Search tips
Search criteria 


Logo of wtpaEurope PMCEurope PMC Funders GroupSubmit a Manuscript
J Acquir Immune Defic Syndr. Author manuscript; available in PMC 2013 September 19.
Published in final edited form as:
PMCID: PMC3776949

Treating curable sexually transmitted infections to prevent HIV in Africa: Still an effective control strategy?



Evidence regarding the effectiveness of sexually transmitted infection(STI) treatment for HIV prevention in Africa is equivocal, leading some policy-makers to question whether it should continue to be promoted for HIV control. We explore whether treating curable STIs remains a cost-effective HIV control strategy in Africa.


The model STDSIM was fitted to the characteristics of four populations in East and West Africa. Over the simulated HIV epidemics, the population-attributable fractions(PAFs) of incident HIV attributable to STIs, the impact of syndromic STI management on HIV incidence and the cost per HIV-infection-averted were evaluated, and compared to an estimate of lifetime HIV treatment costs(US$3,500).


Throughout the HIV epidemics in all cities, the total PAF for all STIs remained high, with ≥50% of HIV transmission attributed to STIs. The PAF for HSV-2 increased during the epidemics, while the PAF for curable STIs and the relative impact of syndromic management decreased. However, the models showed that absolute impact of syndromic management remains high in generalized epidemics, and it remained cost-saving in three of the four populations where the cost per-HIV-infection-averted ranged between US$321-1,665.


Curable STI interventions may remain cost-saving in populations with generalised HIV epidemics, particularly populations with high-risk behaviours or low male circumcision rates.

Keywords: Sexually Transmitted Diseases, Epidemiology, Mathematical Model, Cost Effectiveness, Primary Prevention


HIV continues to spread throughout sub-Saharan Africa(referred to as Africa below)1. Interventions to reduce the transmission of HIV include behaviour change interventions, and interventions that reduce the per-sex-act transmission probability of HIV. Among the latter, prompt diagnosis and treatment of sexually transmitted infections(STIs) takes a prominent place, though evidence regarding the population-level effectiveness of STI treatment for HIV prevention in Africa is equivocal. A randomised controlled trial(RCT) of improved clinic-based syndromic STI treatment in rural Tanzania was shown to reduce HIV incidence in the general population by 38% and to be highly cost-effective2-4. However, since then four other RCTs of various STI treatment strategies including mass and syndromic treatment in the general population have failed to show an impact on HIV incidence5-8. We have shown in previous studies that this may be because behaviour change and the later stage of the HIV epidemic reduces the role of curable STIs in HIV transmission9, 10. This has led some policy-makers to question whether curable STI treatment should continue to be promoted as an HIV control strategy in generalized epidemics in Africa.

We address this question here by extending previous work investigating the heterogeneity of the HIV epidemics across Africa. An empirical study of two East African cities with high HIV prevalence(Ndola, Zambia and Kisumu, Kenya) and two West African cities with relatively low HIV prevalence(Cotonou, Benin and Yaounde, Cameroon)11 identified important differences between the populations12. The prevalence of male circumcision was lower and the prevalence of Herpes simplex virus type-2(HSV-2) infection was higher in the high HIV prevalence cities13, 14. Sexual risk behaviours were not found to be consistently more common in the high HIV prevalence cities12, 15, 16 suggesting that the effects of any differences in sexual behaviours were outweighed by differences in biological cofactors that influence HIV transmission such as male circumcision and HSV-2. This hypothesis was evaluated in a modelling study in which the characteristics of the four populations and the development of the HIV epidemics were simulated17. This showed that the different HIV prevalences could be explained largely by the lower rates of male circumcision in the East African cities compared to the West African cities.

We use these simulations of the development of the HIV epidemics in the four cities17 to estimate the population attributable fractions(PAFs) of STIs for HIV incidence(the proportion of new HIV infections in the population attributable to STI cofactor effects), the impact of treating curable STIs on HIV incidence, and how these effects vary over time.

We then explore the cost-effectiveness of improved syndromic STI case management in preventing HIV transmission. The study in Mwanza showing that this intervention is highly cost-effective4 took place in the 1990s in a population with an early HIV epidemic. To investigate how cost-effectiveness is likely to vary between different populations, and at different phases of the HIV epidemic, we compare the empirical cost-effectiveness estimate for the early HIV epidemic in Mwanza with the predicted cost-effectiveness of the same simulated intervention during the HIV epidemics in the four cities in West and East Africa.



The individual-based stochastic model STDSIM simulates the natural history and transmission of HIV and 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 events18. STDSIM has previously been used to explore the findings of the Mwanza, Rakai and Masaka STI treatment trials and the heterogeneity in the spread of HIV in Africa9, 17.

Default scenarios

Descriptions of the simulated populations used in this study have been published17, 19. In summary, STDSIM was fitted to the demographic, behavioural and epidemiological characteristics of the four cities for the year 1997 and to available data on trends over time17. HIV, HSV-2, chancroid, syphilis, gonorrhoea and chlamydia were simulated. It was assumed that STIs enhance both susceptibility to HIV infection and the infectivity of HIV-infected individuals. For chancroid, a per-contact HIV cofactor effect of 25 was assumed, towards the low end of the range(3 to 300) estimated for genital ulcers from studies among CSW and clients in Nairobi20, to account for possible residual confounding in these estimates21. The assumed per-contact cofactor effects of the other STIs reflected their relative clinical severity22, higher for primary HSV-2(25) than for recurrent HSV-2 ulcers(10) and infectious syphilis(7.5) and lower for the nonulcerative STIs, gonorrhoea and chlamydia(3). These per-contact cofactor magnitudes were larger than relative risks estimated from epidemiological studies22 because the latter reflect cumulative incidence over numerous sexual exposures over an extended follow-up period, during only some of which the STI will have been present. These cofactor magnitudes have been validated where possible by showing that models assuming these parameter values predict cumulative relative risks for HIV acquisition in line with those estimated from epidemiological studies23, and can simultaneously predict the large impact on HIV incidence measured in the syndromic STI treatment trial in Mwanza, Tanzania and the non-significant impacts of the syndromic STI treatment and mass STI treatment interventions tested in Masaka and Rakai, Uganda9, 10. Full details of the simulated STIs and HIV/STI prevalence trends are published elsewhere17, 19.

In line with the available data, previous simulations of the HIV epidemics in the four cities assumed improvements in STI treatment services and condom use over time17. To avoid underestimating the contribution of curable STIs to HIV incidence, we removed all simulated STI treatment services from these model quantifications, to create the default scenarios used in this study. The modelled increases in condom use over time were retained(condom use was simulated to increase gradually from 0% of sexual contacts prior to 1990, to 20-25% of casual contacts and 30-50% of sex-worker contacts by 1995, Table S1, online). Simulated HIV prevalence trends in the default scenarios and observed HIV prevalence trends in the four cities are shown in Figure 1.

Figure 1
Simulated and observed female HIV prevalence

Simulated population-attributable fractions and intervention impacts

To assess whether curable STI play a decreasing role in HIV transmission in generalized epidemics, the simulated population-attributable fractions(PAFs) of incident HIV attributable to the curable STIs(syphilis, chancroid, chlamydia and gonorrhoea), HSV-2 and all STIs combined were calculated using the STDSIM model. The cofactor effects for HIV susceptibility and infectivity for each group of STIs were removed over a two-year period and the PAF was calculated according to the formula24:

equation M1

where IRnocofactor is the HIV incidence among 15-49 year olds over two years in simulations with the cofactor removed, and IRdefaultcofactor is the corresponding incidence in simulations with default cofactor effects.

The simulated intervention was based on data from the syndromic treatment intervention tested in Mwanza, Tanzania2 as quantified in a previous modelling study9. The simulated intervention was assumed to increase the average proportion of symptomatic STI episodes cured, and the proportion of steady partners of treated patients notified and cured, from 0% to 27%9. Treatment of the incurable infection HSV-2 was not simulated in this study but has been explored elsewhere23. The ratio of HIV incidence among adults aged 15–49 years over 2 years in simulations with and without the intervention was used to compare the predicted impact of the intervention between cities.

In separate simulations the PAFs, the intervention impacts and the cost-effectiveness were calculated over two years starting 4, 8, 12, 14 and 16 years into the HIV epidemics in all four cities and also 18 and 20 years into the epidemic in Kisumu. Since epidemic onset in the model was 1980 in Kisumu and 1984 in the other cities to better fit the time-trend in HIV prevalence, the latest period(16 or 20 years into the epidemic) corresponded to the period 2000-1 in all four cities. Results were based on means over 300 simulation runs for each scenario.


The annual incremental cost of the improved syndromic STI treatment intervention in Mwanza was estimated by Gilson et al to be US$54,839 in 1993 prices4. By assuming that the at-risk population size was 72,000(75,000 sexually active adults, of which 4% were already HIV-infected) and that the absolute reduction in annual HIV incidence due to the intervention was 0.35/100pyrs, Gilson et al estimated that 252 HIV infections were averted by the intervention each year, at a cost of US$218 per infection-averted.

To compare this estimate for Mwanza in the early 1990s to that for the four cities over the HIV epidemic, the predicted impacts of the intervention were obtained from the simulations described above for the four cities, and from the results of a previous study that re-estimated the impact of this intervention in the early HIV epidemic in Mwanza that found that the true impact was likely to be slightly lower than the point estimate measured in the trial9.

The costs were based on Gilson4 and adjusted to reflect changes over time(including inflation and changes in treatment regimens) and between countries. Adjustment factors were obtained from a systematic review and regression analysis of studies containing unit STI cost data25. Per capita gross domestic product(GDP) data were obtained from the World Bank’s World Development Indicators26. The adjustment factor for changes over time was 1.0258 per year and the adjustment factor for differences in per capita GDP was 0.9998 per US$. These adjustments were applied to Gilsons’ total fixed costs and unit variable costs. The adjusted unit variable costs were then multiplied by the simulated annual number of STI episodes treated per 75,000 adults in each city in the year in which they were implemented to obtain total variable costs. This was added to their respective fixed costs to obtain the full costs. As in Gilson, the control arm costs were subtracted from the full costs to obtain the incremental costs of the intervention. All costs were then adjusted for inflation to 2001 US$27, 28. Cost per HIV-infection-averted was calculated and compared to a recent estimate of the lifetime treatment costs of an HIV infection in Africa(US$3,500)29.

Generalisability of findings to other settings and sensitivity analysis

The simulated HIV and STI prevalences and the simulated impact of the intervention on HIV incidence will be very sensitive to trends in condom use. To explore the likely generalisability of our findings for settings with differing condom use trends and HIV prevalence, the effect on PAFs, intervention impacts and costs-per-infection-averted in 2000/1, were re-calculated for additional scenarios for each city in which simulated condom use rates were multiplied by zero, 0.5, and 1.5.

To assess the robustness of the results to uncertainties in model input parameters known to affect cost-effectiveness, the Ndola simulations were repeated varying the values of each parameter in turn. For cofactor magnitudes, condom use rates and the proportion of symptomatic STI episodes cured by STI treatment, default parameter values were doubled and halved. For cost adjustments, the effect of each adjustment factor was removed in turn. Alternative scenarios were created that were re-fitted to the observed HIV prevalence data in Ndola using the same procedure used to develop the default scenarios as described above. The relative and absolute impacts of the intervention and the cost per infection-averted were then re-calculated and compared for each parameter in turn. The parameter values used in the sensitivity analysis are shown in Table S3, online.


Simulated population-attributable fractions

The combined contribution of the simulated STIs to HIV transmission decreased over time, but remained high throughout the HIV epidemics in all four cities(Figure 2). The simulated PAFs for combined STIs ranged between 80-87% during years 4-5 of the HIV epidemics, and 50-70% in years 16-17.

Figure 2
Population attributable fraction of STIs for HIV transmission

However, the combined PAF masks different patterns for curable STIs and for HSV-2. In line with our previous findings19, the simulated PAFs for curable STIs decreased markedly during the HIV epidemics, from 54-79% in years 4-5, to 13-33% in years 16-17. This was because of reductions in STI prevalences(Figure 1, online) due to simulated increases in condom use(Table 1, online), and to the spread of HIV from higher to lower risk groups with lower prevalences of curable STIs. Conversely, simulated PAFs for HSV-2 increased during the HIV epidemics, from 8-27% in years 4-5, to 25-37% in years 16-17. This was partly due to the fall in the PAF for curable STIs as HIV spread into lower risk groups in which HSV-2 was still common, but also partly because of simulated increases in HSV-2 ulceration among HSV-2/HIV coinfected individuals due to HIV-related immunosuppression.

Table 1
Annual cost per HIV-infection-averted by syndromic STI case management in an early HIV epidemic in Mwanza, 1992-3 and in the HIV epidemics in the four cities, 2000-1.

Simulated impact of syndromic STD case management

The relative impact of improved syndromic STD case management on HIV incidence fell markedly during the HIV epidemics(Figure 3, Left Panel). The relative reduction in HIV incidence due to the intervention decreased in all four cities from between 39% and 2% in years 4-5, to between 9% and 1% in years 16-17. This was because of reductions in the PAF for curable STIs(Figure 2), particularly those that are most amenable to syndromic treatment such as the highly-symptomatic STI, chancroid.

Figure 3
Relative impact of syndromic STI case management on HIV incidence and estimated cost-effectiveness


The estimated cost per HIV-infection-averted due to syndromic STI treatment in the early epidemic in Mwanza was US$317 in 2001 prices. Table 1 shows this alongside the estimate made by Gilson in 1993 prices(US$218). Our estimate is higher than Gilson’s because it was adjusted for inflation, other changes over time(such as changes in treatment regimen) and was based on the smaller absolute reduction in HIV incidence and smaller number of STIs treated predicted by the model, compared with estimates from the trial data used by Gilson. Adjusting Gilson’s estimate for each of these factors in turn, the cost-per-HIV-infection-averted increased as follows: from US$218 to US$252 after adjustment for inflation between 1993 and 2001; from US$252 to US$307 after adjustment for other changes over time (such as changes in treatment regimen); and from US$307 to US$317 after adjustment for the differences between the modelled and observed impact on HIV incidence and the number of treated STIs.

In 2000-1, the simulated absolute reduction in HIV incidence due to the intervention in the four cities ranged widely between 0.01/100pyrs in Cotonou and 0.41/100pyrs in Ndola(Table 1). Despite the smaller relative impact, the predicted absolute reduction in HIV incidence in Ndola was larger than in the early HIV epidemic in Mwanza during 1992-3(0.28/100pyrs) because of the much higher HIV incidence at this stage of the epidemic in Ndola.

Over the same period, the predicted cost-effectiveness of the intervention varied markedly between the four cities. The cost per-infection-averted was US$321, US$1,499, US$1,665, and US$4,976 in Ndola, Kisumu, Yaounde, and Cotonou respectively(Table 1), suggesting, by comparison with the US$3,500 estimate of the lifetime treatment costs of an HIV infection in Africa29, that the intervention would be cost-saving in three of the four sites. The highest cost per infection-averted was predicted for Cotonou because of the small absolute impact of the intervention. This small impact was due to the combination of the very low STI prevalences(Figure S1, online) and the lower HIV incidence in this city. The intervention was more cost-effective in Kisumu and Yaounde because of the large relative impact of the intervention(Yaounde) and the higher HIV incidence(Kisumu). The intervention was most cost-effective in Ndola where a larger relative impact of the intervention combined with a high HIV incidence.

The trends in the cost-effectiveness of the syndromic treatment intervention over the simulated HIV epidemics are shown in Figure 3(right). Estimated cost-per-infection-averted was very low four years into the HIV epidemics(below US$500) and then increased over time in all cities except Cotonou where it remained high throughout the epidemic.

The increase in the cost per-infection-averted was predicted to occur earlier in Yaounde than in Ndola/Kisumu because of the earlier fall in the relative impact of the intervention on HIV incidence in Yaounde(Fig 3,left). This earlier fall was primarily due to the larger impact of the simulated increase in condom use in 1990 on chancroid prevalence, because the predicted basic reproduction number of chancroid was closer to one in Yaounde than in Kisumu/Ndola and therefore it was more sensitive to small changes in condom use. As reported previously17, chancroid prevalence was predicted to be lower in Yaounde than in Kisumu and Ndola because of the much higher prevalence of male circumcision in Yaounde. Male circumcision has been shown to be associated with a lower risk of chancroid infection30. Similarly, the cost per-infection-averted by the intervention increased rapidly in Kisumu 14 years into the HIV epidemic, as increasing rates of condom use resulted in adequate control of chancroid.

Generalisability of findings to other settings and sensitivity analysis

As expected, the cost-effectiveness of the intervention was sensitive to the simulated trends in condom use(Figure 4A). As condom use rates increased, HIV and STI rates fell and the cost per HIV-infection-averted tended to increase. Our results suggest that in populations like Cotonou with relatively low risk sexual behaviour, near universal male circumcision, and consequently relatively low HIV prevalence, syndromic treatment in the general population may cease to be cost-saving when condom use rates rise above 10-20% of casual contacts and 25-50% of sex-worker contacts. However, in populations with either relatively high risk behaviour or low rates of male circumcision(such as Ndola, Kisumu and Yaounde) our findings suggest that syndromic treatment in the general population may remain cost-saving even with condom use rates as high as 30-37.5% of casual contacts and 45% of sex worker contacts, or higher. In these mature African HIV epidemics, syndromic treatment was always found to be cost-saving if HIV prevalence was above ~5.5%(Figure 4B).

Figure 4
(A) Cost per HIV-infection-averted by trend in condom use rates, and (B) Cost per HIV-infection-averted by projected HIV prevalence in year 2000 for the 16 model scenarios shown in 4A

In line with these findings, the impact and cost-effectiveness predictions for Ndola were shown to be most sensitive to the assumed trend in condom use(Table S3 online). If condom use was doubled, STI treatment was predicted to be less cost-effective because of the lower STI rates(cost per-HIV-infection-averted in 2000-1=US$2,321 and US$276, respectively), and vice versa. By comparison, the cost-effectiveness predictions were shown to be relatively robust to doubling and halving of the magnitude of the STI cofactor effects(US$406 and US$287, respectively) and to the assumptions used to calculate the costs of the intervention(range = US$277 to US$327). Interestingly, the cost-effectiveness of the intervention was also shown to be relatively robust to doubling or halving of the proportion of symptomatic STI episodes cured by the intervention(US$326 and US$372, respectively). This was because, over this relatively short 2-year follow-up period, increases or decreases in the number of HIV infections averted was roughly balanced by increases or decreases in variable costs due to treating more or fewer STI episodes. Results for the other three cities were similar(not shown).


We have explored the changing role of STI cofactors in HIV transmission during the HIV epidemics in four cities in West and East Africa. Throughout the HIV epidemics the contribution of STIs to HIV transmission remained high, with 50% or more of HIV transmission attributed to STIs in all four cities. This relative stability in the overall PAFs concealed opposing trends in the contribution of curable and incurable STIs. The attributable-fraction for curable STIs was predicted to fall during the HIV epidemic, while the attributable-fraction for HSV-2 was predicted to rise.

This has important implications for the(cost-)effectiveness of interventions seeking to target curable STIs. Over time in all four cities, the relative impact of STI treatment on HIV incidence fell, tending to increase the cost per HIV-infection-averted. However, in populations with a rapidly expanding HIV epidemic this increase in cost was offset by the rapidly rising HIV incidence that increased the absolute impact of the intervention. Our findings suggest that in African populations with mature HIV epidemics, STI treatment interventions are likely to remain highly cost-effective and may even be cost-saving, particularly in populations in which safer sexual behaviours have not adequately controlled STIs and HIV incidence remains high.

These results should be interpreted with some caution. The simulated HIV epidemics in the four cities were largely parameterised using data from a cross sectional study carried out in 199711 with limited data on STI trends over time. Although the observed sexual behaviours and STI rates in 1997 and HIV prevalence trends over time were fitted well, actual trends in STI rates may have differed from those simulated. This may affect the quantitative findings of this study, but is less likely to affect the qualitative conclusions.

The magnitudes of STI cofactor effects on HIV transmission remain poorly quantified20, 21. However our sensitivity analysis showed that the cost-effectiveness of the intervention was relatively insensitive to their variation.

Although the simulated STI treatment intervention was based on extensive review of data from the RCT in Mwanza, and adequately explained the observed impacts of this empirical trial and the observed impacts of the STI treatment trials in Masaka and Rakai9, the coverage achieved by this intervention may vary between populations. However, as our sensitivity analysis showed, in the short term this may not affect the cost-effectiveness of the intervention. Over the longer term, increased cure rates would be expected to increase the cost-effectiveness of the intervention.

The adjustments to the cost data were based on the results of an analysis of data from developing countries. Although these adjustments enabled us to estimate the true costs of the intervention more accurately, they are approximations and cannot fully control for differences in implementation and changes over time.

Our findings may help to explain why the STI treatment strategies tested since the trial in Mwanza have failed to show a significant impact on HIV incidence. Our results suggest that even if the STI treatment interventions in Rakai, Masaka and Zimbabwe5, 6, 8 had led to large reductions in STI rates, in these generalised HIV epidemics the relative impact on HIV incidence would have been relatively small and therefore undetectable.

This study suggests that despite the small relative impact on HIV incidence, STI treatment in the general population is likely to remain highly cost-effective and even cost-saving in contemporary African populations with HIV prevalences over ~5.5%, because of the substantial absolute impact on HIV incidence. In African populations with lower HIV prevalences, syndromic STI treatment may not be as cost-effective for HIV prevention, but is recommended for STI control and to reduce HIV transmission among STI patients and their sexual partners.

Rational health policy requires that scarce resources be allocated to interventions with the best cost-effectiveness even if relative impact at population level is modest. Effective STI management may prevent a decreasing fraction of new HIV infections as the epidemic expands, but is an inexpensive intervention with important collateral public health benefits and which effectively protects STI patients from the enhanced risk of HIV acquisition and transmission.

In conclusion, our study has shown that throughout HIV epidemics in sub-Saharan Africa, a large proportion of HIV transmission may remain attributable to other STIs, but that the proportion attributable to curable STIs is likely to fall. Despite this, we have shown that in populations with generalised HIV epidemics, interventions that target curable STIs may remain highly cost-effective and even cost-saving if changes in risk behaviours have not adequately controlled STIs.

Supplementary Material


Table S1. Condom use assumptions. This table shows the condom use assumptions used in the default scenarios. Condoms use was assumed to prevent STI and HIV transmission in 90% of contacts.

Table S2. Sensitivity of population attributable fraction, intervention impact and cost-effectiveness to assumed condom use trends (2000-1, adults 15-49 years). HIVp= HIV prevalence; HDp= chancroid prevalence; HIVi= HIV incidence; PAF= population attributable fraction; STI= sexually transmitted infection; HSV-2=Herpes Simplex Virus, Type 2.

Table S3. Sensitivity of population attributable fraction, intervention impact and cost-effectiveness in Ndola to selected model assumptions (2000-1, adults 15-49 years). Data were not available for blank cells. HIVp= HIV prevalence; HDp= chancroid prevalence; HIVi= HIV incidence; PAF= population attributable fraction; STI= sexually transmitted infection; HSV-2=Herpes Simplex Virus, Type 2.

Figure S1: Simulated prevalence of curable STIs over time in the four cities (%, 15-49 years). Trends reflect increasing condom use in all sites and increased risk behaviour in Yaounde (the timing in 1990, 1993 and 1995 is highlighted with vertical lines).


Sources of support: This study was funded by the Wellcome Trust, Grant No. 069509/Z/02/Z. The funders had no involvement in the design, collection, analysis or interpretation of the data, in writing the report or in the decision to submit.


Previous presentation of results: Some of the material presented in this paper was prepared for and presented at a meeting organised and funded by WHO, UK MRC and UNAIDS: ‘Consultation on STI interventions for preventing HIV: appraisal of the evidence’, Geneva, July 2006.


1. UNAIDS/WHO 2006 Report on the global AIDS epidemic. UNAIDS/ WHO; Geneva: 2006.
2. Grosskurth H, Mosha F, Todd J, et al. Impact of improved treatment of sexually transmitted diseases on HIV infection in rural Tanzania: randomised controlled trial. Lancet. 1995;346(8974):530–536. [PubMed]
3. Hayes R, Grosskurth H, ka-Gina G. Impact of improved treatment of sexually transmitted diseases on HIV infection in rural Tanzania: randomised controlled trial [Letter to Lancet] Lancet. 1995;346:1159–1160. [PubMed]
4. Gilson L, Mkanje R, Grosskurth H, et al. Cost-effectiveness of improved treatment services for sexually transmitted diseases in preventing HIV-1 infection in Mwanza Region, Tanzania. Lancet. 1997;350(9094):1805–1809. [PubMed]
5. Wawer MJ, Sewankambo NK, Serwadda D, et al. Control of Sexually Transmitted diseases for AIDS prevention in Uganda: a randomised community trial. Lancet. 1999;353(9152):525–535. [PubMed]
6. Kamali A, Kinsman J, Nalweyiso N, et al. A community randomized controlled trial to investigate impact of improved STD management and behavioural interventions on HIV incidence in rural Masaka, Uganda: trial design, methods and baseline findings. Tropical Medicine and International Health. 2002 Dec;7(12):1053–1063. 2002. [PubMed]
7. Kaul R, Kimani J, Nagelkerke NJ, et al. Monthly antibiotic chemoprophylaxis and incidence of sexually transmitted infections and HIV-1 infection in Kenyan sex workers: a randomized controlled trial. JAMA. 2004 Jun 2;291(21):2555–2562. [PubMed]
8. Gregson S, Adamson S, Papaya S, et al. Impact and Process Evaluation of Integrated Community and Clinic-Based HIV-1 Control: A Cluster-Randomised Trial in Eastern Zimbabwe. PLoS Med. 2007 Mar 27;4(3):e102. [PMC free article] [PubMed]
9. White RG, Orroth KK, Korenromp EL, et al. Can Population Differences Explain the Contrasting Results of the Mwanza, Rakai, and Masaka HIV/Sexually Transmitted Disease Intervention Trials?: A Modeling Study. J Acquir Immune Defic Syndr. 2004 Dec 1;37(4):1500–1513. [PubMed]
10. Korenromp EL, White RG, Orroth KK, 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 Feb 1;191(Suppl 1):S168–178. [PubMed]
11. Buve A, Carael M, Hayes RJ, 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 Aug;15(Suppl 4):S5–14. [PubMed]
12. Buve A, Carael M, Hayes RJ, et al. The multicentre study on factors determining the differential spread of HIV in four African cities: summary and conclusions. AIDS. 2001 Aug;15(Suppl 4):S127–131. [PubMed]
13. Auvert B, Buvé A, Lagarde E, et al. Male circumcision and HIV infection in four cities in sub-Saharan Africa. AIDS. 2001;15(suppl 4):S31–S40. [PubMed]
14. Weiss H, Buvé A, Robinson N, et al. The epidemiology of HSV-2 infection and its association with HIV infection in four urban African populations. AIDS. 2001;15(suppl 4):S97–S108. [PubMed]
15. Ferry B, Carael M, Buve A, et al. Comparison of key parameters of sexual behaviour in four African urban populations with different levels of HIV infection. AIDS. 2001 Aug;15(Suppl 4):S41–50. [PubMed]
16. Morison L, Weiss HA, Buve A, et al. Commercial sex and the spread of HIV in four cities in sub-Saharan Africa. AIDS. 2001 Aug;15(Suppl 4):S61–69. [PubMed]
17. Orroth KK, Freeman E, Bakker R, et al. Understanding differences across the contrasting epidemics in East and West Africa: results from a simulation model of the Four Cities Study. STI. 2007;83:i5–i16. [PubMed]
18. 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. Mathematical Population Studies. 2000;8(2):135–173.
19. Freeman E, Orroth K, White RG, et al. The proportion of new HIV infections attributable to HSV-2 increases over time: simulations of the changing role of sexually transmitted infections in sub-Saharan African HIV epidemics. STI. 2007;83:i17–i24. [PubMed]
20. Hayes RJ, Schulz KF, Plummer FA. The cofactor effect of genital ulcers on the per-exposure risk of HIV transmission in sub-Saharan Africa. [Review] J Trop Med Hyg. 1995;98(1):1–8. [PubMed]
21. Korenromp EL, De Vlas S, Nagelkerke N, Habbema J. Estimating the magnitude of STD cofactor effects on HIV transmission - how well can it be done? STD. 2001;28(11):613–621. [PubMed]
22. Rottingen JA, Cameron DW, Garnett GP. A systematic review of the epidemiologic interactions between classic sexually transmitted diseases and HIV: how much really is known? Sex Transm Dis. 2001;28(10):579–597. [PubMed]
23. Freeman E. The Role Of Herpes Simplex Virus Type 2 In The Spread And Control Of HIV In Four Sub-Saharan African Cities. Department of Infectious & Tropical Diseases, London School of Hygiene and Tropical Medicine, London University; London: 2006.
24. Orroth KK, White RG, Korenromp EL, et al. 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 Sep;33(9):536–544. [PubMed]
25. Terris-Prestholt F, Vyas S, Kumaranayake L, Mayaud P, Watts C. The costs of treating curable sexually transmitted infections in low- and middle-income countries: a systematic review. Sex Transm Dis. 2006 Oct;33(10 Suppl):S153–166. [PubMed]
26. World Bank World Development Indicators. 2005.
27. Kumaranayake L. The real and the nominal? Making inflationary adjustments to cost and other economic data. Health Policy Plan. 2000 Jun;15(2):230–234. [PubMed]
28. Economic Research Service ERS International Macroeconomic Data Set. Economic Research Service, United States Department of Agriculture; 2005.
29. Stover J, Bertozzi S, Gutierrez JP, et al. The global impact of scaling up HIV/AIDS prevention programs in low- and middle-income countries. Science. 2006 Mar 10;311(5766):1474–1476. [PubMed]
30. Weiss HA, Thomas SL, Munabi SK, Hayes RJ. Male circumcision and risk of syphilis, chancroid, and genital herpes: a systematic review and meta-analysis. Sex Transm Infect. 2006 Apr;82(2):101–109. discussion 110. [PMC free article] [PubMed]