Elsevier

Social Science & Medicine

Volume 66, Issue 2, January 2008, Pages 492-505
Social Science & Medicine

The social epidemiology of tuberculosis in South Africa: A multilevel analysis

https://doi.org/10.1016/j.socscimed.2007.08.026Get rights and content

Abstract

Increased risk of tuberculosis is widely recognized to be associated with increased poverty, yet there have been few analyses of the social determinants of tuberculosis, particularly in high-burden settings. We conducted a multilevel analysis of self-reported tuberculosis disease in a nationally representative sample of South Africans based on the 1998 Demographic and Health Survey (DHS). Individual and household-level demographic, behavioral and socioeconomic risk factors were taken from the DHS; data on community-level socioeconomic status (including measures of absolute wealth and income inequality) were derived from the 1996 national census. Of the 13,043 DHS respondents, 0.5% reported having been diagnosed with tuberculosis disease in the past 12 months and 2.8% reported having been diagnosed with tuberculosis disease in their lifetime. In a multivariate model adjusting for demographic and behavioral risk factors, tuberculosis diagnosis was associated with cigarette smoking, alcohol consumption and low body mass index, as well as a lower level of personal education, unemployment and lower household wealth. In a model including individual- and household-level risk factors, high levels of community income inequality were independently associated with increased prevalence of tuberculosis (adjusted odds ratio for lifetime tuberculosis comparing the most unequal quintile to the middle quintile of inequality: 2.37, 95% confidence interval: 1.59–3.53). These results provide novel insights into the socioeconomic determinants of tuberculosis in developing country settings, although the mechanisms through which income inequality may affect tuberculosis disease require further investigation.

Introduction

Tuberculosis is the world's leading curable cause of infectious disease mortality, with a disproportionate burden of disease falling on low- and middle-income countries (World Health Organisation, 1999). Globally the incidence of tuberculosis is increasing, fueled in part by the concurrent epidemic of HIV/AIDS, particularly in sub-Saharan Africa. Tuberculosis is traditionally regarded as a disease of poverty and many aspects of low socioeconomic status (SES), for example overcrowding and malnutrition, are accepted individual- and household-level risk factors for the disease. The distribution of tuberculosis within populations has also been associated historically with group-level socioeconomic conditions (Dubos & Dubos, 1992; Gandy & Zumla, 2003; McKeown, 1976).

South Africa has a long history of high tuberculosis burden. In 2005, the estimated national incidence of tuberculosis was 600 cases per 100,000 persons per annum, the third highest rate in the world (World Health Organisation, 2007). As in other countries, the tuberculosis epidemic in South Africa can be traced in part to the poor working and living conditions that surrounded early industrialization. Many of these conditions persist to this day, and along with the HIV/AIDS epidemic, have fueled the current high levels of tuberculosis disease in this country (Packard, 1989).

Despite the acknowledgment that tuberculosis has strong socioeconomic determinants, there has been surprisingly little epidemiological research into the pathways through which SES might increase the risk of tuberculosis. Ecologic studies conducted in the United States and Britain have found crude associations between tuberculosis rates across areas and low levels of education, high levels of poverty, less government social support, social deprivation and income inequality (Krieger, Waterman, Chen, Soobader, & Subramanian 2003; Parslow, El-Shimy, Cundall, & McKinney, 2001; Spence, Hotchkiss, Williams, & Davies, 1993; Tocque et al., 1999). Evidence from ecologic studies in Brazil and South Africa supports the existence of this relationship in middle-income countries (Munch et al., 2003; Souza et al., 2000). Finally, individual-level studies of the link between low SES and high risk of tuberculosis have found associations in poorer, high tuberculosis-burden settings, such as southern India and Guinea-Bissau (Gustafson et al., 2004; Shetty, Shemko, Vaz, & D’Souza, 2006), although not in Malawi (Glynn et al., 2000).

Recent decades have seen increasing attention in public health paid to quantitative study of social and economic factors as determinants of health outcomes. This field of social epidemiology recognizes the importance of both individual- and group-level socioeconomic risk factors in determining an individual's health (Berkman & Kawachi, 2000). In particular, it differentiates between two classes of mechanisms by which observed group-level associations may operate causally: a compositional effect occurs when the group-level outcome is due solely to the aggregation of individual-level characteristics of members, while a contextual effect arises when the nature of the environment in which group members live determines their health outcomes (MacIntyre, Ellaway, & Cummins, 2002). Examples of the former include the average number of years of education and mean income in an area; examples of the latter include the level of public services provided by government and the distribution of wealth in an area. Research has tended to focus on compositional effects, perhaps because these can be calculated directly through the aggregation of individual-level data, something not possible for contextual phenomena which are fundamentally properties of groups. In particular, studies of health involving income-based community-level effects more frequently consider poverty in absolute than in relative terms.

Existing epidemiological insights into the relationships between SES and tuberculosis risk come almost exclusively from either individual- or group-level (ecological) studies, yet such studies are unable to distinguish the compositional and contextual effects of socioeconomic factors on tuberculosis risk. A comprehensive investigation of the relationship between SES and tuberculosis needs to jointly consider the effect of both individual- and community-level measures of SES. Such an approach requires a multilevel analysis, incorporating variables at different levels of aggregation, to differentiate between compositional and contextual health effects (Diez-Roux & Aiello, 2005). To better understand the potential compositional and contextual influences of SES on tuberculosis risk, we conducted a multilevel analysis of the relative importance of demographic, behavioral and socioeconomic individual-level risk factors, alongside group-level measures of SES, in determining tuberculosis outcomes in South Africa. Our primary hypothesis was that group-level income inequality is an important determinant of tuberculosis disease, independent of typically used measures of SES.

Section snippets

Methods

This secondary analysis combined data from two cross-sectional surveys, the 1998 South African Demographic and Health Survey (SADHS) and the 1996 South African national census. Demographic and Health Surveys (DHS) are an international series of nationally representative surveys conducted in middle- and lower-income countries (Fisher & Way, 1988). The 1998 SADHS was the first DHS conducted in South Africa and sampling was adjusted to ensure statistically meaningful estimates could be reached for

Results

The study sample consisted of the 13,043 respondents (94.3% of the SADHS population) with information on all variables of interest (Table 1). The sample was 59% female, 75% African, 11% Colored, 10% White and 4% Asian.

In this sample 69 individuals (0.5%) reported having been told they had tuberculosis in the past 12 months. After adjusting for the survey design, the population incidence rate was 422 cases per 100,000 per annum (95% CI: 300–543). A total of 369 respondents (2.8%) reported having

Discussion

This analysis of a nationally representative sample of the South African population found the prevalence of self-reported tuberculosis disease to be associated with lower individual-, household- and community-level SES. A low level of personal education, unemployment and a low level of household wealth were associated with increased odds of tuberculosis. Community-level employment and wealth did not appear to have an independent effect on tuberculosis occurrence once individual and household

Conclusion

This study presents novel insights into the associations between socioeconomic conditions and risk of tuberculosis disease. The use of a social epidemiology framework to analyze risk factors for tuberculosis, particularly in sub-Saharan Africa, is a young endeavor. These findings require further investigation, but highlight the importance of including measures of community-level SES, particularly emergent characteristics such as income inequality, in future research on tuberculosis. Despite

Acknowledgments

The authors would like to thank Dr. Debbie Bradshaw of the South African Medical Research Council and Piet Alberts of Statistics South Africa for their assistance in accessing the South African Demographic and Health Survey and South African Census data, respectively. They would also like to thank Katie Doolan for her advice on data sources, methodology and her comments on a draft of the paper.

References (51)

  • J.R. Glynn et al.

    Tuberculosis: Associations with HIV and socioeconomic status in rural Malawi

    Transactions of the Royal Society of Tropical Medicine and Hygiene

    (2000)
  • D.R. Holtgrave et al.

    Social determinants of tuberculosis case rates in the United States

    American Journal of Preventative Medicine

    (2004)
  • S. MacIntyre et al.

    Place effects on health: How can we conceptualise, operationalise and measure them?

    Social Science & Medicine

    (2002)
  • Actuarial Society of South Africa (2005). AIDS Demographic Model 2003 Lite. Cape Town: Actuarial Society of South...
  • Bank of Canada (n.d.). Year average of exchange rates. Ottawa: Bank of Canada. Retrieved 1 February 2007 from...
  • L.F. Berkman et al.

    A historical framework for social epidemiology

  • F.R. Booysen

    Using demographic and health surveys to measure poverty: An application to South Africa

    Journal for Studies in Economics and Econometrics

    (2002)
  • A. Brandt

    No magic bullet: A social history of venereal disease in the United States since 1880

    (1987)
  • A.S. Bryk et al.

    Hierarchical linear models: Applications and data analysis methods

    (1992)
  • M. Chan-Yeung et al.

    Socio-demographic and geographic indicators and distribution of tuberculosis in Hong Kong: A spatial analysis

    International Journal of Tuberculosis and Lung Disease

    (2005)
  • J.N. Claasen

    The benefits of the CAGE as a screening tool for alcoholism in a closed rural South African community

    South African Medical Journal

    (1999)
  • G. Davey-Smith et al.

    Education and occupational social class: Which is the more important indicator of mortality risk?

    Journal of Epidemiology and Community Health

    (1998)
  • A. Deaton

    Health, inequality and economic development

    Journal of Economic Literature

    (2003)
  • A.V. Diez-Roux et al.

    Multilevel analysis of infectious diseases

    The Journal of Infectious Diseases

    (2005)
  • R. Dubos et al.

    The white plague: Tuberculosis, man, and society

    (1992)
  • P. Farmer

    Infections and inequalities: The modern plagues

    (1999)
  • A.A. Fisher et al.

    The Demographic and Health Surveys program: An overview

    International Family Planning Perspectives

    (1988)
  • P. Gustafson et al.

    Tuberculosis in Bissau: Incidence and risk factors in an urban community in sub-Saharan Africa

    International Journal of Epidemiology

    (2004)
  • W.M. Jakubowiak et al.

    Risk factors associated with default among new pulmonary TB patients and social support in six Russian regions

    International Journal of Tuberculosis and Lung Disease

    (2007)
  • I. Kawachi

    Income inequality and health

  • I. Kawachi et al.

    Social cohesion, social capital, and health

  • I. Kawachi et al.

    When economists and epidemiologists disagree

    Journal of Health Politics, Policy and Law

    (2001)
  • B.P. Kennedy et al.

    Income distribution and mortality: Cross sectional ecological study of the Robin Hood Index in the United States

    British Medical Journal

    (1996)
  • N. Krieger et al.

    Monitoring socioeconomic inequalities in sexually transmitted infections, tuberculosis and violence: Geocoding and choice of area-based socioeconomic measures

    Public Health Reports

    (2003)
  • Cited by (127)

    • Ending tuberculosis in a post-COVID-19 world: a person-centred, equity-oriented approach

      2023, The Lancet Infectious Diseases
      Citation Excerpt :

      These disparities have led to socioeconomic, racial, and ethnic inequities in COVID-19-related outcomes.8 Disparities of a similar nature have long been identified as the fuel that feeds tuberculosis epidemics.9–11 Unfortunately, as illustrated by the COVID-19 response, reliance on existing health systems generally results in a deepening of disparities.12,13

    • Tobacco use and cessation in the context of ART adherence: Insights from a qualitative study in HIV clinics in Uganda

      2021, Social Science and Medicine
      Citation Excerpt :

      To address Finding 5, some of the fundamental causes of anxiety and depression i.e. food insecurity and poverty should be addressed through livelihood support initiatives (Tsai et al., 2012), and tobacco users screened and treated for depression. Following our third finding and given the high incidence and life-threatening nature of TB among people living with HIV in Sub-Saharan Africa, HIV staff training and health education materials could usefully focus on the link between tobacco and alcohol use and TB (Jackson-Morris et al., 2015; van Zyl Smit et al., 2010; Harling et al., 2008; Necho et al., 2021). Stricter tobacco control measures would also support people living with HIV in quitting tobacco and decrease the chances of their relapsing to smoking.

    View all citing articles on Scopus
    View full text