Evaluating large-scale HIV prevention interventions: study design for an integrated mathematical modelling approach

Population-level effectiveness of pre-exposure prophylaxis for HIV prevention among men who have sex with men in Montréal (Canada): a modelling study of surveillance and survey data

INTRODUCTION

HIV pre-exposure prophylaxis (PrEP) has been recommended and partly subsidized in Québec, Canada, since 2013. We evaluated the population-level impact of PrEP on HIV transmission among men who have sex with men (MSM) in Montréal, Québec's largest city, over 2013-2021.

METHODS

We used an agent-based mathematical model of sexual HIV transmission to estimate the fraction of HIV acquisitions averted by PrEP compared to a counterfactual scenario without PrEP. The model was calibrated to local MSM survey, surveillance, and cohort data and accounted for COVID-19 pandemic impacts on sexual activity, HIV prevention, and care. PrEP was modelled from 2013 onwards, assuming 86% individual-level effectiveness. The PrEP eligibility criteria were: any anal sex unprotected by condoms (past 6 months) and either multiple partnerships (past 6 months) or multiple uses of post-exposure prophylaxis (lifetime). To assess potential optimization strategies, we modelled hypothetical scenarios prioritizing PrEP to MSM with high sexual activity (≥11 anal sex partners annually) or aged ⩽45 years, increasing coverage to levels achieved in Vancouver, Canada (where PrEP is free-of-charge), and improving retention.

RESULTS

Over 2013-2021, the estimated annual HIV incidence decreased from 0.4 (90% credible interval [CrI]: 0.3-0.6) to 0.2 (90% CrI: 0.1-0.2) per 100 person-years. PrEP coverage among HIV-negative MSM remained low until 2015 (<1%). Afterwards, coverage increased to a maximum of 10% of all HIV-negative MSM, or about 16% of the 62% PrEP-eligible HIV-negative MSM in 2020. Over 2015-2021, PrEP averted an estimated 20% (90% CrI: 11%-30%) of cumulative HIV acquisitions. The hypothetical scenarios modelled showed that, at the same coverage level, prioritizing PrEP to high sexual activity MSM could have averted 30% (90% CrI: 19%-42%) of HIV acquisitions from 2015-2021. Even larger impacts could have resulted from higher coverage. Under the provincial eligibility criteria, reaching 10% coverage among HIV-negative MSM in 2015 and 30% in 2019, like attained in Vancouver, could have averted up to 63% (90% CrI: 54%-70%) of HIV acquisitions from 2015 to 2021.

CONCLUSIONS

PrEP reduced population-level HIV transmission among Montréal MSM. However, our study suggests missed prevention opportunities and adds support for public policies that reduce PrEP barriers, financial or otherwise, to MSM at risk of HIV acquisition.

An SEIR Model with Time-Varying Coefficients for Analyzing the SARS-CoV-2 Epidemic

In this study, we propose a time-dependent susceptible-exposed-infected-recovered (SEIR) model for the analysis of the SARS-CoV-2 epidemic outbreak in three different countries, the United States, Italy, and Iceland using public data inherent the numbers of the epidemic wave. Since several types and grades of actions were adopted by the governments, including travel restrictions, social distancing, or limitation of movement, we want to investigate how these measures can affect the epidemic curve of the infectious population. The parameters of interest for the SEIR model were estimated employing a composite likelihood approach. Moreover, standard errors have been corrected for temporal dependence. The adoption of restrictive measures results in flatten epidemic curves, and the future evolution indicated a decrease in the number of cases.

Transitions: Novel Study Methods to Understand Early HIV Risk Among Adolescent Girls and Young Women in Mombasa, Kenya, and Dnipro, Ukraine

Transitions aims to understand the human immunodeficiency virus (HIV) risk at critical transition points in the sexual life course of adolescent girls and young women (AGYW) who engage in casual sex, transactional sex, and sex work. In this article, we present the Transitions study methods. The Transitions study has the following objectives: (1) to describe how the characteristics and length of the transition period and access gap vary across two epidemiological contexts (Mombasa, Kenya, and Dnipro, Ukraine); (2) to understand how the risk of HIV varies by length and characteristics of the transition period and access gap across epidemiologic contexts; and (3) to assess the extent to which HIV infections acquired during the transition period and access gap could mitigate the population-level impact of focused interventions for female sex workers and explore the potential marginal benefit of expanding programs to reach AGYW during the transition period and access gap. Cross-sectional biobehavioral data were collected from young women aged 14 to 24 years who were recruited from locations in Mombasa County, Kenya, and Dnipro, Ukraine, where sex work took place. Data are available for 1,299 Kenyan and 1,818 Ukrainian participants. The survey addressed the following areas: timing of transition events (first sex, first exchange of sex for money or other resources, self-identification as sex workers, entry into formal sex work, access to prevention program services); sexual behaviors (condom use, anal sex, sex under the influence of drugs or alcohol); partnerships (regular and first-time clients, regular and first-time transactional sex partners, and husbands and boyfriends); alcohol use; injection and non-injection illicit drug use; experience of violence; access to HIV prevention and treatment program; testing for sexually transmitted and blood-borne infections and HIV; and reproductive health (pregnancies, abortions, contraceptives). HIV and hepatitis C virus prevalence data were based on rapid test results. Mathematical modeling will be used to generate projections of onward HIV transmission at specific transition points in the sexual life course of AGYW. Taken together, these data form a novel data resource providing comprehensive behavioral, structural, and biological data on a high-risk group of AGYW in two distinct sociocultural and epidemiologic contexts.

Evaluating the population impact of hepatitis C direct acting antiviral treatment as prevention for people who inject drugs (EPIToPe) - a natural experiment (protocol)

INTRODUCTION

Hepatitis C virus (HCV) is the second largest contributor to liver disease in the UK, with injecting drug use as the main risk factor among the estimated 200 000 people currently infected. Despite effective prevention interventions, chronic HCV prevalence remains around 40% among people who inject drugs (PWID). New direct-acting antiviral (DAA) HCV therapies combine high cure rates (>90%) and short treatment duration (8 to 12 weeks). Theoretical mathematical modelling evidence suggests HCV treatment scale-up can prevent transmission and substantially reduce HCV prevalence/incidence among PWID. Our primary aim is to generate empirical evidence on the effectiveness of HCV 'Treatment as Prevention' (TasP) in PWID.

METHODS AND ANALYSIS

We plan to establish a natural experiment with Tayside, Scotland, as a single intervention site where HCV care pathways are being expanded (including specialist drug treatment clinics, needle and syringe programmes (NSPs), pharmacies and prison) and HCV treatment for PWID is being rapidly scaled-up. Other sites in Scotland and England will act as potential controls. Over 2 years from 2017/2018, at least 500 PWID will be treated in Tayside, which simulation studies project will reduce chronic HCV prevalence among PWID by 62% (from 26% to 10%) and HCV incidence will fall by approximately 2/3 (from 4.2 per 100 person-years (p100py) to 1.4 p100py). Treatment response and re-infection rates will be monitored. We will conduct focus groups and interviews with service providers and patients that accept and decline treatment to identify barriers and facilitators in implementing TasP. We will conduct longitudinal interviews with up to 40 PWID to assess whether successful HCV treatment alters their perspectives on and engagement with drug treatment and recovery. Trained peer researchers will be involved in data collection and dissemination. The primary outcome - chronic HCV prevalence in PWID - is measured using information from the Needle Exchange Surveillance Initiative survey in Scotland and the Unlinked Anonymous Monitoring Programme in England, conducted at least four times before and three times during and after the intervention. We will adapt Bayesian synthetic control methods (specifically the Causal Impact Method) to generate the cumulative impact of the intervention on chronic HCV prevalence and incidence. We will use a dynamic HCV transmission and economic model to evaluate the cost-effectiveness of the HCV TasP intervention, and to estimate the contribution of the scale-up in HCV treatment to observe changes in HCV prevalence. Through the qualitative data we will systematically explore key mechanisms of TasP real world implementation from provider and patient perspectives to develop a manual for scaling up HCV treatment in other settings. We will compare qualitative accounts of drug treatment and recovery with a 'virtual cohort' of PWID linking information on HCV treatment with Scottish Drug treatment databases to test whether DAA treatment improves drug treatment outcomes.

ETHICS AND DISSEMINATION

Extending HCV community care pathways is covered by ethics (ERADICATE C, ISRCTN27564683, Super DOT C Trial clinicaltrials.gov: NCT02706223). Ethical approval for extra data collection from patients including health utilities and qualitative interviews has been granted (REC ref: 18/ES/0128) and ISCRCTN registration has been completed (ISRCTN72038467). Our findings will have direct National Health Service and patient relevance; informing prioritisation given to early HCV treatment for PWID. We will present findings to practitioners and policymakers, and support design of an evaluation of HCV TasP in England.

FROM INDIVIDUAL HEALTH TO COMMUNITY HEALTH: TOWARDS MULTISCALE MODELING OF DIRECTLY TRANSMITTED INFECTIOUS DISEASE SYSTEMS

In this paper, we present a new method for developing a class of nested multiscale models for directly transmitted infectious disease systems that integrates within-host scale and between-host scale using community pathogen load (CPL) as a new public health measure of a community’s level of infectiousness and as an indicator of the effectiveness of health interventions. The approach develops a multiscale modeling science base for directly transmitted infectious disease systems (where the inside-host environment’s biological entities such as cells, tissues, organs, body fluids, whole body are the reservoir of infective pathogen in the community) that is comparable to an existing multiscale modeling science base for environmentally transmitted infectious diseases (where the outside-host geographical environment’s physical entities such as soil, air, formites/contact surfaces, food and water are the reservoir of infective pathogen in the community) where pathogen load in the environment is explicitly incorporated into the model. This is achieved by assuming that infected hosts in the community are homogeneous and unevenly distributed microbial habitats. We illustrate the utility of this multiscale modeling methodology by evaluating the comparative effectiveness of HIV/AIDS preventive and treatment interventions as a case study.

HIV prevention costs and their predictors: evidence from the ORPHEA Project in Kenya

We estimate costs and their predictors for three HIV prevention interventions in Kenya: HIV testing and counselling (HTC), prevention of mother-to-child transmission (PMTCT) and voluntary medical male circumcision (VMMC). As part of the 'Optimizing the Response of Prevention: HIV Efficiency in Africa' (ORPHEA) project, we collected retrospective data from government and non-governmental health facilities for 2011-12. We used multi-stage sampling to determine a sample of health facilities by type, ownership, size and interventions offered totalling 144 sites in 78 health facilities in 33 districts across Kenya. Data sources included key informants, registers and time-motion observation methods. Total costs of production were computed using both quantity and unit price of each input. Average cost was estimated by dividing total cost per intervention by number of clients accessing the intervention. Multivariate regression methods were used to analyse predictors of log-transformed average costs. Average costs were $7 and $79 per HTC and PMTCT client tested, respectively; and $66 per VMMC procedure. Results show evidence of economies of scale for PMTCT and VMMC: increasing the number of clients per year by 100% was associated with cost reductions of 50% for PMTCT, and 45% for VMMC. Task shifting was associated with reduced costs for both PMTCT (59%) and VMMC (54%). Costs in hospitals were higher for PMTCT (56%) in comparison to non-hospitals. Facilities that performed testing based on risk factors as opposed to universal screening had higher HTC average costs (79%). Lower VMMC costs were associated with availability of male reproductive health services (59%) and presence of community advisory board (52%). Aside from increasing production scale, HIV prevention costs may be contained by using task shifting, non-hospital sites, service integration and community supervision.

A Bayesian approach to estimate changes in condom use from limited human immunodeficiency virus prevalence data

Evaluation of large‐scale intervention programmes against human immunodeficiency virus (HIV) is becoming increasingly important, but impact estimates frequently hinge on knowledge of changes in behaviour such as the frequency of condom use over time, or other self‐reported behaviour changes, for which we generally have limited or potentially biased data. We employ a Bayesian inference methodology that incorporates an HIV transmission dynamics model to estimate condom use time trends from HIV prevalence data. Estimation is implemented via particle Markov chain Monte Carlo methods, applied for the first time in this context. The preliminary choice of the formulation for the time varying parameter reflecting the proportion of condom use is critical in the context studied, because of the very limited amount of condom use and HIV data available. We consider various novel formulations to explore the trajectory of condom use over time, based on diffusion‐driven trajectories and smooth sigmoid curves. Numerical simulations indicate that informative results can be obtained regarding the amplitude of the increase in condom use during an intervention, with good levels of sensitivity and specificity performance in effectively detecting changes. The application of this method to a real life problem demonstrates how it can help in evaluating HIV interventions based on a small number of prevalence estimates, and it opens the way to similar applications in different contexts.

Estimation and Short-Term Prediction of the Course of the HIV Epidemic Using Demographic and Health Survey Methodology-Like Data

BACKGROUND

Mathematical models have played important roles in the understanding of epidemics and in the study of the impacts of various behavioral or medical measures. However, modeling accurately the future spread of an epidemic requires context-specific parameters that are difficult to estimate because of lack of data. Our objective is to propose a methodology to estimate context-specific parameters using Demographic and Health Survey (DHS)-like data that can be used in mathematical modeling of short-term HIV spreading.

METHODS AND FINDINGS

The model splits the population according to sex, age, HIV status, and antiretroviral treatment status. To estimate context-specific parameters, we used individuals' histories included in DHS-like data and a statistical analysis that used decomposition of the Poisson likelihood. To predict the course of the HIV epidemic, sex- and age-specific differential equations were used. This approach was applied to recent data from Kenya. The approach allowed the estimation of several key epidemiological parameters. Women had a higher infection rate than men and the highest infection rate in the youngest age groups (15-24 and 25-34 years) whereas men had the highest infection rate in age group 25-34 years. The immunosuppression rates were similar between age groups. The treatment rate was the highest in age group 35-59 years in both sexes. The results showed that, within the 15-24 year age group, increasing male circumcision coverage and antiretroviral therapy coverage at CD4 ≤ 350/mm3 over the current 70% could have short-term impacts.

CONCLUSIONS

The study succeeded in estimating the model parameters using DHS-like data rather than literature data. The analysis provides a framework for using the same data for estimation and prediction, which can improve the validity of context-specific predictions and help designing HIV prevention campaigns.

Potential impact of existing interventions and of antiretroviral use in female sex workers on transmission of HIV in Burkina Faso: a modeling study

BACKGROUND

The impact and cost-effectiveness of antiretroviral treatment (ART) as prevention is likely to vary depending on the local context. Burkina Faso has a concentrated mature HIV epidemic where female sex workers (FSW) are thought to have driven HIV transmission.

METHODS

A dynamic HIV transmission model was developed using data from the Yerelon FSW cohort in Bobo-Dioulasso and population surveys. Compared with current ART provision [status quo (SQ)], the model estimated the proportion of HIV infections averted or incremental life-years gained per additional person-year of ART over 20 years for ART targeting different subgroups or expanding eligibility to all HIV-infected individuals compared with SQ.

RESULTS

Modeling suggests that condom use within commercial sex has averted 40% of past HIV infections. Continuing SQ averts 35%-47% of new infections over 20 years compared with no ART. Expanding ART eligibility to all HIV-infected individuals and increasing recruitment (80% per year) could avert a further 65% of new infections, whereas targeting full-time FSW or all FSWs achieved less impact but was more efficient in terms of life-years gained per 100 person-years of ART. Local HIV elimination is possible with expanded ART provision to FSWs but requires condom use within commercial sex to be maintained at high levels.

CONCLUSIONS

Increasing FSW recruitment onto ART could be a highly efficient method for reducing HIV transmission in concentrated epidemic settings but should not be undertaken at the expense of existing interventions for FSWs. Specialized clinics providing multiple interventions for FSWs should be a fundamental component of prevention in concentrated epidemics.

Increased HIV prevention program coverage and decline in HIV prevalence among female sex workers in south India

BACKGROUND

As one way of assessing the impact of Avahan, the India AIDS Initiative of the Bill & Melinda Gates Foundation, we examined the association between HIV prevention program indicators and changes in HIV prevalence among female sex workers (FSWs) between 2005 and 2009.

METHODS

We conducted a secondary data analysis from 2 large cross-sectional surveys (2005-2006 and 2008-2009) across 24 districts in south India (n = 11,000 per round). A random-effect multilevel logistic regression analysis was performed using HIV as the outcome, with individual independent variables (from both surveys) at level 1 and district-level FSW-specific program indicators and contextual variables at level 2. Program indicators included their 2006 value, the difference in their values between 2008 and 2006, and the interaction between this difference and study round.

RESULTS

HIV prevalence among FSWs decreased from 17.0% to 14.2% (P < 0.001). This decline varied significantly (P = 0.006) across levels of difference in program coverage (% of FSWs contacted by the program in a given year). Odds ratios comparing HIV prevalence between rounds changed with the level of increase in coverage and were statistically significant with coverage increase ≥ quartile (Q) 1: odds ratio, 0.85 at Q1; 0.78 at Q2; 0.66 at Q3; and 0.51 at Q4.

CONCLUSIONS

These findings suggest that increased program coverage was associated with declining HIV prevalence among FSWs covered by the Avahan program. The triangulation of our results with those from other approaches used in evaluating Avahan suggests a major impact of this intervention on the HIV epidemic in southern India.

Relationship between exposure to the Avahan intervention and levels of reported condom use among men who have sex with men in southern India

BACKGROUND

The Avahan intervention promotes consistent (100%) condom use amongst men who have sex with men in southern India. We assessed how condom use varies with intervention exposure for men who have sex with men in Bangalore.

METHODS

Self-reported condom use and intervention exposure data were derived from a cross-sectional survey. Consistent condom use and condom use at last sex act with all, main, and casual male sex partners were assessed. Binary and continuous variables reflecting intervention exposure (including contact(s) with intervention staff, receiving condoms and seeing condom demonstrations) were used. Multivariable logistic regression was employed to assess the relationship between condom use with each type of partner and each exposure variable independently, controlling for socio-demographic and behavioural factors associated with condom use or intervention exposure.

RESULTS

Condom use with all partners was higher among those who had ever been contacted by, received condoms from, or seen a condom demonstration by intervention staff (adjusted odds ratio >2, p < 0.02 for all). Consistent condom use with all types of partner increased with the number of condom demonstrations seen in the last month (adjusted odds ratio = 2.1 per demonstration, p < 0.025), while condom use at last sex act with a casual (but not main) partner increased with the number of condoms received from the intervention (adjusted odds ratio = 1.4 per condom, p = 0.04).

CONCLUSIONS

Direct contact with Avahan program staff is associated with increased reported condom use among men who have sex with men in Bangalore. Reported consistent condom use and condom use at last sex act are associated with contacts involving demonstrations of correct condom use, and with receiving condoms, respectively.

Community mobilisation and empowerment interventions as part of HIV prevention for female sex workers in Southern India: a cost-effectiveness analysis

BACKGROUND

Most HIV prevention for female sex workers (FSWs) focuses on individual behaviour change involving peer educators, condom promotion and the provision of sexual health services. However, there is a growing recognition of the need to address broader societal, contextual and structural factors contributing to FSW risk behaviour. We assess the cost-effectiveness of adding community mobilisation (CM) and empowerment interventions (eg. community mobilisation, community involvement in programme management and services, violence reduction, and addressing legal policies and police practices), to core HIV prevention services delivered as part of Avahan in two districts (Bellary and Belgaum) of Karnataka state, Southern India.

METHODS

An ingredients approach was used to estimate economic costs in US$ 2011 from an HIV programme perspective of CM and empowerment interventions over a seven year period (2004-2011). Incremental impact, in terms of HIV infections averted, was estimated using a two-stage process. An 'exposure analysis' explored whether exposure to CM was associated with FSW's empowerment, risk behaviours and HIV/STI prevalence. Pathway analyses were then used to estimate the extent to which behaviour change may be attributable to CM and to inform a dynamic HIV transmission model.

FINDINGS

The incremental costs of CM and empowerment were US$ 307,711 in Belgaum and US$ 592,903 in Bellary over seven years (2004-2011). Over a 7-year period (2004-2011) the mean (standard deviation, sd.) number of HIV infections averted through CM and empowerment is estimated to be 1257 (308) in Belgaum and 2775 (1260) in Bellary. This translates in a mean (sd.) incremental cost per disability adjusted life year (DALY) averted of US$ 14.12 (3.68) in Belgaum and US$ 13.48 (6.80) for Bellary--well below the World Health Organisation recommended willingness to pay threshold for India. When savings from ART are taken into account, investments in CM and empowerment are cost saving.

CONCLUSIONS

Our findings suggest that CM and empowerment is, at worst, highly cost-effective and, at best, a cost-saving investment from an HIV programme perspective. CM and empowerment interventions should therefore be considered as core components of HIV prevention programmes for FSWs.

Modelling in concentrated epidemics: informing epidemic trajectories and assessing prevention approaches

PURPOSE OF THE REVIEW

This review summarizes recent mathematical modelling studies conducted among key populations including MSM, people who inject drugs (PWID), and female sex workers (FSWs) in low prevalence settings used as a marker of concentrated epidemics.

RECENT FINDINGS

Most recent studies focused on MSM, Asian settings or high-income countries, studied the transmission dynamics or modelled pre-exposure prophylaxis, treatment as prevention or behavioural interventions specific to each key population (e.g., needle exchange programme or use of low-dead space syringes for PWID). Biological interventions were deemed effective and cost-effective, though still expensive, and often deemed unlikely to result in HIV elimination if used alone. Targeting high-risk individuals even within key populations improved efficiency. Some studies made innovative use of models to formally evaluate HIV prevention programmes, to interpret genetic or co-infection data, and to address methodological questions and validate epidemiological tools.

CONCLUSION

More work is needed to optimize combination prevention focusing on key populations in different settings. The gaps identified include the limited number of studies modelling drug resistance, structural interventions, treatment as prevention among FSWs, and estimating the contribution of key populations to overall transmission in different settings.

Cost-effectiveness of HIV prevention for high-risk groups at scale: an economic evaluation of the Avahan programme in south India

BACKGROUND

Avahan is a large-scale, HIV preventive intervention, targeting high-risk populations in south India. We assessed the cost-effectiveness of Avahan to inform global and national funding institutions who are considering investing in worldwide HIV prevention in concentrated epidemics.

METHODS

We estimated cost effectiveness from a programme perspective in 22 districts in four high-prevalence states. We used the UNAIDS Costing Guidelines for HIV Prevention Strategies as the basis for our costing method, and calculated effect estimates using a dynamic transmission model of HIV and sexually transmitted disease transmission that was parameterised and fitted to locally observed behavioural and prevalence trends. We calculated incremental cost-effective ratios (ICERs), comparing the incremental cost of Avahan per disability-adjusted life-year (DALY) averted versus a no-Avahan counterfactual scenario. We also estimated incremental cost per HIV infection averted and incremental cost per person reached.

FINDINGS

Avahan reached roughly 150 000 high-risk individuals between 2004 and 2008 in the 22 districts studied, at a mean cost per person reached of US$327 during the 4 years. This reach resulted in an estimated 61 000 HIV infections averted, with roughly 11 000 HIV infections averted in the general population, at a mean incremental cost per HIV infection averted of $785 (SD 166). We estimate that roughly 1 million DALYs were averted across the 22 districts, at a mean incremental cost per DALY averted of $46 (SD 10). Future antiretroviral treatment (ART) cost savings during the lifetime of the cohort exposed to HIV prevention were estimated to be more than $77 million (compared with the slightly more than $50 million spent on Avahan in the 22 districts during the 4 years of the study).

INTERPRETATION

This study provides evidence that the investment in targeted HIV prevention programmes in south India has been cost effective, and is likely to be cost saving if a commitment is made to provide ART to all that can benefit from it. Policy makers should consider funding and sustaining large-scale targeted HIV prevention programmes in India and beyond.

FUNDING

Bill & Melinda Gates Foundation.

Two cross-sectional studies in south India assessing the effect of an HIV prevention programme for female sex workers on reducing syphilis among their clients

Objective

To assess the impact of the Avahan HIV prevention programme for female sex workers (FSWs) in south India on reducing syphilis prevalence among their clients, by comparing rates of syphilis over time as reported in two large-scale surveys of FSWs’ clients.

Methods

A random-effect multilevel logistic regression analysis was performed using syphilis as the dependent variable, with individual independent variables (from the two survey rounds) at level 1 and the district-level programme (from the Avahan computerised monitoring and information system) and contextual variables (from Indian government datasets) at level 2. Programme variables included their 2006 value and their difference in value between 2008 and 2006, as well as the interaction between the latter and the study round. The analysis also controlled for baseline syphilis prevalence and its interaction with the study round.

Results

Syphilis decreased significantly among FSWs’ clients, from 4.8% (round 1) to 2.6% (round 2), p<0.001. The OR of the interaction term between the difference in programme coverage of FSWs and the round was 0.98 (p=0.023), suggesting that increased coverage was associated with a reduced incidence of syphilis.

Conclusions

This study suggests that the Avahan intervention programme among FSWs reduced syphilis rates among their clients.

Assessing outcomes of large-scale public health interventions in the absence of baseline data using a mixture of Cox and binomial regressions

Background

Large-scale public health interventions with rapid scale-up are increasingly being implemented worldwide. Such implementation allows for a large target population to be reached in a short period of time. But when the time comes to investigate the effectiveness of these interventions, the rapid scale-up creates several methodological challenges, such as the lack of baseline data and the absence of control groups. One example of such an intervention is Avahan, the India HIV/AIDS initiative of the Bill & Melinda Gates Foundation. One question of interest is the effect of Avahan on condom use by female sex workers with their clients. By retrospectively reconstructing condom use and sex work history from survey data, it is possible to estimate how condom use rates evolve over time. However formal inference about how this rate changes at a given point in calendar time remains challenging.

Methods

We propose a new statistical procedure based on a mixture of binomial regression and Cox regression. We compare this new method to an existing approach based on generalized estimating equations through simulations and application to Indian data.

Results

Both methods are unbiased, but the proposed method is more powerful than the existing method, especially when initial condom use is high. When applied to the Indian data, the new method mostly agrees with the existing method, but seems to have corrected some implausible results of the latter in a few districts. We also show how the new method can be used to analyze the data of all districts combined.

Conclusions

The use of both methods can be recommended for exploratory data analysis. However for formal statistical inference, the new method has better power.

Assessment of the population-level effectiveness of the Avahan HIV-prevention programme in South India: a preplanned, causal-pathway-based modelling analysis

BACKGROUND

Avahan, the India AIDS initiative of the Bill & Melinda Gates Foundation, was a large-scale, targeted HIV prevention intervention. We aimed to assess its overall effectiveness by estimating the number and proportion of HIV infections averted across Avahan districts, following the causal pathway of the intervention.

METHODS

We created a mathematical model of HIV transmission in high-risk groups and the general population using data from serial cross-sectional surveys (integrated behavioural and biological assessments, IBBAs) within a Bayesian framework, which we used to reproduce HIV prevalence trends in female sex workers and their clients, men who have sex with men, and the general population in 24 South Indian districts over the first 4 years (2004-07 or 2005-08 dependent on the district) and the full 10 years (2004-13) of the Avahan programme. We tested whether these prevalence trends were more consistent with self-reported increases in consistent condom use after the implementation of Avahan or with a counterfactual (assuming consistent condom use increased at slower, pre-Avahan rates) using a Bayes factor, which gave a measure of the strength of evidence for the effectiveness estimates. Using regression analysis, we extrapolated the prevention effect in the districts covered by IBBAs to all 69 Avahan districts.

FINDINGS

In 13 of 24 IBBA districts, modelling suggested medium to strong evidence for the large self-reported increase in consistent condom use since Avahan implementation. In the remaining 11 IBBA districts, the evidence was weaker, with consistent condom use generally already high before Avahan began. Roughly 32700 HIV infections (95% credibility interval 17900-61600) were averted over the first 4 years of the programme in the IBBA districts with moderate to strong evidence. Addition of the districts with weaker evidence increased this total to 62800 (32000-118000) averted infections, and extrapolation suggested that 202000 (98300-407000) infections were averted across all 69 Avahan districts in South India, increasing to 606000 (290000-1 193000) over 10 years. Over the first 4 years of the programme 42% of HIV infections were averted, and over 10 years 57% were averted.

INTERPRETATION

This is the first assessment of Avahan to account for the causal pathway of the intervention, that of changing risk behaviours in female sex workers and high-risk men who have sex with men to avert HIV infections in these groups and the general population. The findings suggest that substantial preventive effects can be achieved by targeted behavioural HIV prevention initiatives.

FUNDING

Bill & Melinda Gates Foundation.

Can the UNAIDS modes of transmission model be improved? A comparison of the original and revised model projections using data from a setting in west Africa

Objective:

The UNAIDS modes of transmission model (MoT) is a user-friendly model, developed to predict the distribution of new HIV infections among different subgroups. The model has been used in 29 countries to guide interventions. However, there is the risk that the simplifications inherent in the MoT produce misleading findings. Using input data from Nigeria, we compare projections from the MoT with those from a revised model that incorporates additional heterogeneity.

Methods:

We revised the MoT to explicitly incorporate brothel and street-based sex-work, transactional sex, and HIV-discordant couples. Both models were parameterized using behavioural and epidemiological data from Cross River State, Nigeria. Model projections were compared, and the robustness of the revised model projections to different model assumptions, was investigated.

Results:

The original MoT predicts 21% of new infections occur in most-at-risk-populations (MARPs), compared with 45% (40–75%, 95% Crl) once additional heterogeneity and updated parameterization is incorporated. Discordant couples, a subgroup previously not explicitly modelled, are predicted to contribute a third of new HIV infections. In addition, the new findings suggest that women engaging in transactional sex may be an important but previously less recognized risk group, with 16% of infections occurring in this subgroup.

Conclusion:

The MoT is an accessible model that can inform intervention priorities. However, the current model may be potentially misleading, with our comparisons in Nigeria suggesting that the model lacks resolution, making it challenging for the user to correctly interpret the nature of the epidemic. Our findings highlight the need for a formal review of the MoT.

Positive impact of a large-scale HIV prevention programme among female sex workers and clients in South India

OBJECTIVE

Estimate the potential impact of Avahan, the India AIDS Initiative, among female sex workers (FSWs) and their clients in five districts of Karnataka state, south India.

DESIGN

Examination of time trends in sexually transmitted infection (STI)/HIV prevalence from serial cross-sectional surveys, combined with mathematical modelling.

METHODS

Survey data from each district were used to monitor changes in FSW STI/HIV prevalence during Avahan. A deterministic model, parameterized with district-specific survey data, was used to simulate HIV/HSV-2/syphilis transmission among high-risk groups in each district. Latin hypercube sampling was used to obtain multiple parameter sets that reproduced district-specific HIV prevalence trends. A Bayesian framework tested whether self-reported increases in consistent condom use (CCU) during Avahan were more compatible with FSW HIV prevalence trends than assuming no or slow (preintervention rates) CCU increases, and were used to estimate HIV incidence and infections averted.

RESULTS

Declines in FSW HIV prevalence occurred over 5 years in all districts, and were statistically significant in three. Self-reported increases in CCU were more consistent with observed declines in HIV prevalence in three districts. In all five districts, an estimated 25-64% (32-70%) HIV infections were averted among FSWs (clients) over 5 years. This corresponded to 142-2092 FSW infections averted depending on the district (two-fold to nine-fold more among clients).

CONCLUSION

Empirical HIV prevalence trends combined with Bayesian modelling have provided plausible evidence that Avahan has reduced HIV transmission among FSWs and their clients. If current CCU levels are sustained, FSW HIV prevalence could decline to low levels by 2015, with many more infections averted.

Assessment of HIV-related risky behaviour: a comparative study of face-to-face interviews and polling booth surveys in the general population of Cotonou, Benin

OBJECTIVES

During the 2008 HIV prevalence survey carried out in the general population of Cotonou, Benin, face-to-face interviews (FTFI) were used to assess risky behaviours for HIV and other sexually transmitted infections (STI). We compared sexual behaviours reported in FTFI with those reported in polling booth surveys (PBS) carried out in parallel in an independent random sample of the same population.

METHODS

In PBS, respondents grouped by gender and marital status answered simple questions by putting tokens with question numbers in a green box (affirmative answers) or a red box (negative answers). Both boxes were placed inside a private booth. For each group and question, data were gathered together by type of answer. The structured and gender-specific FTFI guided by trained interviewers included all questions asked during PBS. Pearson χ2 or Fisher's exact test was used to compare FTFI and PBS according to affirmative answers.

RESULTS

Overall, respondents reported more stigmatised behaviours in PBS than in FTFI: the proportions of married women and men who reported ever having had commercial sex were 17.4% and 41.6% in PBS versus 1.8% and 19.6% in FTFI, respectively. The corresponding proportions among unmarried women and men were 16.1% and 25.5% in PBS versus 3.9% and 13.0% in FTFI, respectively. The proportion of married women who reported having had extramarital sex since marriage was 23.6% in PBS versus 4.6% in FTFI.

CONCLUSIONS

PBS are suitable to monitor reliable HIV/STI risk behaviours. Their use should be expanded in behavioural surveillance.

Differences between seven measures of self-reported numbers of clients of female sex workers in southern India: implications for individual- and population-level analysis

Quantifying sexual activity of sub-populations with high-risk sexual behaviour is important in understanding HIV epidemiology. This study examined inconsistency of seven outcomes measuring self-reported clients per month (CPM) of female sex workers (FSWs) in southern India and implications for individual/population-level analysis. Multivariate negative binomial regression was used to compare key social/environmental factors associated with each outcome. A transmission dynamics model was used to assess the impact of differences between outcomes on population-level FSW/client HIV prevalence. Outcomes based on 'clients per last working day' produced lower estimates than those based on 'clients per typical day'. Although the outcomes were strongly correlated, their averages differed by approximately two-fold (range 39.0-79.1 CPM). The CPM measure chosen did not greatly influence standard epidemiological 'risk factor' analysis. Differences across outcomes influenced HIV prevalence predictions. Due to this uncertainty, we recommend basing population-based estimates on the range of outcomes, particularly when assessing the impact of interventions.

Capturing the time-varying drivers of an epidemic using stochastic dynamical systems

Epidemics are often modeled using non-linear dynamical systems observed through partial and noisy data. In this paper, we consider stochastic extensions in order to capture unknown influences (changing behaviors, public interventions, seasonal effects, etc.). These models assign diffusion processes to the time-varying parameters, and our inferential procedure is based on a suitably adjusted adaptive particle Markov chain Monte Carlo algorithm. The performance of the proposed computational methods is validated on simulated data and the adopted model is applied to the 2009 H1N1 pandemic in England. In addition to estimating the effective contact rate trajectories, the methodology is applied in real time to provide evidence in related public health decisions. Diffusion-driven susceptible exposed infected retired-type models with age structure are also introduced.

Polling booth surveys: a novel approach for reducing social desirability bias in HIV-related behavioural surveys in resource-poor settings

This study compared rates of HIV-related sexual risk behaviours reported in individual face-to-face (FTFI) and group anonymous polling booth (PBS) interviews in India. In PBS, respondents grouped by gender and marital status answered yes/no questions by putting tokens with question numbers in colour-coded containers. Data were subsequently collated for each group as a whole, so responses were not traceable back to individuals. Male and female PBS participants reported substantially higher rates of pre-marital, extra-marital, commercial and anal sex than FTFI participants; e.g. 11 vs. 2% married males reported paying for sex; 6 vs. 1% unmarried males reported homosexual anal sex.

HIV treatment as prevention: considerations in the design, conduct, and analysis of cluster randomized controlled trials of combination HIV prevention

The rigorous evaluation of the impact of combination HIV prevention packages at the population level will be critical for the future of HIV prevention. In this review, we discuss important considerations for the design and interpretation of cluster randomized controlled trials (C-RCTs) of combination prevention interventions. We focus on three large C-RCTs that will start soon and are designed to test the hypothesis that combination prevention packages, including expanded access to antiretroviral therapy, can substantially reduce HIV incidence. Using a general framework to integrate mathematical modelling analysis into the design, conduct, and analysis of C-RCTs will complement traditional statistical analyses and strengthen the evaluation of the interventions. Importantly, even with combination interventions, it may be challenging to substantially reduce HIV incidence over the 2- to 3-y duration of a C-RCT, unless interventions are scaled up rapidly and key populations are reached. Thus, we propose the innovative use of mathematical modelling to conduct interim analyses, when interim HIV incidence data are not available, to allow the ongoing trials to be modified or adapted to reduce the likelihood of inconclusive outcomes. The preplanned, interactive use of mathematical models during C-RCTs will also provide a valuable opportunity to validate and refine model projections.

A dose-response relationship between exposure to a large-scale HIV preventive intervention and consistent condom use with different sexual partners of female sex workers in southern India

Background

The Avahan Initiative, a large-scale HIV preventive intervention targeted to high-risk populations including female sex workers (FSWs), was initiated in 2003 in six high-prevalence states in India, including Karnataka. This study assessed if intervention exposure was associated with condom use with FSWs’ sexual partners, including a dose-response relationship.

Methods

Data were from a cross-sectional study (2006-07) of 775 FSWs in three districts in Karnataka. Survey methods accounted for the complex cluster sampling design. Bivariate and multivariable logistic regression was used to separately model the relationships between each of five intervention exposure variables and five outcomes for consistent condom use (CCU= always versus frequently/sometimes/never) with different sex partners, including with: all clients; occasional clients; most recent repeat client; most recent non-paying partner; and the husband or cohabiting partner. Linear tests for trends were conducted for three continuous intervention exposure variables.

Results

FSWs reported highest CCU with all clients (81.7%); CCU was lowest with FSWs’ husband or cohabiting partner (9.6%). In multivariable analysis, the odds of CCU with all clients and with occasional clients were 6.3-fold [95% confidence intervals, CIs: 2.8-14.5] and 2.3-fold [95% CIs: 1.4-4.1] higher among FSWs contacted by intervention staff and 4.9-fold [95% CIs: 2.6-9.3] and 2.3-fold [95% CIs: 1.3-4.1] higher among those who ever observed a condom demonstration by staff, respectively, compared to those who had not. A significant dose-response relationship existed between each of these CCU outcomes and increased duration since first contacted by staff (P=0.001; P=0.006) and numbers of condom demonstrations witnessed (P=0.004; P=0.026); a dose-response relationship was also observed between condom use with all clients and number of times contacted by staff (P=0.047). Intervention exposure was not associated with higher odds of CCU with the most recent repeat client, most recent non-paying partner or with the husband or cohabiting partner.

Conclusion

Study findings suggest that exposure to a large-scale HIV intervention for FSWs was associated with increased CCU with commercial clients. Moreover, there were dose-response relationships between CCU with clients and increased duration since first contacted by staff, times contacted by staff and number of condom demonstrations. Additional program effort is required to increase condom use with non-commercial partners.

Mathematical models for the study of HIV spread and control amongst men who have sex with men

For a quarter of century, mathematical models have been used to study the spread and control of HIV amongst men who have sex with men (MSM). We searched MEDLINE and EMBASE databases up to the end of 2010 and reviewed this literature to summarise the methodologies used, key model developments, and the recommended strategies for HIV control amongst MSM. Of 742 studies identified, 127 studies met the inclusion criteria. Most studies employed deterministic modelling methods (80%). Over time we saw an increase in model complexity regarding antiretroviral therapy (ART), and a corresponding decrease in complexity regarding sexual behaviours. Formal estimation of model parameters was carried out in only a small proportion of the studies (22%) while model validation was considered by an even smaller proportion (17%), somewhat reducing confidence in the findings from the studies. Nonetheless, a number of common conclusions emerged, including (1) identification of the importance of assumptions regarding changes in infectivity and sexual contact rates on the impact of ART on HIV incidence, that subsequently led to empirical studies to gather these data, and (2) recommendation that multiple strategies would be required for effective HIV control amongst MSM. The role of mathematical models in studying epidemics is clear, and the lack of formal inference and validation highlights the need for further developments in this area. Improved methodologies for parameter estimation and systematic sensitivity analysis will help generate predictions that more fully express uncertainty, allowing better informed decision making in public health.

Mathematical models in the evaluation of health programmes

Modelling is valuable in the planning and evaluation of interventions, especially when a controlled trial is ethically or logistically impossible. Models are often used to calculate the expected course of events in the absence of more formal assessments. They are also used to derive estimates of rare or future events from recorded intermediate points. When developing models, decisions are needed about the appropriate level of complexity to be represented and about model structure and assumptions. The degree of rigor in model development and assessment can vary greatly, and there is a danger that existing beliefs inappropriately influence judgments about model assumptions and results.

Uptake of biomedical interventions for prevention of sexually transmitted HIV

PURPOSE OF THE REVIEW

To examine the population-level effects of introducing and/or expanding biomedical interventions for prevention of human immunodeficiency virus (HIV) sexually transmitted infections through mathematical modeling.

RECENT FINDINGS

Successes of several ground-breaking clinical trials have invigorated the field of HIV prevention with new enthusiasm and opportunities for research into and application of biomedical HIV prevention. Mathematical modeling has advanced in tandem with valuable contributions to both investigative science and public health. New models provide qualitative and quantitative insights regarding the epidemiological impact of the uptake of biomedical interventions, singly and/or in combination including treatment of sexually transmitted infections, condom use, male circumcision, antiretroviral treatment and pre-exposure prophylaxis and vaccine for HIV prevention.

SUMMARY

Biomedical interventions are critical for reversing the HIV pandemic. Mathematical modeling is invaluable for informed biomedical HIV prevention research, policy and practice.

Impact of targeted interventions on heterosexual transmission of HIV in India

Background

Targeted interventions (TIs) have been a major strategy for HIV prevention in India. We evaluated the impact of TIs on HIV prevalence in high HIV prevalence southern states (Tamil Nadu, Karnataka, Andhra Pradesh and Maharashtra).

Methods

A quasi-experimental approach was used to retrospectively compare changes in HIV prevalence according to the intensity of targeted intervention implementation. Condom gap (number of condoms required minus condoms supplied by TIs) was used as an indicator of TI intensity. Annual average number of commercial sex acts per female sex worker (FSW) reported in Behavioral Surveillance Survey was multiplied by the estimated number of FSWs in each district to calculate annual requirement of condoms in the district. Data of condoms supplied by TIs from 1995 to 2008 was obtained from program records. Districts in each state were ranked into quartiles based on the TI intensity. Primary data of HIV Sentinel Surveillance was analyzed to calculate HIV prevalence reductions in each successive year taking 2001 as reference year according to the quartiles of TI intensity districts using generalized linear model with logit link and binomial distribution after adjusting for age, education, and place of residence (urban or rural).

Results

In the high HIV prevalence southern states, the number of TI projects for FSWs increased from 5 to 310 between 1995 and 2008. In high TI intensity quartile districts (n = 30), 186 condoms per FSW/year were distributed through TIs as compared to 45 condoms/FSW/year in the low TI intensity districts (n = 29). Behavioral surveillance indicated significant rise in condom use from 2001 to 2009. Among FSWs consistent condom use with last paying clients increased from 58.6% to 83.7% (p < 0.001), and among men of reproductive age, the condom use during sex with non-regular partner increased from 51.7% to 68.6% (p < 0.001). A significant decline in HIV and syphilis prevalence has occurred in high prevalence southern states among FSWs and young antenatal women. Among young (15-24 years) antenatal clinic attendees significant decline was observed in HIV prevalence from 2001 to 2008 (OR = 0.42, 95% CI 0.28-0.62) in high TI intensity districts whereas in low TI intensity districts the change was not significant (OR = 1.01, 95% CI 0.67-1.5).

Conclusion

Targeted interventions are associated with HIV prevalence decline.

Assessment of intervention outcome in the absence of baseline data: 'reconstruction' of condom use time trends using retrospective analysis of survey data

Background

Avahan, the India AIDS Initiative, is a large-scale targeted intervention. Data on condom use prior to Avahan are unavailable. The authors used a novel method of analysing cross-sectional survey data to ‘reconstruct’ condom use rates in presurvey years, and to assess the relationship between Avahan and time trends in condom use among female sex workers (FSWs).

Methods

Among FSWs reporting consistent condom use (CCU) with clients in surveys from 21 districts in southern India (n=7358), data on length of time using condoms, and length of time selling sex, were converted into number of FSWs using condoms (numerator) and selling sex (denominator) by year, to give yearly rates of CCU from 2001 to year of survey. Linear regression with generalised estimating equations was used to assess time trends and to compare the rate of increase in condom use before and after Avahan initiation in 2003.

Results

In all 21 districts, the rate of increase in CCU from 2001 to time of the surveys was highly significant. Overall CCU increased from 27% (22%) with occasional (regular) clients in 2001, to 76% (68%) in 2006, respectively. The yearly rate of increase in CCU was significantly greater after (slope 2003–2006: 12.7% (12.0%) per year for occasional (regular) clients) than prior to Avahan (slope 2001–2003: 5.1% (5.3%) per year for occasional (regular) clients) implementation (p<0.0001).

Conclusions

The findings indicate a positive relationship between implementation of the Avahan programme and rates of CCU increase among FSW. This method of analysis may be useful in other contexts where preintervention data are lacking.

Interim modelling analysis to validate reported increases in condom use and assess HIV infections averted among female sex workers and clients in southern India following a targeted HIV prevention programme

Objectives

This study assesses whether the observed declines in HIV prevalence since the beginning of the ‘Avahan’ India HIV/AIDS prevention initiative are consistent with self-reported increases in condom use by female sex workers (FSWs) in two districts of southern India, and provides estimates of the fraction of new infections averted among FSWs and clients due to increases in condom use in commercial sex after 2004.

Methods

A deterministic compartmental model of HIV/sexually transmitted infection (STI) transmission incorporating heterogeneous sexual behaviour was developed, parameterised and fitted using data from two districts in Karnataka, India. Three hypotheses of condom use among FSWs were tested: (H₀), that condom use increased in line with reported FSW survey data prior to the Avahan initiative but remained constant afterwards; (H₁) that condom use increased following the Avahan initiative, in accordance with survey data; (H₂) that condom use increased according to estimates derived from condom distribution data. The proportion of fits to HIV/STI prevalence data was examined to determine which hypothesis was most consistent.

Results

For Mysore 0/36/82.7 fits were identified per million parameter sets explored under hypothesis H₀/H₁/H₂, respectively, while for Belgaum 9.7/8.3/0 fits were identified. The HIV epidemics in Belgaum and Mysore are both declining. In Mysore, increases in condom use during commercial sex between 2004 and 2009 may have averted 31.2% to 47.4% of new HIV infections in FSWs, while in Belgaum it may have averted 24.8% to 43.2%, if there was an increase in condom use.

Discussion

Increased condom use following the Avahan intervention is likely to have played a role in curbing the HIV epidemic in Mysore. In Belgaum, given the limitations in available data, this method cannot be used alone to decide if there has been an increase in condom use.

Sexual identity and its contribution to MSM risk behavior in Bangaluru (Bangalore), India: the results of a two-stage cluster sampling survey

In India, there are categories of MSM (hijras, kothis, double-deckers, panthis and bisexuals), which are generally associated with different HIV-risk behaviors. Our objective was to quantify differences across MSM identities (n = 357) and assess the extent they conform to typecasts that prevail in policy-orientated discourse. More feminine kothis (26%) and hijras (13%) mostly reported receptive sex, and masculine panthis (15%) and bisexuals (23%) insertive anal sex. However, behavior did not always conform to expectation, with 25% and 16% of the sample reporting both insertive and receptive anal intercourse with known and unknown noncommercial partners, respectively (p < 0.000). Although behavior often complied with stereotyped role and identity, male-with-male sexual practices were fluid. Reification of these categories in an intervention context may hinder our understanding of the differential HIV risk among MSM.

Dynamic modeling of herpes simplex virus type-2 (HSV-2) transmission: issues in structural uncertainty

The sexually transmitted infection (STI) Herpes simplex virus type-2 (HSV-2) is of public health concern because it is a very common frequently unrecognized lifelong infection, which may facilitate HIV transmission. Within HIV/STI modeling, structural uncertainty has received less attention than parametric uncertainty. By merging the compartments of a "complex" model, a "simple" HSV-2 model is developed. Sexual interactions between female sex workers (FSWs) and clients are modeled using data from India. Latin Hypercube Sampling selects from parameter distributions and both models are run for each of the 10,000 parameter sets generated. Outputs are compared (except for 2,450 unrealistic simulations). The simple model is a good approximation to the complex model once the HSV-2 epidemic has reached 60% of the equilibrium prevalence (95% of the 7,550 runs produced <10% relative error). The simple model is a reduced version of the complex model that retains details implicitly. For late-stage epidemics, the simple model gives similar prevalence trends to the complex model. As HSV-2 epidemics in many populations are advanced, the simple model is accurate in most instances, although the complex model may be preferable for early epidemics. The analysis highlights the issue of structural uncertainty and the value of reducing complexity.

Baseline integrated behavioural and biological assessment among most at-risk populations in six high-prevalence states of India: design and implementation challenges

OBJECTIVE

This paper presents key methodological approaches and challenges in implementing and analysing the first round of the integrated biobehavioural assessment of most-at-risk populations, conducted in conjunction with evaluation of Avahan, the India AIDS initiative.

METHODS

The survey collected data on HIV risk behaviours, sexually transmitted infections and HIV prevalence in 29 districts in six high-prevalence states of India. Groups included female sex workers and clients, men who have sex with men, injecting drug users and truck drivers. Strategies for overcoming some challenges of the large-scale surveys among vulnerable populations, including sampling hidden populations, involvement of the communities targeted by the survey, laboratory and quality control in remote, non-clinic field settings, and data analysis and data use are presented.

DISCUSSION

Satisfying the need for protocols, guidelines and tools that allowed for sufficient standardization, while being tailored enough to fit diverse local situations on such a large scale, with so many implementing partners, emerged as a major management challenge. A major lesson from the first round is the vital importance of investing upfront time in tailoring the sampling methods, data collection instruments, and analysis plan to match measurement objectives.

CONCLUSION

Despite the challenges, the integrated biobehavioural assessment was a huge achievement, and was largely successful in providing previously unavailable information about the HIV situation among populations that are critical to the curtailment of HIV spread in India. Lessons from the first round will be used to evolve the second round into an exercise with increased evaluative capability for Avahan.

Prevalence and determinants of HIV and sexually transmitted infections in a general population-based sample in Mysore district, Karnataka state, southern India

OBJECTIVES

To estimate HIV and sexually transmitted infection (STI) prevalence within the general population of Mysore district, and to examine differences in the distribution of risk factors associated with HIV prevalence stratified by sex.

METHODS

A community-based study was conducted in Mysore, Karnataka state, southern India, between October 2005 and November 2006; final sample size 4653. A face-to-face interview was conducted, and blood and urine specimens collected to measure HIV and STI prevalences. Risk factors for HIV among men and women were examined using weighted and clustered logistic regression.

RESULTS

Weighted HIV prevalence was 0.8% [95% confidence interval (CI) 0.52-1.09] overall and 0.7% (0.35-1.08) and 0.9% (0.51-1.37) in rural and urban populations, respectively. The prevalence of syphilis, gonorrhoea and chlamydial infection was 2.8% for men and 1.8% for women. In multivariate analysis, higher HIV prevalence was associated with ever having used a condom [odds ratio (OR) 2.75, 95% CI 1.01-7.47] and number of lifetime partners for men (OR 6.9, 95% CI 2.18-21.91). For women, HIV infection was associated with condom use at last sexual intercourse (OR 10.51, 95% CI 2.05-53.79), number of lifetime partners and reporting 'don't know' for whether ever had anal sex (OR 9.10, 95% CI 1.14-72.34).

CONCLUSIONS

HIV prevalence in the general population of Mysore was found to be comparable to recent prevalence estimates for Karnataka state, and also similar to recent prevalence estimates from antenatal clinic attenders for the district. Few modifiable risk factors for HIV infection were identified. There is evidence from this study that high-risk behaviour may have been underreported, but the prevalence of STI was generally low.

HIV, sexually transmitted infections and sexual behaviour of male clients of female sex workers in Andhra Pradesh, Tamil Nadu and Maharashtra, India: results of a cross-sectional survey

OBJECTIVE

To characterize and describe patterns of HIV, sexually transmitted infections (STI) and sexual behaviour of male clients of female sex workers (FSW).

METHODS

A cross-sectional study was conducted among 4821 clients of FSW from 12 districts in three states in India: Andhra Pradesh, Maharashtra and Tamil Nadu. A structured questionnaire was administered to elicit demographic characteristics and sexual behavioural patterns. Blood and urine samples were tested for HIV, syphilis and herpes simplex type 2 serology, gonococcal and chlamydial infection.

RESULTS

The median age of clients surveyed was 30 years; 57% were married and 64% had a steady sexual partner; 61% had sex with more than one FSW and 39% had four or more sexual encounters in the past month. The prevalence of HIV ranged from 2.0% to 10.9%, syphilis ranged from 3.1% to 10.1%; gonorrhoea and chlamydia ranged between 0% and 4.5%. Clients older than 30 years [odds ratio (OR) 1.65; 95% confidence interval (CI) 1.33-2.05] and having a different mix of commercial and non-commercial partners (OR 1.56; 95% CI 1.25-1.96) had a higher volume of sex acts with FSW. Inconsistent condom use with FSW was significantly associated with older clients (OR 4.2; 95% CI 3.33-5.29), illiteracy (OR 1.39; 95% CI 1.14-1.69), age <18 years at first paid sex (OR 1.83; 95% CI 1.24-2.70) and having different FSW partners in the past month (OR 1.64; 95% CI 1.23-2.18).

CONCLUSION

Clients of FSW studied here constitute a significant bridge group for HIV and other STI, because of their high volume of different sexual partners and very low consistent condom use. HIV prevention programmes need to address male clients.

Evaluation design for large-scale HIV prevention programmes: the case of Avahan, the India AIDS initiative

BACKGROUND

Closing the HIV prevention gap to prevent HIV infections requires rapid, worldwide rollout of large-scale national programmes. Evaluating such programmes is challenging and complex, requiring clarity of evaluation purpose and evidential approaches substantively different to those employed for pilots and small programmes.

OBJECTIVES

This paper describes the evaluation design for the implementation phase of Avahan, the India AIDS initiative, a large HIV prevention programme funded by the Bill and Melinda Gates Foundation. Avahan, which began in December 2003, has a 10-year charter to impact the Indian epidemic and its response by implementing an HIV prevention programme targeting core and bridge groups in 83 districts of six Indian states, transferring the programme to the Government of India, and disseminating programme learning.

METHODS

The foundation commissioned an external process to design Avahan's evaluation framework. An independent advisory group oversees and guides course corrections in the execution of this framework.

RESULTS

Avahan's evaluation framework comprises: trend and synthetic analysis of data from core, bridge and household biobehavioural surveys in a subset of intervention districts, denominator estimates and programme monitoring from all intervention districts, and government's antenatal surveillance (two sites per district in all districts); bespoke transmission dynamics modelling to estimate infections averted (subset of districts); cost effectiveness studies (subset of districts). In addition, there are other knowledge-building and quality-monitoring activities.

CONCLUSION

Rather than a small set of monofocal outcome measures, scaled programmes require nuanced evaluations that approximate programmatic scale by collecting data with different levels of geographical scope, synthesize multiple data and methods to arrive at a composite picture, and can cope with continuous environmental and programme evolution.

Declines in risk behaviour and sexually transmitted infection prevalence following a community-led HIV preventive intervention among female sex workers in Mysore, India

OBJECTIVE

To investigate the impact on sexual behaviour and sexually transmitted infections (STI) of a comprehensive community-led intervention programme for reducing sexual risk among female sex workers (FSW) in Mysore, India. The key programme components were: community mobilization and peer-mediated outreach; increasing access to and utilization of sexual health services; and enhancing the enabling environment to support programme activities.

METHODS

Two cross-sectional surveys among random samples of FSW were conducted 30 months apart, in 2004 and 2006.

RESULTS

Of over 1000 women who sell sex in Mysore city, 429 participated in the survey at baseline and 425 at follow-up. The median age was 30 years, median duration in sex work 4 years, and the majority were street based (88%). Striking increases in condom use were seen between baseline and follow-up surveys: condom use at last sex with occasional clients was 65% versus 90%, P < 0001; with repeat clients 53% versus 66%, P < 0.001; and with regular partners 7% versus 30%, P < 0.001. STI prevalence declined from baseline to follow-up: syphilis 25% versus 12%, P < 0.001; trichomonas infection 33% versus 14%, P < 0.001; chlamydial infection 11% versus 5%, P = 0.001; gonorrhoea 5% versus 2%, P = 0.03. HIV prevalence remained stable (26% versus 24%), and detuned assay testing suggested a decline in recent HIV infections.

CONCLUSION

This comprehensive HIV preventive intervention empowering FSW has resulted in striking increases in reported condom use and a concomitant reduction in the prevalence of curable STI. This model should be replicated in similar urban settings across India.

Using mathematical modelling to investigate the plausibility of attributing observed antenatal clinic declines to a female sex worker intervention in Karnataka state, India

OBJECTIVES

To determine whether the 32% and 52% decline in ANC HIV prevalence among female antenatal clinic (ANC) attenders, observed in Avahan districts between 2004 and 2006, and 2007 respectively, in the state of Karnataka could be due to a HIV preventive intervention targeted at female sex workers and their clients.

METHODS

An exhaustive sensitivity analysis, based on an age and parity structured mathematical model of HIV transmission in a general and ANC population, was undertaken to estimate intervention impact in different concentrated HIV epidemics representative of those in Karnataka districts. To assess if the large reduction in ANC HIV prevalence could be solely due to the intervention, we simulated a very optimistic intervention.

RESULTS

If 100% of FSWs were reached and condom use between clients and FSWs increased instantaneously to over 80% of sex acts, the expected intervention decline (50th, (10th, 90th) percentiles) among the overall and 15-19 year old ANC population after three years of intense intervention activity was 21% (14%, 27%) and 27% (19%, 35%); with a predicted time required to produce a 30% intervention decline being approximately 5 (4.0, 6.4) and approximately 3.6 (2.8, 4.8) years, respectively. To achieve this magnitude of decline, the client and FSW HIV prevalence needed to decrease by 33% (28%, 38%) and 44% (38%, 50%), respectively, after three years.

CONCLUSION

Despite the optimistic prevention parameters assumed, our results suggest that the large observed changes in ANC HIV prevalence are very unlikely to already be entirely caused by the FSW targeted intervention. Interpretation of HIV trends in ANC populations should involve triangulation of observed biological and behavioural trends in high-risk groups, modeling studies and documentation of possible sources of bias.

Impact of an intensive HIV prevention programme for female sex workers on HIV prevalence among antenatal clinic attenders in Karnataka state, south India: an ecological analysis

OBJECTIVES

To examine the impact of an intensive HIV preventive intervention (IPI) among female sex workers (FSW) on community HIV transmission, as represented by HIV prevalence among young antenatal clinic (ANC) attenders in Karnataka state, south India.

METHODS

The IPI was initiated in 18 of the 27 districts in Karnataka in 2003, and was generally at scale by mid-2005, covering over 80% of the urban FSW population. We examined trends over time in HIV prevalence from annual HIV surveillance conducted among ANC attenders in Karnataka under the age of 25 years from 2003 to 2007, comparing the IPI with the other districts.

RESULTS

Overall, HIV prevalence among ANC attenders under 25 years of age declined from 1.40% to 0.77%. In a multivariate model, the decline in HIV prevalence in the IPI districts compared to the other districts was statistically significant (P = 0.01), with an adjusted annual odds ratio of 0.88 (95% CI 0.79-0.97). The decline in standardized HIV prevalence in the IPI districts over the period was 56%, compared to 5% in the non-IPI districts.

CONCLUSIONS

Although this analysis is limited by lack of precise comparative data on intervention coverage and intensity, it supports the notion that scaled-up, intensive, targeted HIV preventive interventions among high-risk groups can have a measurable and relatively rapid impact on HIV transmission in the general population, particularly young sexually active populations as represented by ANC attenders. Such focused intervention programmes should be rapidly taken to scale in all HIV epidemics, and especially in concentrated epidemics such as in India.