Leveraging Routinely Collected Program Data to Inform Extrapolated Size Estimates for Key Populations in Namibia: Small Area Estimation Study

BACKGROUND

Estimating the size of key populations, including female sex workers (FSW) and men who have sex with men (MSM), can inform planning and resource allocation for HIV programs at local and national levels. In geographic areas where direct population size estimates (PSEs) for key populations have not been collected, small area estimation (SAE) can help fill in gaps using supplemental data sources known as auxiliary data. However, routinely collected program data have not historically been used as auxiliary data to generate subnational estimates for key populations, including in Namibia.

OBJECTIVE

To systematically generate regional size estimates for FSW and MSM in Namibia, we used a consensus-informed estimation approach with local stakeholders that included the integration of routinely collected HIV program data provided by key populations' HIV service providers.

METHODS

We used quarterly program data reported by key population implementing partners, including counts of the number of individuals accessing HIV services over time, to weight existing PSEs collected through bio-behavioral surveys using a Bayesian triangulation approach. SAEs were generated through simple imputation, stratified imputation, and multivariable Poisson regression models. We selected final estimates using an iterative qualitative ranking process with local key population implementing partners.

RESULTS

Extrapolated national estimates for FSW ranged from 4777 to 13,148 across Namibia, comprising 1.5% to 3.6% of female individuals aged between 15 and 49 years. For MSM, estimates ranged from 4611 to 10,171, comprising 0.7% to 1.5% of male individuals aged between 15 and 49 years. After the inclusion of program data as priors, the estimated proportion of FSW derived from simple imputation increased from 1.9% to 2.8%, and the proportion of MSM decreased from 1.5% to 0.75%. When stratified imputation was implemented using HIV prevalence to inform strata, the inclusion of program data increased the proportion of FSW from 2.6% to 4.0% in regions with high prevalence and decreased the proportion from 1.4% to 1.2% in regions with low prevalence. When population density was used to inform strata, the inclusion of program data also increased the proportion of FSW in high-density regions (from 1.1% to 3.4%) and decreased the proportion of MSM in all regions.

CONCLUSIONS

Using SAE approaches, we combined epidemiologic and program data to generate subnational size estimates for key populations in Namibia. Overall, estimates were highly sensitive to the inclusion of program data. Program data represent a supplemental source of information that can be used to align PSEs with real-world HIV programs, particularly in regions where population-based data collection methods are challenging to implement. Future work is needed to determine how best to include and validate program data in target settings and in key population size estimation studies, ultimately bridging research with practice to support a more comprehensive HIV response.

Triangulating Truth and Reaching Consensus on Population Size, Prevalence, and More: Modeling Study

BACKGROUND

Population size, prevalence, and incidence are essential metrics that influence public health programming and policy. However, stakeholders are frequently tasked with setting performance targets, reporting global indicators, and designing policies based on multiple (often incongruous) estimates of these variables, and they often do so in the absence of a formal, transparent framework for reaching a consensus estimate.

OBJECTIVE

This study aims to describe a model to synthesize multiple study estimates while incorporating stakeholder knowledge, introduce an R Shiny app to implement the model, and demonstrate the model and app using real data.

METHODS

In this study, we developed a Bayesian hierarchical model to synthesize multiple study estimates that allow the user to incorporate the quality of each estimate as a confidence score. The model was implemented as a user-friendly R Shiny app aimed at practitioners of population size estimation. The underlying Bayesian model was programmed in Stan for efficient sampling and computation.

RESULTS

The app was demonstrated using biobehavioral survey-based population size estimates (and accompanying confidence scores) of female sex workers and men who have sex with men from 3 survey locations in a country in sub-Saharan Africa. The consensus results incorporating confidence scores are compared with the case where they are absent, and the results with confidence scores are shown to perform better according to an app-supplied metric for unaccounted-for variation.

CONCLUSIONS

The utility of the triangulator model, including the incorporation of confidence scores, as a user-friendly app is demonstrated using a use case example. Our results offer empirical evidence of the model's effectiveness in producing an accurate consensus estimate and emphasize the significant impact that the accessible model and app offer for public health. It offers a solution to the long-standing problem of synthesizing multiple estimates, potentially leading to more informed and evidence-based decision-making processes. The Triangulator has broad utility and flexibility to be adapted and used in various other contexts and regions to address similar challenges.

Estimation of the population size of gay, bisexual and other men who have sex with men in Canada, 2020

Background

Gay or bisexual (GB) and other men who have sex with men (MSM) are disproportionately affected by human immunodeficiency virus (HIV) globally and domestically in Canada. Reliable and recent population size estimates are necessary to allocate resources to meet prevention needs and for modelling the HIV epidemic. However, previous direct estimates did not account for GB men who would not reveal their sexual identity to a government survey, nor MSM not identifying as GB. The objective of this study was to develop two national population size estimates of gay, bisexual and other men who have sex with men (gbMSM) in 2020. First, GB men based on identity, regardless of sexual experience, and MSM who do not identify as GB but reported anal sex with a man in the past 1-5 years ("Identity-or-Behaviour" estimate). Second, an estimate of gbMSM who reported past 6-12 months anal sex with a man ("Behaviour-only" estimate).

Methods

Estimates for males aged 15 years and older were drawn from Statistics Canada's population size estimates, the Canadian Community Health Survey and the Community-Based Research Centre's Sex Now Survey. Estimated proportions of GB identity, those not likely to disclose GB identity and MSM who do not identify as GB but who reported past 1-5 years anal sex were applied. Past 6-12 months anal sex history was subsequently used to limit estimates to those sexually active anally.

Results

It was estimated that 3.5% of the male population in Canada aged 15 years and older identified as GB. Of GB males, 86.5% were likely to disclose their sexual identity to a government survey. A further 0.1% of non-GB identified males reported past year anal sex with a man. The national Identity-or-Behaviour gbMSM population size in 2020 was estimated at 669,613 people, equivalent to 4.3% of the Canadian male population aged 15 years and older. The estimate of Behaviour-only gbMSM was 412,186, representing 2.6% of the Canadian male population aged 15 years and older.

Conclusion

Using data from multiple sources, a model applied to estimate the population size of gbMSM, accounting for populations previously not included in prior estimates, has been described.

Prevalence of opioid dependence in New South Wales, Australia, 2014-16: Indirect estimation from multiple data sources using a Bayesian approach

AIMS

To estimate the prevalence of, and number of unobserved people with opioid dependence by sex and age group in New South Wales (NSW), Australia.

DESIGN

We applied a Bayesian statistical modelling approach to opioid agonist treatment records linked to adverse event rate data. We estimated prevalence from three types of adverse event separately: opioid mortality, opioid-poisoning hospitalizations and opioid-related charges. We extended the model and produced prevalence estimates from a 'multi-source' model based on all three types of adverse event data.

SETTING, PARTICIPANTS AND MEASUREMENTS

This study was conducted in NSW, Australia, 2014-16 using data from the Opioid Agonist Treatment and Safety (OATS) study, which included all people who had received treatment for opioid dependence in NSW. Aggregate data were obtained on numbers of adverse events in NSW. Rates of each adverse event type within the OATS cohort were modelled. Population data were provided by State and Commonwealth agencies.

FINDINGS

Prevalence of opioid dependence among those aged 15-64 years in 2016 was estimated to be 0.96% (95% credible interval [CrI] = 0.82%, 1.12%) from the mortality model, 0.75% (95% CrI = 0.70%, 0.83%) from hospitalizations, 0.95% (95% CrI = 0.90%, 0.99%) from charges and 0.92% (95% CrI = 0.88%, 0.96%) from the multi-source model. Of the estimated 46 460 (95% CrI = 44 680, 48 410) people with opioid dependence in 2016 from the multi-source model, approximately one-third (16 750, 95% CrI = 14 960, 18 690) had no record of opioid agonist treatment within the last 4 years. From the multi-source model, prevalence in 2016 was estimated to be 1.24% (95% CrI = 1.18%, 1.31%) in men aged 15-44, 1.22% (95% CrI = 1.14%, 1.31%) in men 45-64, 0.63% (95% CrI = 0.59%, 0.68%) in women aged 15-44 and 0.56% (95% CrI = 0.50%, 0.63%) in women aged 45-64.

CONCLUSIONS

A Bayesian statistical approach to estimate prevalence from multiple adverse event types simultaneously calculates that the estimated prevalence of opioid dependence in NSW, Australia in 2016 was 0.92%, higher than previous estimates.

On the equivalence of one-inflated zero-truncated and zero-truncated one-inflated count data likelihoods

One-inflation in zero-truncated count data has recently found considerable attention. There are currently two views in the literature. In the first approach, the untruncated model is considered as one-inflated whereas in the second approach the truncated model is viewed as one-inflated. Here, we show that both models have identical model spaces as well as identical maximum likelihoods. Consequences of population size estimation are illuminated, and the findings are illustrated at hand of two case studies.

Using Google Trends to Inform the Population Size Estimation and Spatial Distribution of Gay, Bisexual, and Other Men Who Have Sex With Men: Proof-of-concept Study

Background

We must triangulate data sources to understand best the spatial distribution and population size of marginalized populations to empower public health leaders to address population-specific needs. Existing population size estimation techniques are difficult and limited.

Objective

We sought to identify a passive surveillance strategy that utilizes internet and social media to enhance, validate, and triangulate population size estimates of gay, bisexual, and other men who have sex with men (gbMSM).

Methods

We explored the Google Trends platform to approximate an estimate of the spatial heterogeneity of the population distribution of gbMSM. This was done by comparing the prevalence of the search term “gay porn” with that of the search term “porn.”

Results

Our results suggested that most cities have a gbMSM population size between 2% and 4% of their total population, with large urban centers having higher estimates relative to rural or suburban areas. This represents nearly a double up of population size estimates compared to that found by other methods, which typically find that between 1% and 2% of the total population are gbMSM. We noted that our method was limited by unequal coverage in internet usage across Canada and differences in the frequency of porn use by gender and sexual orientation.

Conclusions

We argue that Google Trends estimates may provide, for many public health planning purposes, adequate city-level estimates of gbMSM population size in regions with a high prevalence of internet access and for purposes in which a precise or narrow estimate of the population size is not required. Furthermore, the Google Trends platform does so in less than a minute at no cost, making it extremely timely and cost-effective relative to more precise (and complex) estimates. We also discuss future steps for further validation of this approach.

Noninvasive Genetic Assessment Is an Effective Wildlife Research Tool When Compared with Other Approaches

Wildlife research has been indispensable for increasing our insight into ecosystem functioning as well as for designing effective conservation measures under the currently high rates of biodiversity loss. Genetic and genomic analyses might be able to yield the same information on, e.g., population size, health, or diet composition as other wildlife research methods, and even provide additional data that would not be possible to obtain by alternative means. Moreover, if DNA is collected non-invasively, this technique has only minimal or no impact on animal welfare. Nevertheless, the implementation rate of noninvasive genetic assessment in wildlife studies has been rather low. This might be caused by the perceived inefficiency of DNA material obtained non-invasively in comparison with DNA obtained from blood or tissues, or poorer performance in comparison with other approaches used in wildlife research. Therefore, the aim of this review was to evaluate the performance of noninvasive genetic assessment in comparison with other methods across different types of wildlife studies. Through a search of three scientific databases, 113 relevant studies were identified, published between the years 1997 and 2020. Overall, most of the studies (94%) reported equivalent or superior performance of noninvasive genetic assessment when compared with either invasive genetic sampling or another research method. It might be also cheaper and more time-efficient than other techniques. In conclusion, noninvasive genetic assessment is a highly effective research approach, whose efficacy and performance are likely to improve even further in the future with the development of optimized protocols.

Female sex workers population size estimation in Rwanda using a three-source capture-recapture method

Establishing accurate population size estimates (PSE) is important for prioritising and planning provision of services. Multiple source capture−recapture sampling method increases PSE accuracy and reliability. In August 2018, the three-source capture−recapture (3S-CRC) method was employed with a stringent assumption of sample independence to estimate the number of female sex workers (FSW) in Rwanda. Using Rwanda 2017 FSW hotspots mapping data, street and venue-based FSW were sampled at the sector level of each province and tagged with two unique gifts. Each capture was completed within one week to minimise FSW migration between provinces and recall bias. The three captures had 1042, 1204 and 1488 FSW. There were 111 FSW recaptured between captures 1 and 2; 237 between captures 2 and 3; 203 between captures 1 and 3 and 46 captured in all three. The PSE for street and venue-based FSW in Rwanda lies within 95% credible set: 8328–22 806 with corresponding median of 13 716 FSW. The 3S-CRC technique was low-cost and relatively easy to use for PSE in hard-to-reach populations. This estimate provides the basis for determining the denominators to assess HIV programme performance towards FSW and epidemic control and warrants further PSE for home- and cyber-based FSW in Rwanda.

Population Size Estimation Methods: Searching for the Holy Grail

Accurate size estimates of key populations (eg, sex workers, people who inject drugs, transgender people, and men who have sex with men) can help to ensure adequate availability of services to prevent or treat HIV infection; inform HIV response planning, target setting, and resource allocation; and provide data for monitoring and evaluating program outcomes and impact. A gold standard method for population size estimation does not exist, but quality of estimates could be improved by using empirical methods, multiple data sources, and sound statistical concepts. To highlight such methods, a special collection of papers in JMIR Public Health and Surveillance has been released under the title “Key Population Size Estimations.” We provide a summary of these papers to highlight advances in the use of empirical methods and call attention to persistent gaps in information.

Software Application Profile: The Anchored Multiplier calculator-a Bayesian tool to synthesize population size estimates

Estimating the number of people in hidden populations is needed for public health research, yet available methods produce highly variable and uncertain results. The Anchored Multiplier calculator uses a Bayesian framework to synthesize multiple population size estimates to generate a consensus estimate. Users submit point estimates and lower/upper bounds which are converted to beta probability distributions and combined to form a single posterior probability distribution. The Anchored Multiplier calculator is available as a web browser-based application. The software allows for unlimited empirical population size estimates to be submitted and combined according to Bayes Theorem to form a single estimate. The software returns output as a forest plot (to visually compare data inputs and the final Anchored Multiplier estimate) and a table that displays results as population percentages and counts. The web application 'Anchored Multiplier Calculator' is free software and is available at [http://globalhealthsciences.ucsf.edu/resources/tools] or directly at [http://anchoredmultiplier.ucsf.edu/].

Assessing Bias in Population Size Estimates Among Hidden Populations When Using the Service Multiplier Method Combined With Respondent-Driven Sampling Surveys: Survey Study

BACKGROUND

Population size estimates (PSEs) for hidden populations at increased risk of HIV, including female sex workers (FSWs), are important to inform public health policy and resource allocation. The service multiplier method (SMM) is commonly used to estimate the sizes of hidden populations. We used this method to obtain PSEs for FSWs at 9 sites in Zimbabwe and explored methods for assessing potential biases that could arise in using this approach.

OBJECTIVE

This study aimed to guide the assessment of biases that arise when estimating the population sizes of hidden populations using the SMM combined with respondent-driven sampling (RDS) surveys.

METHODS

We conducted RDS surveys at 9 sites in late 2013, where the Sisters with a Voice program (the program), which collects program visit data of FSWs, was also present. Using the SMM, we obtained PSEs for FSWs at each site by dividing the number of FSWs who attended the program, based on program records, by the RDS-II weighted proportion of FSWs who reported attending this program in the previous 6 months in the RDS surveys. Both the RDS weighting and SMM make a number of assumptions, potentially leading to biases if the assumptions are not met. To test these assumptions, we used convergence and bottleneck plots to assess seed dependence of RDS-II proportion estimates, chi-square tests to assess if there was an association between the characteristics of FSWs and their knowledge of program existence, and logistic regression to compare the characteristics of FSWs attending the program with those recruited to RDS surveys.

RESULTS

The PSEs ranged from 194 (95% CI 62-325) to 805 (95% CI 456-1142) across 9 sites from May to November 2013. The 95% CIs for the majority of sites were wide. In some sites, the RDS-II proportion of women who reported program use in the RDS surveys may have been influenced by the characteristics of selected seeds, and we also observed bottlenecks in some sites. There was no evidence of association between characteristics of FSWs and knowledge of program existence, and in the majority of sites, there was no evidence that the characteristics of the populations differed between RDS and program data.

CONCLUSIONS

We used a series of rigorous methods to explore potential biases in our PSEs. We were able to identify the biases and their potential direction, but we could not determine the ultimate direction of these biases in our PSEs. We have evidence that the PSEs in most sites may be biased and a suggestion that the bias is toward underestimation, and this should be considered if the PSEs are to be used. These tests for bias should be included when undertaking population size estimation using the SMM combined with RDS surveys.

Upper bound estimators of the population size based on ordinal models for capture-recapture experiments

Capture-recapture studies have attracted a lot of attention over the past few decades, especially in applied disciplines where a direct estimate for the size of a population of interest is not available. Epidemiology, ecology, public health, and biodiversity are just a few examples. The estimation of the number of unseen units has been a challenge for theoretical statisticians, and considerable progress has been made in providing lower bound estimators for the population size. In fact, it is well known that consistent estimators for this cannot be provided in the very general case. Considering a case where capture-recapture studies are summarized by a frequency of frequencies distribution, we derive a simple upper bound of the population size based on the cumulative distribution function. We introduce two estimators of this bound, without any specific parametric assumption on the distribution of the observed frequency counts. The behavior of the proposed estimators is investigated using several benchmark datasets and a large-scale simulation experiment based on the scheme discussed by Pledger.

Estimation of the Population Size of Men Who Have Sex With Men in Vietnam: Social App Multiplier Method

Background

Although the prevalence of HIV among men who have sex with men (MSM) in Vietnam has been increasing in recent years, there are no estimates of the population size of MSM based on tested empirical methods.

Objective

This study aimed to estimate the size of the MSM population in 12 provinces in Vietnam and extrapolate from those areas to generate a national population estimate of MSM. A secondary aim of this study was to compare the feasibility of obtaining the number of users of a mobile social (chat and dating) app for MSM using 3 different approaches.

Methods

This study used the social app multiplier method to estimate the size of MSM populations in 12 provinces using the count of users on a social app popular with MSM in Vietnam as the first data source and a questionnaire propagated through the MSM community using respondent-driven sampling as the second data source. A national estimation of the MSM population is extrapolated from the results in the study provinces, and the percentage of MSM reachable through online social networks is clarified.

Results

The highest MSM population size among the 12 provinces is estimated in Hanoi and the lowest is estimated in Binh Dinh. On average, 37% of MSM in the provinces surveyed had used the social app Jack’d in the last 30 days (95% CI 27-48). Extrapolation of the results from the study provinces with reliable estimations results in an estimated national population of 178,000 MSM (95% CI 122,000-512,000) aged 15 to 49 years in Vietnam. The percentage of MSM among adult males aged 15 to 49 years in Vietnam is 0.68% (95% CI 0.46-1.95).

Conclusions

This study is the first attempt to empirically estimate the population of MSM in Vietnam and highlights the feasibility of reaching a large proportion of MSM through a social app. The estimation reported in this study is within the bounds suggested by the Joint United Nations Programme on HIV/AIDS. This study provides valuable information on MSM population sizes in provinces where reliable estimates were obtained, which they can begin to work with in program planning and resource allocation.

Kuantim mi tu ("Count me too"): Using Multiple Methods to Estimate the Number of Female Sex Workers, Men Who Have Sex With Men, and Transgender Women in Papua New Guinea in 2016 and 2017

Background

Female sex workers (FSW), men who have sex with men (MSM), and transgender women (TGW) are at high risk of acquiring HIV in many settings, such as Papua New Guinea (PNG). An understanding of the approximate size of these populations can inform resource allocation for HIV services for FSW, MSM, and TGW.

Objective

An objective of this multi-site survey was to conduct updated population size estimations (PSE) of FSW and MSM/TGW.

Methods

Respondent-driven sampling (RDS) biobehavioral surveys of FSW and MSM/TGW were conducted in 3 major cities—(1) Port Moresby, (2) Lae, and (3) Mount Hagen—between June 2016 and December 2017. Eligibility criteria for FSW included: (1) ≥12 years of age, (2) born female, (3) could speak English or Tok Pisin (PNG Pidgin), and (4) had sold or exchanged sex with a man in the past six months. Eligibility for MSM/TGW included: (1) ≥12 years of age, (2) born male, (3) could speak English, or Tok Pisin, and (4) had engaged in oral or anal sex with another person born male in the past six months. PSE methods included unique object multiplier, service multiplier, and successive sampling-population size estimation (SS-PSE) using imputed visibility. Weighted data analyses were conducted using RDS-Analyst and Microsoft Excel.

Results

Sample sizes for FSW and MSM/TGW in Port Moresby, Lae, and Mount Hagen included: (1) 673 and 400, (2) 709 and 352, and (3) 709 and 111 respectively. Keychains were used for the unique object multiplier method and were distributed 1 week before the start of each RDS survey. HIV service testing data were only available in Port Moresby and Mount Hagen and SS-PSE estimates were calculated for all cities. Due to limited service provider data and uncertain prior size estimation knowledge, unique object multiplier weighted estimations were chosen for estimates. In Port Moresby, we estimate that there are 16,053 (95% CI 8232-23,874) FSW and 7487 (95% CI 3975-11,000) MSM/TGW, approximately 9.5% and 3.8% of the female and male populations respectively. In Lae, we estimate that there are 6105 (95% CI 4459-7752) FSW and 4669 (95% CI 3068-6271) MSM/TGW, approximately 14.4% and 10.1% of the female and male populations respectively. In Mount Hagen, we estimate that there are 2646 (95% CI 1655-3638) FSW and 1095 (95% CI 913-1151) MSM/TGW using service multiplier and successive sampling, respectively. This is approximately 17.1% and 6.3% of the female and male populations respectively.

Conclusions

As the HIV epidemic in PNG rapidly evolves among key populations, PSE should be repeated to produce current estimates for timely comparison and future trend analysis.

Population Size Estimations Among Hidden Populations Using Respondent-Driven Sampling Surveys: Case Studies From Armenia

Background

Estimates of the sizes of hidden populations, including female sex workers (FSW), men who have sex with men (MSM), and people who inject drugs (PWID), are essential for understanding the magnitude of vulnerabilities, health care needs, risk behaviors, and HIV and other infections.

Objective

This article advances the successive sampling-population size estimation (SS-PSE) method by examining the performance of a modification allowing visibility to be jointly modeled with population size in the context of 15 datasets. Datasets are from respondent-driven sampling (RDS) surveys of FSW, MSM, and PWID from three cities in Armenia. We compare and evaluate the accuracy of our imputed visibility population size estimates to those found for the same populations through other unpublished methods. We then suggest questions that are useful for eliciting information needed to compute SS-PSE and provide guidelines and caveats to improve the implementation of SS-PSE for real data.

Methods

SS-PSE approximates the RDS sampling mechanism via the successive sampling model and uses the order of selection of the sample to provide information on the distribution of network sizes over the population members. We incorporate visibility imputation, a measure of a person’s propensity to participate in the study, given that inclusion probabilities for RDS are unknown and social network sizes, often used as a proxy for inclusion probability, are subject to measurement errors from self-reported study data.

Results

FSW in Yerevan (2012, 2016) and Vanadzor (2016) as well as PWID in Yerevan (2014), Gyumri (2016), and Vanadzor (2016) had great fits with prior estimations. The MSM populations in all three cities had inconsistencies with expert prior values. The maximum low prior value was larger than the minimum high prior value, making a great fit impossible. One possible explanation is the inclusion of transgender individuals in the MSM populations during these studies. There could be differences between what experts perceive as the size of the population, based on who is an eligible member of that population, and what members of the population perceive. There could also be inconsistencies among different study participants, as some may include transgender individuals in their accounting of personal network size, while others may not. Because of these difficulties, the transgender population was split apart from the MSM population for the 2018 study.

Conclusions

Prior estimations from expert opinions may not always be accurate. RDS surveys should be assessed to ensure that they have met all of the assumptions, that variables have reached convergence, and that the network structure of the population does not have bottlenecks. We recommend that SS-PSE be used in conjunction with other population size estimations commonly used in RDS, as well as results of other years of SS-PSE, to ensure generation of the most accurate size estimation.

Population Size Estimation of Venue-Based Female Sex Workers in Ho Chi Minh City, Vietnam: Capture-Recapture Exercise

Background

There is limited population size estimation of female sex workers (FSWs) in Ho Chi Minh City (HCMC)—the largest city in Vietnam. Only 1 population size estimation among venue-based female sex workers (VFSWs) was conducted in 2012 in HCMC. Appropriate estimates of the sizes of key populations are critical for resource allocation to prevent HIV infection.

Objective

The aim of this study was to estimate the population size of the VFSWs from December 2016 to January 2017 in HCMC, Vietnam.

Methods

A multistage capture-recapture study was conducted in HCMC. The capture procedures included selection of districts using stratified probability proportion to size, mapping to identify venues, approaching all VFSWs to screen their eligibility, and then distribution of a unique object (a small pink makeup bag) to all eligible VFSWs in all identified venues. The recapture exercise included equal probability random selection of a sample of venues from the initial mapping and then approaching FSWs in those venues to determine the number and proportion of women who received the unique object. The proportion and associated confidence bounds, calculated using sampling weights and accounting for study design, were then divided by the number of objects distributed to calculate the number of VFSWs in the selected districts. This was then multiplied by the inverse of the proportion of districts selected to calculate the number of VFSWs in HCMC as a whole.

Results

Out of 24 districts, 6 were selected for the study. Mapping identified 573 venues across which 2317 unique objects were distributed in the first capture. During the recapture round, 103 venues were selected and 645 VFSWs were approached and interviewed. Of those, 570 VFSWs reported receiving the unique object during the capture round. Total estimated VFSWs in the 6 selected districts were 2616 (95% CI 2445-3014), accounting for the fact that only 25% (6/24) of total districts were selected gives an overall estimate of 10,465 (95% CI 9782-12,055) VFSWs in HCMC.

Conclusions

The capture-recapture exercise provided an estimated number of VFSWs in HCMC. However, for planning HIV prevention and care service needs among all FSWs, studies are needed to assess the number of sex workers who are not venue-based, including those who use social media platforms to sell services.

Estimating the Population Size of Female Sex Workers in Three South African Cities: Results and Recommendations From the 2013-2014 South Africa Health Monitoring Survey and Stakeholder Consensus

Background

Robust population size estimates of female sex workers and other key populations in South Africa face multiple methodological limitations, including inconsistencies in surveillance and programmatic indicators. This has, consequently, challenged the appropriate allocation of resources and benchmark-setting necessary to an effective HIV response. A 2013-2014 integrated biological and behavioral surveillance (IBBS) survey from South Africa showed alarmingly high HIV prevalence among female sex workers in South Africa’s three largest cities of Johannesburg (71.8%), Cape Town (39.7%), and eThekwini (53.5%). The survey also included several multiplier-based population size estimation methods.

Objective

The objective of our study was to present the selected population size estimation methods used in an IBBS survey and the subsequent participatory process used to estimate the number of female sex workers in three South African cities.

Methods

In 2013-2014, we used respondent-driven sampling to recruit independent samples of female sex workers for IBBS surveys in Johannesburg, Cape Town, and eThekwini. We embedded multiple multiplier-based population size estimation methods into the survey, from which investigators calculated weighted estimates and ranges of population size estimates for each city’s female sex worker population. Following data analysis, investigators consulted civil society stakeholders to present survey results and size estimates and facilitated stakeholder vetting of individual estimates to arrive at consensus point estimates with upper and lower plausibility bounds.

Results

In total, 764, 650, and 766 female sex workers participated in the survey in Johannesburg, Cape Town, and eThekwini, respectively. For size estimation, investigators calculated preliminary point estimates as the median of the multiple estimation methods embedded in the IBBS survey and presented these to a civil society-convened stakeholder group. Stakeholders vetted all estimates in light of other data points, including programmatic experience, ensuring inclusion only of plausible point estimates in median calculation. After vetting, stakeholders adopted three consensus point estimates with plausible ranges: Johannesburg 7697 (5000-10,895); Cape Town 6500 (4579-9000); eThekwini 9323 (4000-10,000).

Conclusions

Using several population size estimates methods embedded in an IBBS survey and a participatory stakeholder consensus process, the South Africa Health Monitoring Survey produced female sex worker size estimates representing approximately 0.48%, 0.49%, and 0.77% of the adult female population in Johannesburg, Cape Town, and eThekwini, respectively. In data-sparse environments, stakeholder engagement and consensus is critical to vetting of multiple empirically based size estimates procedures to ensure adoption and utilization of data-informed size estimates for coordinated national and subnational benchmarking. It also has the potential to increase coherence in national and key population-specific HIV responses and to decrease the likelihood of duplicative and wasteful resource allocation. We recommend building cooperative and productive academic-civil society partnerships around estimates and other strategic information dissemination and sharing to facilitate the incorporation of additional data as it becomes available, as these additional data points may minimize the impact of the known and unknown biases inherent in any single, investigator-calculated method.

Underreporting in HIV-related high-risk behaviors: comparing the results of multiple data collection methods in a behavioral survey of prisoners in Iran

We explored the potentials of using three indirect methods including crosswise, proxy respondent method, and network scale-up (NSU) in comparison to direct questioning in collecting sensitive and socially stigmatized HIV-related risk behaviors information from prisoners (N=265). Participants reported more sexual contact in prison for their friends than they did for themselves (10.6% vs. 3.8% in men, 13.7% vs. 0% in women). In men, NSU provided lower estimates than direct questioning, while in women NSU estimates were higher. Different data collection methods provide different estimates, and collectively offer a more comprehensive picture of HIV-related risk behaviors in prisons.

Quantity Versus Quality: A Survey Experiment to Improve the Network Scale-up Method

The network scale-up method is a promising technique that uses sampled social network data to estimate the sizes of epidemiologically important hidden populations, such as sex workers and people who inject illicit drugs. Although previous scale-up research has focused exclusively on networks of acquaintances, we show that the type of personal network about which survey respondents are asked to report is a potentially crucial parameter that researchers are free to vary. This generalization leads to a method that is more flexible and potentially more accurate. In 2011, we conducted a large, nationally representative survey experiment in Rwanda that randomized respondents to report about one of 2 different personal networks. Our results showed that asking respondents for less information can, somewhat surprisingly, produce more accurate size estimates. We also estimated the sizes of 4 key populations at risk for human immunodeficiency virus infection in Rwanda. Our estimates were higher than earlier estimates from Rwanda but lower than international benchmarks. Finally, in this article we develop a new sensitivity analysis framework and use it to assess the possible biases in our estimates. Our design can be customized and extended for other settings, enabling researchers to continue to improve the network scale-up method.

Estimation of the size of the female sex worker population in Rwanda using three different methods

HIV prevalence is disproportionately high among female sex workers compared to the general population. Many African countries lack useful data on the size of female sex worker populations to inform national HIV programmes. A female sex worker size estimation exercise using three different venue-based methodologies was conducted among female sex workers in all provinces of Rwanda in August 2010. The female sex worker national population size was estimated using capture–recapture and enumeration methods, and the multiplier method was used to estimate the size of the female sex worker population in Kigali. A structured questionnaire was also used to supplement the data. The estimated number of female sex workers by the capture–recapture method was 3205 (95% confidence interval: 2998–3412). The female sex worker size was estimated at 3348 using the enumeration method. In Kigali, the female sex worker size was estimated at 2253 (95% confidence interval: 1916–2524) using the multiplier method. Nearly 80% of all female sex workers in Rwanda were found to be based in the capital, Kigali. This study provided a first-time estimate of the female sex worker population size in Rwanda using capture–recapture, enumeration, and multiplier methods. The capture–recapture and enumeration methods provided similar estimates of female sex worker in Rwanda. Combination of such size estimation methods is feasible and productive in low-resource settings and should be considered vital to inform national HIV programmes.

Estimation of capture probabilities using generalized estimating equations and mixed effects approaches

Modeling individual heterogeneity in capture probabilities has been one of the most challenging tasks in capture–recapture studies. Heterogeneity in capture probabilities can be modeled as a function of individual covariates, but correlation structure among capture occasions should be taking into account. A proposed generalized estimating equations (GEE) and generalized linear mixed modeling (GLMM) approaches can be used to estimate capture probabilities and population size for capture–recapture closed population models. An example is used for an illustrative application and for comparison with currently used methodology. A simulation study is also conducted to show the performance of the estimation procedures. Our simulation results show that the proposed quasi-likelihood based on GEE approach provides lower SE than partial likelihood based on either generalized linear models (GLM) or GLMM approaches for estimating population size in a closed capture–recapture experiment. Estimator performance is good if a large proportion of individuals are captured. For cases where only a small proportion of individuals are captured, the estimates become unstable, but the GEE approach outperforms the other methods.