Re: S Gabrysch et al: The role of context: neighbourhood characteristics strongly influence HIV risk in young women in Ndola, Zambia. TMIH 2008, February issue
Article first published online: 6 MAY 2008
© 2008 Blackwell Publishing Ltd
Tropical Medicine & International Health
Volume 13, Issue 7, page 959, July 2008
How to Cite
Muula, A. S. (2008), Re: S Gabrysch et al: The role of context: neighbourhood characteristics strongly influence HIV risk in young women in Ndola, Zambia. TMIH 2008, February issue. Tropical Medicine & International Health, 13: 959. doi: 10.1111/j.1365-3156.2008.02097_1.x
- Issue published online: 6 MAY 2008
- Article first published online: 6 MAY 2008
Gabrysch et al (2008) reported on the importance of context in our understanding of HIV in Zambia by their secondary analysis and presenting of data. While such direction of focus should be welcomed, there are a number of areas where the reader may wish for clarity in conceptualization of the problem and presentation.
The first limitation of the analysis is that the authors focus on “neighborhoods” without explicitly giving their definition of the term. Neighborhoods can be demarcated by mere geography, by administrative or political criteria, by historical attributes or by the study participants’ own definition of what constitutes their neighborhood. It seems the authors did not sufficiently consider these differences (Diez Roux 2001). In HIV research, a neighborhood defined by study participants as the area within which they normally find sexual partners may be conceptually more relevant than an administrative neighborhood such as a census track or township.
In Table 2, the authors report male circumcision as a binary variable i.e. >8% and <8%, justifying the cut-off because the levels were “very low.” But variables that may have low prevalence values may still be categorized differently than binary. Even if the prevalence of circumcision was higher, the variable could still be dichotomized. What should matter is the underlying assumption as the role of this variable within the broader conceptual framework of the study. As the authors only used <8% and >8%; in which category was 8%, which is not reported? It would appear reasonable to dichotomize this variable at the median prevalence of circumcision (for the 16 “neighborhoods”) rather than at an arbitrary value.
The paper’s terminology is sometimes imprecise and thus has the potential to limit the reader’s understanding. For instance, neighborhoods are categorized as having “a market nearby”, electricity or water uniformly distributed and being located near a health center. Was universal access to water in a neighborhood required for it to be described as having water (versus not)? If not, what were the cut-off values? How near is near? Is nearness (by distance) to market similar to nearness to health center? Whose assessment matters - the study participants’, the interviewer’s or measurements taken and then categorized at analysis? These are important considerations because if the focus is on context, we should be able to hypothesize which risk behaviours an individual is exposed to because of his or her nearness to a community facility.
Finally, it appears the authors have used aggregate measures to highlight compositional aspects of the “neighborhoods” rather than context (Diez Roux 2001). In the Discussion section, there are limited suggestions as to how “contextual” factors such as water and electricity availability, closeness to health center or market are related to being HIV-infected. How would being close to a health center affect the risk of infection? Does residence in a poor neighborhood affects one’s behaviours? Attention to these concerns would have shed more light on the mediating factors that readers need to consider in a neighborhood study.
- 2001) Investigating neighborhood and area effects on health. American Journal of Public Health 91, 1783–1789. (
- 2008) Study Group on Heterogeneity of HIV Epidemics in African Cities The role of context: Neighbourhood characteristics strongly influence HIV risk in young women in Ndola, Zambia. Tropical Medicine and International Health 13, 162–170. , & (