, 4.94), and having low social support (β = - 0.077; 95% CI - 0.14, 0.010) were associated with increased TB stigma among TB presumptives. Among TB Patients, identifying as Black African (β = - 2.90; 95% CI - 4.74, - 1.04) and knowing the correct causes of TB (β = - 2.93; 95% CI - 4.92, - 0.94) were associated with decreased TB stigma, while understanding the relationship between HIV and TB disease (β = 2.48; 95% CI 1.05, 3.90) and higher HIV stigma (β = 0.32; 95% CI 0.21, 0.42) were associated with increased TB stigma.

conclusionsTB stigma interventions should be developed for TB presumptives, as stigma may increase initial-loss-to-follow up. Given that stigma may be driven by numerous factors throughout the TB cascade, adaptive stigma reduction interventions may be required.