Hargreaves JR, Bonell CP, Boler T, Boccia D, Birdthistle I, Fletcher A, et al. Systematic review exploring time trends in the association between educational attainment and risk of HIV infection in sub-Saharan Africa. AIDS 2008 Jan 30;22(3):403-14.
To assess the evidence that the association between educational attainment and risk of HIV infection has changed over time in sub-Saharan Africa
Systematic review of published, peer-reviewed articles. The search strategy included a comprehensive search of the major biomedical databases (e.g., PubMed, EMBASE), hand searching key journals (e.g., AIDS, International Journal of STD & AIDS, JAIDS) and checking reference lists of relevant articles. Articles were identified that reported original data comparing individually measured educational attainment and HIV status among at least 300 individuals representative of the general population of countries or regions of sub-Saharan Africa. Studies restricted to high-risk and other sub-sections of the population were excluded. Articles not adjusting for setting, age, and sex or inappropriately adjusting for sexual behavior or STDs on the causal pathway to HIV infection were also excluded. A search term for "sexual behaviour" was included. The same search strategy, not restricted to sub-Saharan Africa nor including sexual behaviour, was used by the authors in a review of this question in 2001.(1) A formal meta-analysis was not done because the association was expected to differ across differing populations.
Approximately 3,800 abstracts and 1,200 full papers were reviewed. Thirty-six articles were included in the study (18 with data through 1996 were reported in the 2001 publication), that contained data on 72 discrete populations from 11 countries between 1987 and 2003, representing over 200,000 individuals, 32 populations from data prior to 1996 and 40 from 1996 onwards. Thirty-four were cross-sectional studies from general population surveys, samples of young people, army recruits, residents of a sugar estate, and antenatal clinic attendees. One factory-based cohort study and one case-control study nested within a cluster-randomized trial were included. Situations were identified in which successive cross-sectional studies covered the same population with the same methodology and analysis at different points in time.
Overall, more than 200,000 participants were included from all studies combined. Twenty-five of the 72 discrete populations were male, 35 were female, and 12 were mixed sex. Twenty-six were urban settings, 41 were rural, and 5 were mixed. Fifty-four were from general population surveys.
Analyses compared levels of HIV infection among individuals grouped by two to five categories of educational attainment, measured as years of school attended or grade achieved. The primary outcome was the odds ratio and 95% confidence interval for HIV infection between the most educated and the least educated within each population, as most studies did not assess linearity or trend by education level. In each case, the result was classified as a statistically significant (p<0.05) increased risk of HIV, a significantly lower risk, or no association. The year 1996 was chosen a priori as a cut-off to describe changes over time because most of the articles included in the authors' previous review presented data from before 1996 and because about half the data were collected before and after this date. Trends over time in the odds ratio were reported for those studies with repeat data on the same population.
During each wave, participants received HIV testing, along with pre- and post-test counseling, to determine HIV prevalence and to estimate HIV incidence in the community. A structured questionnaire was administered by trained interviewers to determine key demographics and injecting behaviors. Aside from the variations in the recruitment strategies, data collection methods were the same for all sites.
Comparing the highest and lowest educational categories, increased educational attainment was not statistically significantly associated with risk of HIV infection in 44 populations, was associated with an increased risk of HIV infection in 20 populations and was associated with a decreased risk of infection in 8 populations. Populations with data collected prior to 1996 were more likely to show a higher risk of HIV infection among the most educated (15/32 populations) than those with data from 1996 onwards (5/40). Only one pre-1996 population, compared with 7/40 from 1996 onwards, had a lower risk of HIV infection among the most educated.
Eligible data from different time periods were available for 13 populations from five countries: Malawi (1 population), Tanzania (2), Uganda (3), Zambia (6) and Zimbabwe (1). Downward trends in HIV prevalence in the most educated, together with a reduction in the relative HIV prevalence between the most and least educated were seen in rural Kagera, Tanzania; women from Masaka, Uganda; antenatal clinic users in Fort Portal, Uganda, and in Manicaland, Zimbabwe; and urban populations in Zambia. In general, from 1996 onwards, changes in the adjusted odds ratio for HIV infection comparing groups of the highest and lowest levels of educational attainment tended to shift towards no association or to a lower risk in the most educated.
HIV infections appear to be shifting towards higher prevalence among the least educated in sub-Saharan Africa, in a reverse of previous patterns. Studies on data collected prior to 1996 generally found either no association or the highest risk of HIV infection among the most educated. Studies conducted from 1996 onwards were more likely to find a lower risk of HIV infections among the most educated. Where data over time were available, HIV prevalence fell more consistently among highly educated groups than among less-educated groups, in whom HIV prevalence sometimes rose while overall population prevalence was falling. In several populations, associations suggesting greater HIV risk in the more educated at earlier time points were replaced by weaker associations later. The authors suggest that new HIV infections occurring in the late 1990s and into the 2000s have been occurring disproportionately among the least educated members of society in sub-Saharan Africa.
Using the QUOROM scale for evaluation of systematic reviews, this study is of good quality. The objectives are clearly described and information on the data sources is detailed in a table in the report. The authors conducted a comprehensive search for published articles. The search strategy, inclusion/exclusion criteria, and data abstraction procedures are delineated. A discussion of the methodological quality of included studies, however, was not well described. Outcome measures were detailed, however, a standardized measure of exposure (educational attainment) was not available and a variety of scales were used so that it was only possible to identify associations between risk of HIV infection and relative level of education in each study. Thus, the distinction between most educated and least educated groups varied between studies. Repeated studies over time in the same population were available in only 13 of the 72 populations. Statistical measures of the significance of time trends are not included
HIV has widely been characterized as a disease of poverty; however, the association between socioeconomic status and risk of HIV infection is much more complex. The authors previously published a systematic review on the association between educational attainment and risk of HIV infection in developing countries.(1) All but one of the studies from sub-Saharan Africa included in this review found either no association between educational attainment and HIV infection or that individuals with the highest levels of education were more likely to be infected. The authors hypothesized, along with others, that as the public health response to HIV grew in sub-Saharan Africa, the most educated, empowered members of society would be the first to adopt protective behaviors.(1,2) They further hypothesized that in the absence of effective interventions that engaged the least educated, HIV would increasingly affect this group over time. The hypothesis that change occurs earlier in those with more education is also supported by well-established theory, such as the diffusion of innovations model.(3) The present study serves as an update to the previous systematic review and attempts to provide data to evaluate these hypotheses.
The authors suggest that the results of the review provide evidence that the epidemiology of HIV infection in sub-Saharan Africa may be changing. This change highlights the importance of monitoring future trends within surveillance systems across sub-Saharan Africa. Policy responses that ensure HIV-prevention measures reach all strata of society and increased education levels are urgently needed.
- Hargreaves JR, Glynn JR. Educational attainment and HIV-1 infection in developing countries: a systematic review. Trop Med Int Health 2002;7(6):489-98.
- Over M, Piot P. HIV infection and sexually transmitted diseases. In: Disease Control Priorities in Developing Countries. Edited by Jamison DT, Mosley WH, Mensham AR, Bobadilla JL. Oxford: Oxford University Press; 1993. Pp. 455-527. No abstract available.
- Rogers EM. Diffusion of Innovations. New York: The Free Press; 1962. No abstract available.