TB and poverty

TB and Poverty

Tuberculosis, a respiratory and infectious disease, has historically been associated with poverty. A number of studies indicate that two broad mechanisms explain the relationship between poverty and TB incidence: the likelihood of being exposed to the disease and the immunological status of the individual.

The aim of this paper is to identify the intermediary factors between poverty and tuberculosis in South Africa using a large representative household survey. Our contention is that the identification of these intermediary risk factors could inform the debate on where best to target policy interventions.

We use a pooled version of Statistics South Africa’s General Household Survey 2002 – 2011. Due to the large sample size achieved by pooling these surveys, this combined version of the GHS provides a rare opportunity to use a collection of representative household surveys to better understand the relationship between poverty and the prevalence patterns of this disease.

Poverty quintiles are estimated using household assets and characteristics associated with poverty to compile an index (employing Multiple Correspondence Analysis). We then add intermediary risk variables associated to an initial baseline regression model (with poverty as the only predictor of tuberculosis infection) to determine how much they contribute to tuberculosis prevalence.

Descriptive analysis shows that those in the poorest quintile are 9 times more likely to report that they are suffering from TB or associated symptoms than those in the top quintile.

Regression analysis shows that TB prevalence has a positive and significant association with urban location, age, being black or Coloured, living in the Eastern Cape, Free State, KwaZulu-Natal or Northwest province (compared to the base case which is the Western Cape) and overcrowded living conditions. More education, being female and living in Limpopo (rather than the Western Cape) significantly reduce the likelihood of having TB. The addition of these intermediary factors significantly reduces the coefficient on poverty and provides preliminary evidence to show that policies targeting these specific transmission channels may significantly reduce the vulnerability of the poor to this disease. We find no evidence of significant differences in access to care and also no significant patterns in the reported levels of satisfaction across poverty quintiles for TB sufferers. TB sufferers from poor households are less likely to consult doctors and to utilise private care.

The work has been presented at the International Health Economics Association conference in Sydney in July 2013. The work is due to be published as a working paper soon. Please email carinebrunsdon@sun.ac.za if you want to be included on our health group’s distribution list for new working papers.


ReSEP members involved: Ronelle Burger, Eldridge Moses