Looking upstream to prevent HIV transmission: can interventions with sex workers alter the course of HIV epidemics in Africa as they did in Asia?

Steen, Richarda,b; Hontelez, Jan A.C.a,d; Veraart, Andrac; White, Richard G.e; de Vlas, Sake J.a

doi: 10.1097/QAD.0000000000000176
Epidemiology and Social

Background: High rates of partner change in ‘upstream’ sex work networks have long been recognized to drive ‘downstream’ transmission of sexually transmitted infections (STIs). We used a stochastic microsimulation model (STDSIM) to explore such transmission dynamics in a generalized African HIV epidemic.

Methods: We refined the quantification of sex work in Kisumu, Kenya, from the 4-cities study. Interventions with sex workers were introduced in 2000 and epidemics projected to 2020. We estimated the contribution of sex work to transmission, and modelled standard condom and STI interventions for three groups of sex workers at feasible rates of use and coverage.

Results: Removing transmission from sex work altogether would have resulted in 66% lower HIV incidence (range 54–75%) and 56% lower prevalence (range 44–63%) after 20 years. More feasible interventions reduced HIV prevalence from one-fifth to one-half. High rates of condom use in sex work had the greatest effect, whereas STI treatment contributed to HIV declines at lower levels of condom use. Interventions reaching the 40% of sex workers with most clients reduced HIV transmission nearly as much as less targeted approaches attempting to reach all sex workers. Declines were independent of antiretroviral therapy rollout and robust to realistic changes in parameter values.

Conclusion: ‘Upstream’ transmission in sex work remains important in advanced African HIV epidemics even in the context of antiretroviral therapy. As in concentrated Asian epidemics, feasible condom and STI interventions that reach the most active sex workers can markedly reduce the size of HIV epidemics. Interventions targeting ‘transactional’ sex with fewer clients have less impact.

Author Information

aDepartment of Public Health, Erasmus MC, University Medical Center Rotterdam, Rotterdam, The Netherlands

bCentre for Health Policy, Faculty of Health Sciences, School of Public Health, University of Witwatersrand, South Africa

cCreating 010, Rotterdam University of Applied Sciences, Rotterdam

dNijmegen International Center for Health System Analysis and Education, Department of Primary and Community Care, Radboud University Nijmegen Medical Centre, The Netherlands

eLondon School of Hygiene and Tropical Medicine, London, United Kingdom.

Correspondence to Richard Steen, PA, MPH, Erasmus MC, University Medical Center Rotterdam, Rotterdam, The Netherlands. Tel: +39 333 838 8534; e-mail: steenr7@gmail.com.

Received 4 August, 2013

Revised 5 December, 2013

Accepted 5 December, 2013

Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal's Website (http://www.AIDSonline.com).

Article Outline
Back to Top | Article Outline


The concept of ‘core transmission’ is fundamental to the epidemiology of sexually transmitted infections (STIs) [1]. The basic reproductive number (R0) that determines whether an STI will spread or die out within a population depends on rates of sexual partner change [2]. As sexual behaviour is highly heterogeneous, however, with most people having few sexual partners, high rates of partner change in sub-populations are required to sustain transmission. A high reproductive number within active sub-populations can serve as an engine for epidemic growth, whereas wider sexual mixing – ‘bridging’ between high and lower-risk populations – facilitates dissemination to others at lower risk.

HIV prevalence rates among sex workers and clients are generally several times higher than among lower-risk women and men [3]. Higher prevalence, more frequent contact with uninfected, susceptible individuals and higher transmission efficiency – due to high prevalence of ulcerative and other STI cofactors – potentially raise HIV incidence in core-bridge networks orders of magnitude above that in the general population. Such a large incidence differential is analogous to a transmission pump that can sustain overall prevalence at high levels. The implication for prevention, perhaps as relevant for HIV today as it was for cholera in nineteenth century London, is that epidemics may be controlled by intervening effectively at the level of the pump [4].

Empirical evidence from Thailand, Cambodia and elsewhere in Asia – where interrupting core transmission resulted in rapid deflation of HIV epidemics – supports this approach [5–7]. Common components of such interventions include peer-based outreach to attain high coverage, condom programming and STI services that address both symptomatic and asymptomatic infections. These have also been successful in parts of Africa – Kinshasa, Nairobi, Abidjan, South African mining communities – although implemented on a more limited geographic scale [8–13]. Programmes that address structural conditions of sex work and offer more comprehensive HIV and reproductive health services may be even more effective [14–16].

Yet, exception has been made for highly generalized epidemics in sub-Saharan Africa (SSA). Some argue that since most infections now take place between couples or regular partners, the fraction attributable to sex work must be low [17]. The concept of ‘self-sustaining’ generalized epidemics – initially ignited by, but now largely independent of sex work – has taken hold, leading to a wide diffusion of prevention resources in an effort to cover large populations with a range of interventions [18].

Mathematical modelling has been used to estimate the effect of individual interventions – from ‘behaviour change’ to STI treatment, male circumcision and antiretroviral therapy (ART) – on HIV epidemic spread in SSA [19–22]. Implicit in many models is the assumption of reaching high levels of coverage in the general population. Less attention has been paid to modelling the effect of targeting itself – focusing a few proven prevention interventions on small sub-populations at high risk of HIV acquisition and transmission – as a prevention strategy [13,23,24].

The study uses an established microsimulation model (STDSIM) to explore the effect that targeting sex workers could have on HIV epidemics in SSA, using an updated and refined version of an existing model quantification for Kisumu, Kenya [25–27]. Condom use and STI treatment interventions are simulated at varying levels of coverage among sex worker populations, assuming no change in interventions or behaviour among the general population. Our aim is to explore transmission dynamics and related intervention effects in a generalized epidemic setting, not to make specific predictions for Kisumu. To do this, we take a counterfactual approach – simulating introduction of a range of interventions with sex workers in the year 2000, where the model fit is close to data points from the 4-cities study – and estimate the impact on HIV epidemics 10 and 20 years later. This approach also permits estimation of changes in HIV incidence and prevalence separately from ART, which is a more recent factor with independent effects on both [22]. Programmatic implications, including potential synergies of implementing targeted interventions together with ART rollout, are also examined.

Back to Top | Article Outline


Model description

STDSIM, a stochastic individual-based microsimulation model, has been used extensively to study HIV and STI epidemics in SSA [28–32]. STDSIM simulates the natural history and transmission of HIV and other STIs in a population consisting of individuals with characteristics that can change over time. The formation and dissolution of heterosexual relationships and transmission of STIs during contacts between sexual partners are modelled as stochastic events. STDSIM allows the simultaneous and interactive simulation of several different STIs, with different HIV cofactor effects, and against which different prevention and treatment interventions can be applied.

An STDSIM quantification was previously fit to demographic, behavioural and epidemiological data for Kisumu as part of the 4-cities study [23–27]. We used the same quantification with four improvements. First, we updated HIV transmission probabilities used in 4-cities modelling to lower values that resulted from a recent systematic review (Table S1 in supplementary materials, http://links.lww.com/QAD/A462) [33]. Secondly, in the quantification of sex work, we used the mean number of clients per week (1.6) and centile ranges from survey data, rather than the median (1.0) as reported by Morison et al.[27] and used by Orroth et al.[25]. Also, assumptions about the coverage and effectiveness of STI treatment were revised downward based on recent programme evaluation data [34]. Finally, we refined the simulation of commercial sex to include three categories of sex workers with different rates of partner change (numbers of clients), as explained in the following paragraph. More details on the above changes are provided in the supplementary materials (http://links.lww.com/QAD/A462).

Data suggest that approximately 3% of women aged 15–49 in Kisumu engage in some form of sex work [35]. Behavioural data derive from population-based surveys using standardized questionnaires with a broad definition of sex work – any exchange of sex for money, gifts or favours in the past year. Yet, sex workers in the 4-cities study reported widely divergent mean numbers of clients in the past week – 1.6 in Kisumu, 3.3 in Yaounde, 4.0 in Ndola and 11.9 in Cotonou. We explored heterogeneity of client numbers by disaggregating into three sub-populations of sex workers, maintaining an overall sex worker population size of 2–3% and a mean of 1.6 clients per week. On the basis of the observed geometric distribution of the survey data, we defined a low activity group (L = 60% of all sex workers) with 0.6 clients, a medium group (M = 30%) with 2.2 clients and a high group (H = 10%) with 5.7 clients per week. In programmatic terms, these groups may approximate commonly observed patterns of sex work, from overt, full-time, self-identified ‘professional’ (H) and their somewhat less active peers (M), to covert, occasional, non-identifying or ‘transactional’ (L). We aggregate H and M groups in some analyses to compare the effect of targeting the 40% of more highly active sex workers (HM) to that of reaching all sex workers (HML).

Client demand – the proportion of men who buy sex and their number of sex worker contacts per year – was calculated to arrive at the sex worker activity levels described above. Client population sizes were maintained from the 4-cities quantification (31% of men 15–49), as was their distribution among married and unmarried men. We increased the frequency of visiting sex workers from 12 to 18 times per year for the more active group (5% of men) and from one to three times per year for a less active group (26% of men) after examining alternative assumptions (see Table S3 in supplementary materials, http://links.lww.com/QAD/A462). HIV is introduced first as outside infections (at an ongoing rate of 1.9% for male and 1.7% for female immigrants) in 1980, and in four sex workers in 1984.

Back to Top | Article Outline
Simulated interventions

Simulations were run to explore the contribution of sex work to HIV and STI transmission and the potential effect of targeting sex workers with different condom and STI treatment interventions. First, the hypothetical independence of HIV epidemics from sex work was explored by completely removing the possibility of transmission through sex work abruptly in 2000. To this end, sex work continues in the model but we assume 100% condom protection without failure. Although unrealistic, these assumptions permit exploring the plausibility of epidemics persisting where sex work exists but has no role in transmission – the null hypothesis behind sex work as a continuing transmission factor. Simulations examine the effect of removing transmission potential for all sex workers and for subsets of highly active sex workers. The population-attributable fraction (PAF) of sex work in HIV transmission in these settings was estimated for the years 2010 and 2020 using the formula 1 −  (HIVincidence without sex work/HIVincidence with sex work).

Second, after restoring transmission potential to commercial sex, we introduced standard, feasible interventions at different levels of utilization and coverage among sex workers assuming no change in other interventions or behaviour among the general population. These include the following:

* Targeted condom programmes, which have been shown to raise reported condom use in commercial sex to 80–90% or more [7,14,15].

* Targeted STI treatment, which includes both sensitive STI screening and periodic presumptive treatment (PPT), or either of them, covering both symptomatic and asymptomatic STIs. Coverage rates of 60% or more have been described for such interventions [36,37].

For each scenario, we modelled interventions – first condom use alone, then condom use with STI treatment – at different levels of coverage, and determined HIV incidence and prevalence in the general population 10 and 20 years after their introduction. As before, we estimate the effect of feasible interventions applied to different sex worker sub-populations.

After making the above changes, STI transmission probabilities were refitted within a priori ranges as was done in the original 4-cities model (Table S2 in supplemental materials, http://links.lww.com/QAD/A462) in order to arrive at similar STI prevalences to those observed. STI co-factor estimates for HIV transmission and acquisition were not changed.

Simulation results were based on the average of 400 runs for each scenario to minimize stochastic variability [25]. The epidemiology of HIV transmission at population level was examined by linking HIV transmission events over time from sex workers and clients to subsequent ‘downstream’ infections in other partners.

Alternative explanations for findings were explored in scenario analysis varying key parameters including sex worker and client population sizes and partner change rates, multiplying each parameter value by 3/2 and 2/3, respectively (Table S3 in supplemental materials, http://links.lww.com/QAD/A462). A scenario that includes assumptions about ART rollout based on Kenyan programme data was run to explore whether this influenced the main findings. ART coverage until 2010 was fit to coverage levels reported by WHO using two sub-models representing an individual's demand for ART and the capacity of the health system to meet this demand (section S3.2 in supplemental materials, http://links.lww.com/QAD/A462) [22].

Back to Top | Article Outline


The baseline model quantification (Fig. 1) describes simulated HIV prevalence trends from 1980 to 2020 compared to data points from antenatal clinic surveys.

Figure 2 shows that complete blocking of transmission in sex work in 2000 would have resulted in 66% lower HIV incidence (54–75% lower in 95% of runs) and 56% lower prevalence (44–63% lower in 95% of runs) by 2020 compared to the same scenario without intervention. Overall HIV incidence declines from 3.6% in 2000 to 1.8% in 2010 and 1.1% in 2020; HIV prevalence falls from 24.9% in 2000 to 16.0% in 2010 and 9.4% in 2020 (continuing to below 5% by 2033 and 2.1% in 2050). The estimated PAF for all sex work (HML) is 45.6% in 2010 (PAF2010) and 66.4% in 2020 (PAF2020). HIV incidence reductions are almost identical for the HM sub-group (PAF2010 = 43.7%, PAF2020 = 64.4%), suggesting that most transmission is attributable to sex work with high numbers of clients. In contrast, blocking transmission in the L sub-group with lowest client numbers has only a marginal effect on the epidemic.

The response to more realistic interventions in sex work is shown in Fig. 3. Increases in condom use and STI screening resulted in reduced transmission and a progressive decline in HIV prevalence in all scenarios. At 70–85% condom use, HIV prevalence is reduced by 25–29% after 10 years and 46–52% after 20 years. Again, intervening in high (H) and medium (M) activity sex work accounts for nearly all of the estimated HIV declines.

The addition of STI treatment interventions (including screening for asymptomatic infections or PPT) adds little benefit in terms of HIV reductions when rates of condom use are high (70–85%). However, at lower rates (40–55%), adding STI treatment is roughly equivalent to increasing condom use 15 percentage points – 40% condom use with STI treatment comparable to 55% condom use alone.

Finally, we ran all simulations in a quantification that included progressive rollout of ART as has been described for Kenya (Fig. 4) [34]. ART scale-up started in 2003; by 2010, 50% of those eligible (at CD4+ cell counts of ≤350 cells/μl) were on treatment, with universal access (coverage of >80%) achieved in 2014 and maintained thereafter.

Compared to ART alone, HIV prevalence in 2020 was 23, 42 and 46% lower when condom use among more highly active sex workers (HM) increased to 55, 70 and 85% in 2000, and HIV incidence declined 33, 53 and 63% more than in the ART-only scenario. The complementary of the two interventions is also apparent – with ART, the effect on HIV prevalence of raising condom use to 85% (HM: 8.4% by 2020) is greater than that of completely blocking transmission without ART (HM: 10.0%; Fig. 2b).

Back to Top | Article Outline


The study explores the role of sex work and the potential impact of targeted prevention interventions in a generalized African HIV epidemic. The simulation of sex work in the 4-cities model was refined based on available data without changing assumptions about general population sexual behaviour. Our baseline parameters and assumptions thus reflect early conditions and interventions in Kisumu, which have undoubtedly changed since 2000. The estimates from this study were used to generate insights into transmission dynamics and potential interventions in a generic generalized epidemic setting, not to describe HIV and STI trends in Kisumu today.

Our findings suggest that sex work – in the absence of effective interventions – remains an important driver of HIV transmission, even in generalized or ‘hyper-endemic’ epidemics such as Kisumu's. By analysing transmission chains (HIV transmission events), we estimate that approximately 60% of HIV infections can be linked to sex work in 2000, 20 years after the start of the epidemic, subsequently levelling off to nearly 50% in 2010 and beyond. Without this continued infusion of new infections from sex work after 2000, the model estimates that HIV prevalence in the general population would be 30% lower in 2010 and 56% lower in 2020.

As in Asia – where preventing transmission in sex work has already reversed large HIV epidemics in several countries – interrupting ‘upstream’ transmission would have obvious benefits. These include facilitating ART scale-up and universal access to services by reducing the number of people who need them, as well as raising the possibility of eliminating new HIV infections altogether. We estimate that HIV incidence in the modelled scenario could be reduced to 0.1% (and prevalence below 2%) by 2040 by increasing condom use rates in sex work, beginning in 2013, to levels comparable to those reported from Asia (85%), while continuing to scale up ART (explorations beyond period shown in Figure S2, http://links.lww.com/QAD/A462). Given the magnitude of such potential benefits, further research into their relevance to African settings should be a high priority.

We also examined the effect of interventions targeted to different groups of sex workers. Most behavioural surveys define sex work broadly as the exchange of sex for material gain, in cash or in kind, yet often report that many respondents who meet this definition neither self-identify as sex workers nor have many partners. Such surveys, in casting a wide net, tend to ‘dilute’ the core group with large numbers of ‘sex workers’ with relatively low rates of partner change. We disaggregated sex workers by reported client numbers to analyse the effect of targeting different groups. Although interventions have a large impact when applied to all sex workers, nearly all of this effect remains if only the 40% of sex workers with highest rates of partner change are reached. From a programme perspective, this would be similar to targeting only the most overt sex workers – who are often easiest to identify – without reaching the more numerous and hard-to-reach ‘transactional’ or ‘occasional’ sex workers. Blocking transmission from these sex workers – who account for a disproportionate number (79%) of client contacts and most sex work-linked transmission events (91%) – completely would be expected to reduce general population HIV prevalence by 28% at 10 years to 53% after 20 years, with overall incidence reductions of 44 and 64%, respectively.

It is, of course, unrealistic to assume that transmission in sex work, even among smaller sub-populations, can be blocked completely. We therefore modelled several basic interventions among sex workers that have been shown to be feasible and effective. Condom use as expected is the most effective way to interrupt transmission in commercial sex networks. Empirical data from Kinshasa, Nairobi, Abidjan, Cotonou, Thailand, Cambodia, India and elsewhere associate condom use rates of 80–90% reported by sex workers with declining HIV rates [8–11,14–16]. Actual use may be lower than reported, however. Simulations at different levels of condom use (85, 70 and 55%) in our model showed large impact on transmission, even when only highly active sex workers were targeted. Again, from a programme perspective, this is an important finding as the feasibility of increasing coverage and condom use among a subset of easily identifiable sex workers is much greater than trying to reach all sex workers.

Sexually transmitted infection screening and treatment interventions were simulated as an adjunct to condom programmes, assuming effective treatment of both symptomatic and asymptomatic infections, with 60% sex worker coverage every 3 months. Such interventions can be implemented either as laboratory-based screening (with sensitive diagnostics) or PPT, feasible interventions with well documented outcomes [3,12]. The simulated STI intervention in the model was found to be highly synergistic with condom use – raising the impact of 40% condom use to the equivalent of 55% condom use alone, if 60% of sex workers also received STI screening or presumptive treatment. As expected, the additional benefit of STI treatment diminishes as rates of condom use in sex work increase reducing STI/HIV exposures.

We also explored alternative explanations for the findings by varying several key assumptions (supplementary materials, http://links.lww.com/QAD/A462). Sex worker partner change rates reinforce findings, with higher rates contributing more to transmission. Findings are also resilient to varying client population sizes and contact rates. In addition, the analysis including ART rollout using Kenyan data (Fig. 4) showed an additive, complementary effect on HIV incidence and prevalence. This is an important area for further research.

In conclusion, our findings suggest that, similar to concentrated epidemics in Asia, generalized African HIV epidemics may not be self-sustaining in the absence of high-incidence ‘upstream’ transmission from sex work. Moreover, feasible interventions that focus on small but highly active sex work networks may be highly efficient in controlling transmission in such epidemics. Basic interventions that address condom use and curable STIs in sex work appear to be core elements of an effective prevention response even in contexts where ART is being rolled out. Intervention-linked research should be prioritized to confirm these findings in other generalized epidemic settings.

Back to Top | Article Outline


Author contributions: R.S. contributed to the study design and is responsible for conducting the primary data analyses and drafting the article.

J.A.C.H. developed the ART module, conducted the ART analyses, reviewed the entire manuscript and contributed substantially to revisions.

A.V. conducted an analysis of HIV transmission events, reviewed the entire manuscript and contributed to revisions.

R.G.W. contributed substantially to the conception and design of the study, data analysis, manuscript review and revision.

S.J.deV. contributed to the conception and design of the study, provided technical guidance in data analysis, reviewed the manuscript and contributed substantially to revisions.

Author access to data: R.S. and S.J.deV. had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Back to Top | Article Outline
Conflicts of interest

The authors state no conflict of interest.

Back to Top | Article Outline


1. Brunham RC, Plummer FA. A general model of sexually transmitted disease epidemiology and its implications for control. Med Clin North Am 1990; 74:1339–1352.
2. Bremermann HJ, Anderson RW. Mathematical models of HIV infection. I. Threshold conditions for transmission and host survival. J Acquir Immune Defic Syndr 1990; 3:1129–1134.
3. Hank Pizer, Vuylsteke B, Das A, Dallabetta G, Laga M. Kenneth H, Mayer. Preventing HIV among sex workers. HIV prevention: a comprehensive approach 2008.
4. Snow J. On the mode of communication of cholera. London:Wilson and Ogilvy; 1849.
5. Ruxrungtham K, Brown T, Phanuphak P. HIV/AIDS in Asia. Lancet 2004; 364:69–82.
6. Phoolcharoen W. HIV/AIDS prevention in Thailand: success and challenges. Science 1998; 280:1873–1874.
7. Ng M, Gakidou E, Levin-Rector A, Khera A, Murray CJL, Dandona L. Assessment of population-level effect of Avahan, an HIV-prevention initiative in India. Lancet 2011; 378:1643–1652. doi: 10.1016/S0140-6736(11)61390-1
8. Laga M, Alary M, Nzila N, Manoka AT, Tuliza M, Behets F, et al. Condom promotion, sexually transmitted disease treatment, and declining incidence of HIV-1 infection in female Zairean sex workers. Lancet 1994; 344:246–248.
9. Ngugi EN, Chakkalackal M, Sharma A, Bukusi E, Njoroge B, Kimani J, et al. Sustained changes in sexual behavior by female sex workers after completion of a randomized HIV prevention trial. J Acquir Immune Defic Syndr 2007; 45:588–594.
10. Ghys PD, Diallo MO, Ettiegne-Traore V, Bukusi E, Njoroge B, Kimani J, et al. Increase in condom use and decline in HIV and sexually transmitted diseases among female sex workers in Abidjan, Cote d’Ivoire, 1991–1998. AIDS 2002; 16:251–258.
11. Alary M, Mukenge-Tshibaka L, Bernier F, Geraldo N, Lowndes CM, Meda H, et al. Decline in the prevalence of HIV and sexually transmitted diseases among female sex workers in Cotonou, Benin, 1993–1999. AIDS 2002; 16:463–470.
12. Steen R, Chersich M, Gerbase A, Neilsen G, Wendland A, Ndowa F, et al. Periodic presumptive treatment of curable STIs among sex workers: a systematic review. AIDS 2012; 26:437–445.
13. Vickerman P, Ndowa F, O’Farrell N, Steen R, Alary M, Delany-Moretlwe S. Using mathematical modelling to estimate the impact of periodic presumptive treatment on the transmission of sexually transmitted infections and HIV amongst female sex workers. Sex Transm Infect 2010; 86:163–168.
14. Jana S, Basu I, Rotheram-Borus MJ, Newman PA. The Sonagachi Project: a sustainable community intervention program. AIDS Educ Prev 2004; 16:405–414.
15. Rojanapithayakorn W. The 100% condom use programme in Asia. Reprod Health Matters 2006; 14:41–52.
16. Sopheab H, Fylkesnes K, Lim Y, Godwin P. Community action for preventing HIV in Cambodia: evaluation of a 3-year project. Health Policy Plan 2008; 23:277–287.
18. Potts M, Halperin DT, Kirby D. Reassessing HIV prevention. Science 2008; 320:749–750.
19. Korenromp EL, Van Vliet C, Bakker R, De Vlas SJ, Habbema JDF. HIV spread and partnership reduction for different patterns of sexual behaviour: a study with the microsimulation model STDSIM. Math Popul Studies 2000; 8:135–173.
20. Orroth KK, White RG, Korenromp EL, Bakker R, Changalucha J, Habbema JD, Hayes RJ. Empirical observations underestimate the proportion of human immunodeficiency virus infections attributable to sexually transmitted diseases in the Mwanza and Rakai sexually transmitted disease treatment trials: simulation results. Sex Transm Dis 2006; 33:536–544.
21. White RG, Glynna JR, Orroth KK, Freeman EE, Bakker R, Weiss HA, et al. Male circumcision for HIV prevention in sub-Saharan Africa: who, what and when?. AIDS 2008; 22:1841–1850.
22. Hontelez JA, Lurie MN, Bärnighausen T, Bakker R, Baltussen R, Tanser FET-AL>. Elimination of HIV in South Africa through expanded access to antiretroviral therapy: a model comparison study. PLoS 2013; 10:e1001534.
23. Boily M-C, Pickles M, Vickerman P, Buzdugan R, Isac S, Deering KN, et al. Using mathematical modelling to investigate the plausibility of attributing observed antenatal clinic declines to a female sex worker intervention in Karnataka state, India. AIDS 2008; 22 (Suppl 5):149–164.
24. Mishra S, Steen R, Gerbase A, Lo Y-R, Boily M-C. Impact of high-risk sex and focused interventions in heterosexual HIV epidemics: a systematic review of mathematical models. PLoS One 2012; 7:e50691.
25. Orroth KK, Freeman EE, Bakker R, Buve A, Glynn JR, Boily MC, et al. Understanding the differences between contrasting HIV epidemics in east and west Africa: results from a simulation model of the Four Cities Study. Sex Transm Infect 2007; 83 (Suppl 1):i5–i16.
26. Buvé A, Caraël M, Hayes RJ, Auvert B, Ferry B, Robinson NJ, et al. Multicentre study on factors determining differences in rate of spread of HIV in sub-Saharan Africa: methods and prevalence of HIV infection. AIDS 2001; 15 (Suppl 4):S5–S14.
27. Morison L, Weiss HA, Buvé A, Carael M, Abega SC, Kaona F, et al. Commercial sex and the spread of HIV in four cities in sub-Saharan Africa. AIDS 2001; 15 (Suppl 4):S61–S69.
28. Korenromp EL, White RG, Orroth KK, Bakker R, Kamali A, Serwadda D, et al. Determinants of the impact of sexually transmitted infection treatment on prevention of HIV infection: a synthesis of evidence from the Mwanza, Rakai, and Masaka Intervention Trials. J Infect Dis 2005; 191:S168–S178.
29. White RG, Orroth KK, Glynn JR, Freeman EE, Bakker R, Habbema JD, et al. Treating curable sexually transmitted infections to prevent HIV in Africa: still an effective control strategy?. J Acquir Immune Defic Syndr 2008; 47:346–353.
30. White RG, Freeman EE, Orroth KK, Bakker R, Weiss H, O’Farrell N, et al. Population-level effect of HSV-2 therapy on HIV incidence in sub-Saharan Africa. Sex Transm Infect 2008; 84 (Suppl 2):ii12–ii18.
31. Hontelez JA, de Vlas SJ, Baltussen R, Newell ML, Bakker R, Tanser F, et al. The impact of antiretroviral treatment on the age composition of the HIV epidemic in sub-Saharan Africa. AIDS 2012; 26 (Suppl 1):S19–S30.
32. Vissers DC, De Vlas SJ, Bakker R, Urassa M, Voeten HA, Habbema JD. The impact of mobility on HIV control: a modelling study. Epidemiol Infect 2011; 139:1845–1853.
33. Boily MC, Baggaley RF, Wang L, Masse B, White RG, Hayes RJ, et al. Heterosexual risk of HIV-1 infection per sexual act: systematic review and meta-analysis of observational studies. Lancet Infect Dis 2009; 9:118–129.
34. Cheluget B, Joesoef MR, Marum LH, Wandera C, Ryan CA, Decock KM, Chebet KL. Changing patterns in sexually transmitted disease syndromes in Kenya after the introduction of a syndromic management program. Sex Transm Dis 2004; 31:522–525.
35. Vandepitte J, Lyerla R, Dallabetta G, Crabbe F, Alary M, Buve A. Estimates of the number of female sex workers in different regions of the world. Sex Transm Infect 2006; 82 (Suppl 3):iii18–iii25.
36. Steen R, Chersich M, de Vlas SJ. Periodic presumptive treatment of curable sexually transmitted infections among sex workers: recent experience with implementation. Curr Opin Infect Dis 2012; 25:100–106.
37. Steen R, Mogasale V, Wi T, Singh AK, Das A, Daly C, et al. Pursuing scale and quality in STI interventions with sex workers: initial results from Avahan India AIDS Initiative. Sex Transm Infect 2006; 82:381–385.

generalized HIV epidemic; HIV; modelling; sex work

Supplemental Digital Content

Back to Top | Article Outline
© 2014 Lippincott Williams & Wilkins, Inc.