An Empirical Analysis of the Impact of Recruitment Patterns on RDS Estimates among a Socially Ordered Population of Female Sex Workers in China

The Development and the Assessment of Sampling Methods for Hard-to-Reach Populations in HIV Surveillance

Due to stigma or legal issues, populations with higher HIV risk are often hard to reach, which impedes accurate population estimation of HIV burden. To better sample hard-to-reach populations (HTRPs) for HIV surveillance, various sampling methods have been designed and/or used since HIV epidemic following the first reported AIDS cases in 1981. This paper describes the development and the assessment (i.e., validity and reproducibility) of approximately eight sampling methods (e.g., convenience sampling, snowball sampling, time location sampling, and respondent-driven sampling) for HTRPs in HIV surveillance, with a focus on respondent-driven sampling (RDS). Compared to other methods, RDS has been greatly assessed. However, current evidence is still inadequate for RDS to be considered the best option for sampling HTRPs. The field must continue to assess RDS and to develop new sampling approaches or modifications to existing approaches.

Using Social Networks to Sample Migrants and Study the Complexity of Contemporary Immigration: An Evaluation Study

We test the effectiveness of a link-tracing sampling approach-network sampling with memory (NSM)-to recruit samples of rare immigrant populations with an application among Chinese immigrants in the Raleigh-Durham area of North Carolina. NSM uses the population network revealed by data from the survey to improve the efficiency of link-tracing sampling and has been shown to substantially reduce design effects in simulated sampling. Our goals are to (1) show that it is possible to recruit a probability sample of a locally rare immigrant group using NSM and achieve high response rates; (2) demonstrate the feasibility of the collection and benefits of new forms of network data that transcend kinship networks in existing surveys and can address unresolved questions about the role of social networks in migration decisions, the maintenance of transnationalism, and the process of social incorporation; and (3) test the accuracy of the NSM approach for recruiting immigrant samples by comparison with the American Community Survey. Our results indicate feasibility, high performance, cost-effectiveness, and accuracy of the NSM approach to sample immigrants for studies of local immigrant communities. This approach can also be extended to recruit multisite samples of immigrants at origin and destination.

A Bayesian framework for modelling the preferential selection process in respondent-driven sampling

In sampling designs that utilize peer recruitment, the sampling process is partially unknown and must be modelled to make inference about the population and estimate standard outcomes like prevalence. We develop a Bayesian model for the recruitment process for respondent-driven sampling (RDS), a network sampling methodology used worldwide to sample hidden populations that are not reachable by conventional sampling techniques, including those at high risk for HIV/AIDS. Current models for the RDS sampling process typically assume that recruitment occurs randomly given the population social network, but this is likely untrue in practice. To model preferential selection on covariates, we develop a sequential two-sided rational choice framework, which allows generative probabilistic network models to be created for the RDS sampling process. In the rational choice framework, members of the population make recruitment and participation choices based on observable nodal and dyadic covariates to maximize their utility given constraints. Inference is made about recruitment preferences given the observed recruitment chain in a Bayesian framework by incorporating the latent utilities and sampling from the joint posterior distribution via Markov chain Monte Carlo. We present simulation results and apply the model to an RDS study of Francophone migrants in Rabat, Morocco.

Implementing Respondent-Driven Sampling to Recruit Women Who Exchange Sex in New York City: Factors Associated with Recruitment and Lessons Learned

Respondent-driven sampling (RDS) relies on productive peer recruitment to capture hidden populations. Domestic studies have identified characteristics of productive recruitment among RDS samples of men who have sex with men and persons who use drugs, but not of women who exchange sex, a group vulnerable to HIV infection. We examined sociodemographic-, behavioral-, exchange-sex-, and protocol-related factors associated with recruitment among seeds (n = 25) and peers (n = 297) in the 2016 New York City National HIV Behavioral Surveillance Study cycle focused on women who exchange sex. Recruiter productivity was significantly associated with not having been recently incarcerated, lower rate of HIV testing, and larger exchange sex networks among seeds, and with HIV-prevention services usage among peers. We describe challenges and lessons learned from implementing RDS in this population. Our study identifies seed characteristics and protocol improvements researchers can utilize when implementing future RDS studies among women who exchange sex.

The development of respondent-driven sampling (RDS) inference: A systematic review of the population mean and variance estimates

BACKGROUND

Respondent-driven sampling (RDS) is a successful data collection method used in hard-to-reach populations, like those experiencing or at high risk of drug dependence. Since its introduction in 1997, identifying appropriate methods for estimating population means and sampling variances has been challenging and numerous approaches have been developed for making inferences about these quantities. To guide researchers and practitioners in deciding which approach to use, this article reviews the literature on these methodological developments.

METHODS

A systematic review using four electronic databases was conducted in order to summarize the progress of RDS inference over the last 20 years and to provide insight to researchers on using the appropriate estimators in analyzing RDS data. Two independent reviewers selected the relevant abstracts and articles; thirty-two studies were included. The content of the studies was further categorized into developing and evaluating RDS mean and variance estimators.

RESULTS

The population mean estimator RDSI^EGO and the sampling variance estimators associated with tree boot strapping were identified as promising methods as the most robust population mean and variance estimate, respectively; as these estimators rely on a fewer assumptions.

CONCLUSIONS

RDS holds substantial promise as a sampling method for understanding populations at high risk. The varied approaches to inference with RDS data each rely on different assumptions, but some require fewer assumptions than others and provide more robust and accurate inferences, when their corresponding assumptions are met.

Using Social Networks to Understand and Overcome Implementation Barriers in the Global HIV Response

BACKGROUND

Despite the development of several efficacious HIV prevention and treatment methods in the past 2 decades, HIV continues to spread globally. Uptake of interventions is nonrandomly distributed across populations. Such inequality is socially patterned and reinforced by homophily arising from both social selection (becoming friends with similar people) and influence (becoming similar to friends).

METHODS

We conducted a narrative review to describe how social network analysis methods-including egocentric, sociocentric, and respondent-driven sampling designs-provide tools to measure key populations, to understand how epidemics spread, and to evaluate intervention take-up.

RESULTS

Social network analysis-informed designs can improve intervention effectiveness by reaching otherwise inaccessible populations. They can also improve intervention efficiency by maximizing spillovers, through social ties, to at-risk but susceptible individuals. Social network analysis-informed designs thus have the potential to be both more effective and less unequal in their effects, compared with social network analysis-naïve approaches. Although social network analysis-informed designs are often resource-intensive, we believe they provide unique insights that can help reach those most in need of HIV prevention and treatment interventions.

CONCLUSION

Increased collection of social network data during both research and implementation work would provide important information to improve the roll-out of existing studies in the present and to inform the design of more data-efficient, social network analysis-informed interventions in the future. Doing so will improve the reach of interventions, especially to key populations, and to maximize intervention impact once delivered.

Empirical evidence of recruitment bias in a network study of people who inject drugs

Background: Epidemiologic surveys of people who inject drugs (PWID) can be difficult to conduct because potential participants may fear exposure or legal repercussions. Respondent-driven sampling (RDS) is a procedure in which subjects recruit their eligible social contacts. The statistical validity of RDS surveys of PWID and other risk groups depends on subjects recruiting at random from among their network contacts. Objectives: We sought to develop and apply a rigorous definition and statistical tests for uniform network recruitment in an RDS survey. Methods: We undertook a detailed study of recruitment bias in a unique RDS study of PWID in Hartford, CT, the USA in which the network, individual-level covariates, and social link attributes were recorded. A total of n=527 participants (402 male, 123 female, and two individuals who did not specify their gender) within a network of 2626 PWID were recruited. Results: We found strong evidence of recruitment bias with respect to age, homelessness, and social relationship characteristics. In the discrete model, the estimated hazard ratios regarding the significant features of recruitment time and choice of recruitee were: alter's age 1.03 [1.02, 1.05], alter's crack-using status 0.70 [0.50, 1.00], homelessness difference 0.61 [0.43, 0.87], and sharing activities in drug preparation 2.82 [1.39, 5.72]. Under both the discrete and continuous-time recruitment regression models, we reject the null hypothesis of uniform recruitment. Conclusions: The results provide the evidence that for this study population of PWID, recruitment bias may significantly alter the sample composition, making results of RDS surveys less reliable. More broadly, RDS studies that fail to collect comprehensive network data may not be able to detect biased recruitment when it occurs.

Age bias in survey sampling and implications for estimating HIV prevalence in men who have sex with men: insights from mathematical modelling

Respondent-driven sampling (RDS) is widely used to estimate HIV prevalence in men who have sex with men (MSM). Mathematical models that are calibrated to these data may be compromised if they fail to account for selection biases in RDS surveys. To quantify the potential extent of this bias, an agent-based model of HIV in South Africa was calibrated to HIV prevalence and sexual behaviour data from South African studies of MSM, first reweighting the modelled MSM population to match the younger age profile of the RDS surveys (age-adjusted analysis) and then without reweighting (unadjusted analysis). The model estimated a median HIV prevalence in South African MSM in 2015 of 34.6% (inter-quartile range (IQR): 31.4-37.2%) in the age-adjusted analysis, compared with 26.1% (IQR: 24.1-28.4%) in the unadjusted analysis. The median lifetime risk of acquiring HIV in exclusively homosexual men was 88% (IQR: 82-92%) in the age-adjusted analysis, compared with 76% (IQR: 64-85%) in the unadjusted analysis. These results suggest that RDS studies may under-estimate the exceptionally high HIV prevalence rates in South African MSM because of over-sampling of younger MSM. Mathematical models that are calibrated to these data need to control for likely over-sampling of younger MSM.

Identification of Homophily and Preferential Recruitment in Respondent-Driven Sampling

Respondent-driven sampling (RDS) is a link-tracing procedure used in epidemiologic research on hidden or hard-to-reach populations in which subjects recruit others via their social networks. Estimates from RDS studies may have poor statistical properties due to statistical dependence in sampled subjects' traits. Two distinct mechanisms account for dependence in an RDS study: homophily, the tendency for individuals to share social ties with others exhibiting similar characteristics, and preferential recruitment, in which recruiters do not recruit uniformly at random from their network alters. The different effects of network homophily and preferential recruitment in RDS studies have been a source of confusion and controversy in methodological and empirical research in epidemiology. In this work, we gave formal definitions of homophily and preferential recruitment and showed that neither is identified in typical RDS studies. We derived nonparametric identification regions for homophily and preferential recruitment and showed that these parameters were not identified unless the network took a degenerate form. The results indicated that claims of homophily or recruitment bias measured from empirical RDS studies may not be credible. We applied our identification results to a study involving both a network census and RDS on a population of injection drug users in Hartford, Connecticut (2012-2013).

Social Network Clustering and the Spread of HIV/AIDS Among Persons Who Inject Drugs in 2 Cities in the Philippines

INTRODUCTION

The Philippines has seen rapid increases in HIV prevalence among people who inject drugs. We study 2 neighboring cities where a linked HIV epidemic differed in timing of onset and levels of prevalence. In Cebu, prevalence rose rapidly from below 1% to 54% between 2009 and 2011 and remained high through 2013. In nearby Mandaue, HIV remained below 4% through 2011 then rose rapidly to 38% by 2013.

OBJECTIVES

We hypothesize that infection prevalence differences in these cities may owe to aspects of social network structure, specifically levels of network clustering. Building on previous research, we hypothesize that higher levels of network clustering are associated with greater epidemic potential.

METHODS

Data were collected with respondent-driven sampling among men who inject drugs in Cebu and Mandaue in 2013. We first examine sample composition using estimators for population means. We then apply new estimators of network clustering in respondent-driven sampling data to examine associations with HIV prevalence.

RESULTS

Samples in both cities were comparable in composition by age, education, and injection locations. Dyadic needle-sharing levels were also similar between the 2 cities, but network clustering in the needle-sharing network differed dramatically. We found higher clustering in Cebu than Mandaue, consistent with expectations that higher clustering is associated with faster epidemic spread.

CONCLUSIONS

This article is the first to apply estimators of network clustering to empirical respondent-driven samples, and it offers suggestive evidence that researchers should pay greater attention to network structure's role in HIV transmission dynamics.

NEW SURVEY QUESTIONS AND ESTIMATORS FOR NETWORK CLUSTERING WITH RESPONDENT-DRIVEN SAMPLING DATA

Respondent-driven sampling (RDS) is a popular method for sampling hard-to-survey populations that leverages social network connections through peer recruitment. While RDS is most frequently applied to estimate the prevalence of infections and risk behaviors of interest to public health, such as HIV/AIDS or condom use, it is rarely used to draw inferences about the structural properties of social networks among such populations because it does not typically collect the necessary data. Drawing on recent advances in computer science, we introduce a set of data collection instruments and RDS estimators for network clustering, an important topological property that has been linked to a network's potential for diffusion of information, disease, and health behaviors. We use simulations to explore how these estimators, originally developed for random walk samples of computer networks, perform when applied to RDS samples with characteristics encountered in realistic field settings that depart from random walks. In particular, we explore the effects of multiple seeds, without replacement versus with replacement, branching chains, imperfect response rates, preferential recruitment, and misreporting of ties. We find that clustering coefficient estimators retain desirable properties in RDS samples. This paper takes an important step toward calculating network characteristics using nontraditional sampling methods, and it expands the potential of RDS to tell researchers more about hidden populations and the social factors driving disease prevalence.

HIV Prevalence and Risk Factors Among Male Foreign Migrants in Cape Town, South Africa

While migration has been shown to be a risk factor for HIV, variation in HIV prevalence by subgroups of migrants needs further exploration. This paper documents the HIV prevalence and key characteristics among male foreign migrants in Cape Town, South Africa and the effectiveness of respondent-driven sampling (RDS) to recruit this population. Participants in this cross-sectional study completed a behavioral risk-factor questionnaire and provided a dried blood sample for HIV analysis. Overall HIV prevalence was estimated to be 8.7 % (CI 5.4-11.8) but varied dramatically by country of origin. After adjusting for country of origin, HIV sero-positivity was positively associated with older age (p = 0.001), completing high school (p = 0.025), not having enough money for food (p = 0.036), alcohol use (p = 0.049), and engaging in transactional sex (p = 0.022). RDS was successful in recruiting foreign migrant men. A better understanding of the timing of HIV acquisition is needed to design targeted interventions for migrant men.

A comparison of network sampling designs for a hidden population of drug users: Random walk vs. respondent-driven sampling

Both random walk and respondent-driven sampling (RDS) exploit social networks and may reduce biases introduced by earlier methods for sampling from hidden populations. Although RDS has become much more widely used by social researchers than random walk (RW), there has been little discussion of the tradeoffs in choosing RDS over RW. This paper compares experiences of implementing RW and RDS to recruit drug users to a network-based study in Houston, Texas. Both recruitment methods were implemented over comparable periods of time, with the same population, by the same research staff. RDS methods recruited more participants with less strain on staff. However, participants recruited through RW were more forthcoming than RDS participants in helping to recruit members of their social networks. Findings indicate that, dependent upon study goals, researchers' choice of design may influence participant recruitment, participant commitment, and impact on staff, factors that may in turn affect overall study success.

Estimating uncertainty in respondent-driven sampling using a tree bootstrap method

Respondent-driven sampling (RDS) is a network-based form of chain-referral sampling used to estimate attributes of populations that are difficult to access using standard survey tools. Although it has grown quickly in popularity since its introduction, the statistical properties of RDS estimates remain elusive. In particular, the sampling variability of these estimates has been shown to be much higher than previously acknowledged, and even methods designed to account for RDS result in misleadingly narrow confidence intervals. In this paper, we introduce a tree bootstrap method for estimating uncertainty in RDS estimates based on resampling recruitment trees. We use simulations from known social networks to show that the tree bootstrap method not only outperforms existing methods but also captures the high variability of RDS, even in extreme cases with high design effects. We also apply the method to data from injecting drug users in Ukraine. Unlike other methods, the tree bootstrap depends only on the structure of the sampled recruitment trees, not on the attributes being measured on the respondents, so correlations between attributes can be estimated as well as variability. Our results suggest that it is possible to accurately assess the high level of uncertainty inherent in RDS.

Generalizing the Network Scale-Up Method: A New Estimator for the Size of Hidden Populations

The network scale-up method enables researchers to estimate the size of hidden populations, such as drug injectors and sex workers, using sampled social network data. The basic scale-up estimator offers advantages over other size estimation techniques, but it depends on problematic modeling assumptions. We propose a new generalized scale-up estimator that can be used in settings with non-random social mixing and imperfect awareness about membership in the hidden population. Further, the new estimator can be used when data are collected via complex sample designs and from incomplete sampling frames. However, the generalized scale-up estimator also requires data from two samples: one from the frame population and one from the hidden population. In some situations these data from the hidden population can be collected by adding a small number of questions to already planned studies. For other situations, we develop interpretable adjustment factors that can be applied to the basic scale-up estimator. We conclude with practical recommendations for the design and analysis of future studies.

Sampling Migrants from their Social Networks: The Demography and Social Organization of Chinese Migrants in Dar es Salaam, Tanzania

The streams of Chinese migration to Africa are growing in tandem with rising Chinese investments and trade flows in and to the African continent. In spite of the high profile of this phenomenon in the media, there are few rich and broad descriptions of Chinese communities in Africa. Reasons for this include the rarity of official statistics on foreign-born populations in African censuses, the absence of predefined sampling frames required to draw representative samples with conventional survey methods and difficulties to reach certain segments of this population. Here, we use a novel network-based approach, Network Sampling with Memory, which overcomes the challenges of sampling 'hidden' populations in the absence of a sampling frame, to recruit a sample of recent Chinese immigrants in Dar es Salaam, Tanzania and collect information on the demographic characteristics, migration histories and social ties of members of this sample. These data reveal a heterogeneous Chinese community composed of "state-led" migrants who come to Africa to work on projects undertaken by large Chinese state-owned enterprises and "independent" migrants who come on their own accord to engage in various types of business ventures. They offer a rich description of the demographic profile and social organization of this community, highlight key differences between the two categories of migrants and map the structure of the social ties linking them. We highlight needs for future research on inter-group differences in individual motivations for migration, economic activities, migration outcomes, expectations about future residence in Africa, social integration and relations with local communities.

Statistical design and analysis plan for an impact evaluation of an HIV treatment and prevention intervention for female sex workers in Zimbabwe: a study protocol for a cluster randomised controlled trial

Background

Pragmatic cluster-randomised trials should seek to make unbiased estimates of effect and be reported according to CONSORT principles, and the study population should be representative of the target population. This is challenging when conducting trials amongst ‘hidden’ populations without a sample frame. We describe a pair-matched cluster-randomised trial of a combination HIV-prevention intervention to reduce the proportion of female sex workers (FSW) with a detectable HIV viral load in Zimbabwe, recruiting via respondent driven sampling (RDS).

Methods

We will cross-sectionally survey approximately 200 FSW at baseline and at endline to characterise each of 14 sites. RDS is a variant of chain referral sampling and has been adapted to approximate random sampling. Primary analysis will use the ‘RDS-2’ method to estimate cluster summaries and will adapt Hayes and Moulton’s ‘2-step’ method to adjust effect estimates for individual-level confounders and further adjust for cluster baseline prevalence. We will adapt CONSORT to accommodate RDS. In the absence of observable refusal rates, we will compare the recruitment process between matched pairs. We will need to investigate whether cluster-specific recruitment or the intervention itself affects the accuracy of the RDS estimation process, potentially causing differential biases. To do this, we will calculate RDS-diagnostic statistics for each cluster at each time point and compare these statistics within matched pairs and time points. Sensitivity analyses will assess the impact of potential biases arising from assumptions made by the RDS-2 estimation.

Discussion

We are not aware of any other completed pragmatic cluster RCTs that are recruiting participants using RDS. Our statistical design and analysis approach seeks to transparently document participant recruitment and allow an assessment of the representativeness of the study to the target population, a key aspect of pragmatic trials. The challenges we have faced in the design of this trial are likely to be shared in other contexts aiming to serve the needs of legally and/or socially marginalised populations for which no sampling frame exists and especially when the social networks of participants are both the target of intervention and the means of recruitment.

The trial was registered at Pan African Clinical Trials Registry (PACTR201312000722390) on 9 December 2013.

Network Structure and Biased Variance Estimation in Respondent Driven Sampling

This paper explores bias in the estimation of sampling variance in Respondent Driven Sampling (RDS). Prior methodological work on RDS has focused on its problematic assumptions and the biases and inefficiencies of its estimators of the population mean. Nonetheless, researchers have given only slight attention to the topic of estimating sampling variance in RDS, despite the importance of variance estimation for the construction of confidence intervals and hypothesis tests. In this paper, we show that the estimators of RDS sampling variance rely on a critical assumption that the network is First Order Markov (FOM) with respect to the dependent variable of interest. We demonstrate, through intuitive examples, mathematical generalizations, and computational experiments that current RDS variance estimators will always underestimate the population sampling variance of RDS in empirical networks that do not conform to the FOM assumption. Analysis of 215 observed university and school networks from Facebook and Add Health indicates that the FOM assumption is violated in every empirical network we analyze, and that these violations lead to substantially biased RDS estimators of sampling variance. We propose and test two alternative variance estimators that show some promise for reducing biases, but which also illustrate the limits of estimating sampling variance with only partial information on the underlying population social network.

Brief Report: Respondent-driven Sampling Estimators Under Real and Theoretical Recruitment Conditions of Female Sex Workers in China

We compare the performance of multiple respondent-driven sampling estimators under different sample recruitment conditions in hidden populations of female sex workers in the midst of China's ongoing epidemic of sexually transmitted infections. We first examine empirically calibrated simulations grounded in survey data to evaluate the relative performance of each estimator under ideal sampling conditions consistent with respondent-driven sampling assumptions and under conditions that mimic observed respondent-driven sampling recruitment processes. One estimator, which incorporates respondents' reports on their network of contacts, substantially out-performs the others under all conditions. We then apply the estimators to empirical samples of female sex workers collected in two Chinese cities that include unique data on respondents' networks. These empirical results are consistent with the simulation results, suggesting that traditional respondent-driven sampling estimators overestimate the proportion of female sex workers working in low tiers of sex work and are likely to overstate the sexually transmitted infection risk profiles of these populations.

HIV Prevalence and Awareness of Positive Serostatus Among Men Who Have Sex With Men and Transgender Women in Bogotá, Colombia

OBJECTIVES

We estimated HIV prevalence among men who have sex with men (MSM) and transgender women in Bogotá, Colombia, and explored differences between HIV-positive individuals who are aware and unaware of their serostatus.

METHODS

In this cross-sectional 2011 study, we used respondent-driven sampling (RDS) to recruit 1000 MSM and transgender women, who completed a computerized questionnaire and received an HIV test.

RESULTS

The RDS-adjusted prevalence was 12.1% (95% confidence interval [CI] = 8.7, 15.8), comparable to a previous RDS-derived estimate. Among HIV-positive participants, 39.7% (95% CI = 25.0, 54.8) were aware of their serostatus and 60.3% (95% CI = 45.2, 75.5) were unaware before this study. HIV-positive-unaware individuals were more likely to report inadequate insurance coverage, exchange sex (i.e., sexual intercourse in exchange for money, goods, or services), and substance use than other participants. HIV-positive-aware participants were least likely to have had condomless anal intercourse in the previous 3 months. Regardless of awareness, HIV-positive participants reported more violence and forced relocation experiences than HIV-negative participants.

CONCLUSIONS

There is an urgent need to increase HIV detection among MSM and transgender women in Bogotá. HIV-positive-unaware group characteristics suggest an important role for structural, social, and individual interventions.

Sexual Mixing in Shanghai: Are Heterosexual Contact Patterns Compatible With an HIV/AIDS Epidemic?

China's HIV prevalence is low, mainly concentrated among female sex workers (FSWs), their clients, men who have sex with men, and the stable partners of members of these high-risk groups. We evaluate the contribution to the spread of HIV of China's regime of heterosexual relations, of the structure of heterosexual networks, and of the attributes of key population groups with simulations driven by data from a cross-sectional survey of egocentric sexual networks of the general population of Shanghai and from a concurrent respondent-driven sample of FSWs. We find that the heterosexual network generated by our empirically calibrated simulations has low levels of partner change, strong constraints on partner selection by age and education, and a very small connected core, mainly comprising FSWs and their clients and characterized by a fragile transmission structure. This network has a small HIV epidemic potential but is compatible with the transmission of bacterial sexually transmitted infections (STIs), such as syphilis, which are less susceptible to structural breaks in transmission of infection. Our results suggest that policies that force commercial sex underground could have an adverse effect on the spread of HIV and other STIs.

Diagnostics for Respondent-driven Sampling

Respondent-driven sampling (RDS) is a widely used method for sampling from hard-to-reach human populations, especially populations at higher risk for HIV. Data are collected through peer-referral over social networks. RDS has proven practical for data collection in many difficult settings and is widely used. Inference from RDS data requires many strong assumptions because the sampling design is partially beyond the control of the researcher and partially unobserved. We introduce diagnostic tools for most of these assumptions and apply them in 12 high risk populations. These diagnostics empower researchers to better understand their data and encourage future statistical research on RDS.

Respondent driven sampling is an effective method for engaging methamphetamine users in HIV prevention research in South Africa

BACKGROUND

South Africa, in the midst of the world's largest HIV epidemic, has a growing methamphetamine problem. Respondent driven sampling (RDS) is a useful tool for recruiting hard-to-reach populations in HIV prevention research, but its use with methamphetamine smokers in South Africa has not been described. This study examined the effectiveness of RDS as a method for engaging methamphetamine users in a Cape Town township into HIV behavioral research.

METHODS

Standard RDS procedures were used to recruit active methamphetamine smokers from a racially diverse peri-urban township in Cape Town. Effectiveness of RDS was determined by examining social network characteristics (network size, homophily, and equilibrium) of recruited participants.

RESULTS

Beginning with eight seeds, 345 methamphetamine users were enrolled over 6 months, with a coupon return rate of 67%. The sample included 197 men and 148 women who were racially diverse (73% Coloured, 27% Black African) and had a mean age of 28.8 years (SD=7.2). Social networks were adequate (mean network size >5) and mainly comprised of close social ties. Equilibrium on race was reached after 11 waves of recruitment, and after ≤3 waves for all other variables of interest. There was little to moderate preference for either in- or out-group recruiting in all subgroups.

CONCLUSIONS

Results suggest that RDS is an effective method for engaging methamphetamine users into HIV prevention research in South Africa. Additionally, RDS may be a useful strategy for seeking high-risk methamphetamine users for HIV testing and linkage to HIV care in this and other low resource settings.

Epidemiological challenges to the assessment of HIV burdens among key populations: respondent-driven sampling, time-location sampling and demographic and health surveys

PURPOSE OF REVIEW

Measuring the burden of HIV among key populations is subject to many challenges. Sufficient quantities of valid HIV prevalence and programme coverage data are required to effectively respond to the epidemic.

RECENT FINDINGS

Ability to validate exposure to unprotected sex through the innovative use of prostate-specific antigen provides confirmation of condom use. A new weighting scheme based on frequency of venue attendance for time location samples should improve validity of data obtained with this method. Two new proportion estimators, new diagnostic methods, a new population size estimator and new analysis software will provide more robust results from respondent-driven sampling (RDS).

SUMMARY

Analytical advances have improved the potential quality of results from surveys using time location and RDS. However, data from sufficient numbers of sites over sufficient number of years are still needed to provide clear national pictures of distribution and trends of HIV infection.

Stickiness of respondent-driven sampling recruitment chains

Respondent-driven sampling (RDS) is an increasingly popular chain-referral sampling method. Although it has proved effective at generating samples of hard to reach populations—meaning populations for which sampling frames are not available because they are hidden or socially stigmatized like sex workers or injecting drug users—quickly and cost-effectively, the ease of collecting the sample comes with a cost: bias or inefficiency in the estimates of population parameters (Gile & Handcock, 2010; Goel & Salganik, 2010). One way that RDS can produce inefficient estimates is if one or more of the recruitment chains gets stuck among members of a cohesive subpopulation, preventing the RDS sampling process from exploring other areas of the network. If that happens, members of the population subgroup recruit one another repeatedly, leading to an increase in sample size without increasing the diversity of the sample. This type of stickiness is particularly likely when hidden populations are stratified, and the stratified groups are organized into venues that provide opportunities to recruit other members of the same stratum. Female sex workers (FSW) in China, who are stratified into tiers of sex work that are correlated with marital status, age, and risk behaviors, are a prime example (Merli et al., 2014; Yamanis et al., 2013). Chinese FSW recruit clients from venues such as karaoke bars, massage parlors, or street corners. At larger venues, sex workers who participate in an RDS study might recruit other members of the same venue into the study at a higher rate than expected, leading to inefficient estimates. In short, the chain could get stuck in a venue.

Challenges to recruiting population representative samples of female sex workers in China using Respondent Driven Sampling

We explore the network coverage of a sample of female sex workers (FSWs) in China recruited through Respondent Drive Sampling (RDS) as part of an effort to evaluate the claim of RDS of population representation with empirical data. We take advantage of unique information on the social networks of FSWs obtained from two overlapping studies--RDS and a venue-based sampling approach (PLACE)--and use an exponential random graph modeling (ERGM) framework from local networks to construct a likely network from which our observed RDS sample is drawn. We then run recruitment chains over this simulated network to assess the assumption that the RDS chain referral process samples participants in proportion to their degree and the extent to which RDS satisfactorily covers certain parts of the network. We find evidence that, contrary to assumptions, RDS oversamples low degree nodes and geographically central areas of the network. Unlike previous evaluations of RDS which have explored the performance of RDS sampling chains on a non-hidden population, or the performance of simulated chains over previously mapped realistic social networks, our study provides a robust, empirically grounded evaluation of the performance of RDS chains on a real-world hidden population.