Application of the Respondent-Driven Sampling methodology in a biological and behavioral surveillance survey among female sex workers, Brazil, 2016 / Aplicação da metodologia Respondent-Driven Sampling em pesquisa biológica e comportamental com mulheres trabalhadoras do sexo, Brasil, 2016

ABSTRACT Introduction: Respondent-Driven Sampling (RDS) has been used in surveys with key populations at risk of HIV infection, such as female sex workers (FSW). This article describes the application of the RDS method among FSW in 12 Brazilian cities, during a survey carried out in 2016. Methodology: A biological and behavioral surveillance study carried out in 12 Brazilian cities, with a minimum sample of 350 FSW in each city. Tests were performed for HIV, syphilis, and hepatitis B and C infections. A social-behavioral questionnaire was also applied. Results: The sample was comprised of 4,328 FSW. For data analysis, the sample was weighted according to each participant's network size (due consideration to the implications of RDS complex design and to the effects of homophilia are recommended). Discussion: Although RDS methods for obtaining a statistical sample are based on strong statistical assumptions, allowing for an estimation of statistical parameters, with each new application the method has been rethought. In the analysis of whole-sample data, estimators were robust and compatible with those found in 2009. However, there were significant variations according to each city. Conclusion: The achieved sample size was of great relevance for assessing progress and identifying problems regarding the prevention and treatment of Sexually Transmitted Infections. New RDS studies with more time and operational resources should be envisaged. This could further network development.

RESUMO Introdução: O método de amostragem Respondent-Driven Sampling (RDS) tem sido utilizado em inquéritos com populações-chave sob maior risco de infecção pelo HIV, como as mulheres trabalhadoras do sexo (MTS). Este artigo tem o objetivo de descrever a implementação do RDS entre MTS em 12 cidades brasileiras em 2016. Metodologia: Trata-se de um estudo de vigilância biológica e comportamental realizado em 12 cidades brasileiras, com amostra mínima de 350 MTS em cada cidade. Foram realizados testes para infecções por HIV, sífilis, hepatites B e C, e aplicou-se questionário sociocomportamental. Resultados: Participaram 4.328 MTS. Para a análise dos dados, foi elaborada uma ponderação amostral considerando o tamanho da rede de cada participante; recomenda-se que o desenho complexo de amostragem por RDS e o efeito de homofilia sejam considerados. Discussão: Apesar de o RDS ser fundamentado em pressupostos estatísticos para obtenção de uma amostra probabilística e possibilitar estimação de parâmetros estatísticos, ele vem sendo repensado a cada nova aplicação. Na análise dos dados na totalidade da amostra, os estimadores mostraram-se robustos e coerentes aos encontrados em 2009. Entretanto, constataram-se grandes variações por cidade. Conclusão: O tamanho amostral alcançado foi de grande relevância para avaliar avanços e identificar problemas a respeito da prevenção e assistência às infecções sexualmente transmissíveis. Ressalta-se a necessidade de pensar estudos RDS com maior tempo e recursos para implementação, o que poderia permitir um melhor desenvolvimento das redes.

Evaluation on the effectiveness for respondent-driven sampling method among men who have sex with men in Beijing, 2017 / 中华流行病学杂志

Objective To evaluate the randomness and representativeness of respondent-driven sampling (RDS) tool in conducting the investigation in MSM population,in Beijing,2017.Methods RDS tool was used to recruit MSM population for a face-to-face interview with structured questionnaire and serological tests.Results A total of 600 MSM people were sampled and interviewed.The median number of personal network of seeds was 10,which was higher than other MSM people recruited.The numbers of recruitments by wave presented a skewed positive distribution and the highest number was in the fourth wave.It was also dramatically varied from different seeds.Three seeds had the longest chains and had recruited 184,113 and 92 MSM people,respectively.In contrast,five seeds recruited less than 10 MSM people.Two college students were the most non-generative seeds and each recruited only 1 MSM person.After five to nine waves of sampling,the major demographic characteristics reached equilibrium.Both convergence and bottleneck plots of major demographic characteristics reached convergence,although the plots on marriage and education did not.The homophiles of characteristics were all closed to 1,except for education.The HIV positive rate appeared as 7.9％ (95％CI:4.4％-11.4％). Conclusions Results from this study showed that RDS could be used as a feasible sampling method for the study on MSM population with major demographic characteristics reached equilibrium.The process of recruitment appeared controllable and reasonable,showing that this could represent the MSM population in Beijing,in some degree.

Evaluation on the effectiveness for respondent-driven sampling method among men who have sex with men in Beijing, 2017 / 中华流行病学杂志

Objective To evaluate the randomness and representativeness of respondent-driven sampling (RDS) tool in conducting the investigation in MSM population,in Beijing,2017.Methods RDS tool was used to recruit MSM population for a face-to-face interview with structured questionnaire and serological tests.Results A total of 600 MSM people were sampled and interviewed.The median number of personal network of seeds was 10,which was higher than other MSM people recruited.The numbers of recruitments by wave presented a skewed positive distribution and the highest number was in the fourth wave.It was also dramatically varied from different seeds.Three seeds had the longest chains and had recruited 184,113 and 92 MSM people,respectively.In contrast,five seeds recruited less than 10 MSM people.Two college students were the most non-generative seeds and each recruited only 1 MSM person.After five to nine waves of sampling,the major demographic characteristics reached equilibrium.Both convergence and bottleneck plots of major demographic characteristics reached convergence,although the plots on marriage and education did not.The homophiles of characteristics were all closed to 1,except for education.The HIV positive rate appeared as 7.9％ (95％CI:4.4％-11.4％). Conclusions Results from this study showed that RDS could be used as a feasible sampling method for the study on MSM population with major demographic characteristics reached equilibrium.The process of recruitment appeared controllable and reasonable,showing that this could represent the MSM population in Beijing,in some degree.

Estimating The Size Of The Drug Using Population In Three Deep-South Provinces Of Thailand: Results From A Service Multiplier And Respondent Driven Sampling (RDS) Method

Objective: The deep-south provinces of Thailand, the border area where the population is majority Malay ethnic, have faced with armed-conflicts for the decade and are a major route for drug trafficking. Several studies have reported concerns about drug problems among local population but the size of drug using population has not been studied. The objective of this study is to estimate the size of drug using population in the deep-south provinces of Thailand. Methods: A cross-sectional survey, including interviews was conducted in 2016. Males aged 18-40 years who had used any drugs in past six months were recruited. Respondent-driven sampling method was used to reach the target population and multiplier method to estimate the size of drug using population. Results: The estimated number of male using drug population in three deep-south provinces was 13,545, making the prevalence of 50.2 per 1,000 male population of this age group, twice as high as the national prevalence. Most drug users were in Pattani (60.6%), followed by Narathiwas and Yala provinces. Kratom is the illicit drug with the highest number of current users (85.2% of all users). The majority of current users (60.6%) used more than one type of illicit drug, with kratom and methamphetamine being the most common combination (25.6%). Conclusion: This study depicts the situation of drug abuse in deep-south Thailand. Kratom was the most popular substance. The high prevalence of drug users in the area should be of concern by stakeholders and interventions to minimize and control the abuse are needed.

Vulnerabilidade e prevalência de HIV e sífilis em usuários de drogas no Recife: resultados de um estudo respondent-driven sampling / Vulnerability and prevalence of HIV and syphilis in drug users, Recife: results of a respondent-driven sampling based study

Os usuários de drogas têm sido desproporcionalmente afetados pelo HIV, desempenhando papel relevante na epidemia de aids no Brasil. Objetivos: Analisar o perfil e indicadores de vulnerabilidade, prevalência e fatores associados à infecção pelo HIV em usuários de drogas do Recife recrutados pelo método respondent-driven sampling (RDS). Materiais e métodos: Estudo sociocomportamental e de soroprevalência, com base no critério da Organização Pan-americana de Saúde para definição de usuários de drogas de alto risco de adquirir/transmitir HIV. Foram feitas entrevista com auxílio de computador com áudio, e realizados testes rápidos para HIV e sífilis, além de VDRL. Bairros de moradia e de uso de drogas foram georreferenciados, os dados ponderados pelo estimador RDS II e os fatores associados avaliados pela odds ratio. Resultados: Foram recrutados 400 usuários de drogas, residentes em 48 por cento dos bairros do Recife, distribuídos nas suas seis regiões político administrativas, e que faziam uso de drogas em 71 por cento dos bairros. A maioria homens (71 por cento), jovens, solteiros, com filhos e baixa condição socioeconômica, 22 por cento em situação de rua. População estigmatizada e criminalizada, com frequente vivência de violência física e sexual. Usavam crack 96 por cento dos usuários, pelo menos quatro dias por semana e consumiam múltiplas drogas. Metade iniciou há mais de seis anos o uso de drogas ilícitas, exceto maconha, antes dos 18 anos de idade, e não faz uso em locais públicos. Referiram trocar sexo por dinheiro ou drogas 27,5 por cento dos homens e 58,6 por cento das mulheres. Metade dos que foram a serviços de saúde, em decorrência de sintomas de DST, não receberam orientação quanto a medidas de prevenção. A cobertura de testagem para HIV foi de 39 por cento, abaixo da verificada para os homens que fazem sexo com outros homens e mulheres profissionais do sexo...

Drug users have been disproportionately affected by HIV, playing an importantrole in the aids epidemic in Brazil. Objectives: To analyze the profile and vulnerabilityindicators, prevalence and factors associated with HIV infection of drug users in Reciferecruited by the respondent-driven sampling (RDS) method. Materials and methods: ASocial behavior and seroprevalence study, using respondent-driven sampling, based on thecriterion of the Pan American Health Organization definition for drug users at high risk ofacquiring/transmitting HIV. Interviews were recorded using computer audio, and HIV andsyphilis rapid tests and VDRL were performed. Local housing and drug use weregeoreferenced were georeferenced, data were weighted according to RDS II estimator andassociated factors assessed by the odds ratio. Results: A total of 400 drug users wererecruited, residing in 48% of neighborhood of Recife distributed in its six politicaladministrative regions, and used drugs in 71% of neighborhood. The majority were male(71%), young, single, with children and low socioeconomic status, with 22% living on thestreets, suffering from stigma and criminalization, with frequent experience of physical andsexual violence. Approximately 96% used crack, at least four days a week and consumingmultiple drugs. Half the sample began with illicit drugs, except marijuana, before reaching 18years, have taken drugs for more than six years and do not taken drugs in public places...

Nuevas formas de muestreo para minorías y poblaciones ocultas: muestras por encuestado conducido en una población de inmigrantes sudamericanos / New forms of sampling for minority and hidden populations: respondent samples conducted in a south american immigrant population

El Respondent-Driven Sampling (RDS) es un método de muestreo para poblaciones ocultas o de difícil acceso. Se trata de un procedimiento que permite realizar estimaciones de la representatividad de la muestra en aquellos grupos en que desconocemos el marco muestral. En este estudio, se presenta tanto la descripción teórica de dicha forma de muestreo como el reporte de su aplicación práctica en el caso de minorías étnicas (inmigrantes bolivianos, peruanos y colombianos) en el norte de Chile. El estudio comenzó con ocho semillas y se reclutó una muestra total de 109 personas provenientes de Colombia, Perú y Bolivia. La información resultante del procedimiento RDS informa de un patrón similar de reclutamiento entre hombres y mujeres, reflejando que los tamaños de las redes de ambos grupos no difieren significativamente.

The Respondent-Driven Sampling (RDS) is a method of sampling for hidden or hard-to-reach populations. This is a procedure for estimating the representativeness of the sample in those groups that is unknow the sampling frame. This study presents both the theoretical description of this form of sampling as the report of its practical application in the case of ethnic minorities (immigrants Bolivians, Peruvians and Colombians) in northern Chile. The study started with 8 seeds and recruited a total sample of 109 people from Colombia, Peru and Bolivia. The information resulting from the procedure RDS reports a similar pattern of recruitment for men and women, reflecting the sizes of the networks of both groups did not differ significantly.

El muestreo dirigido por los entrevistados (MDE) para acceder a poblaciones en condiciones de vulnerabilidad frente al VIH: su aplicación en grupos de hombres que tienen sexo con hombres: [revisión] / Respondent-driven sampling (rds) as a new method to access vulnerable populations to hiv: its application in men who have sex with men: [review]

La infección por el VIH afecta a cerca de 30 millones de personas, de las cuales las más vulnerables son los trabajadores del sexo, usuarios de drogas inyectables, personas transgénero y los hombres que tienen sexo con otros hombres (HSH). Los estudios de ITS y VIH realizados con muestreos intencionales como la bola de nieve no logran alcanzar representatividad ni posibilidad de inferencia al resto de la población. A esta limitación se suma el desconocimiento de los tamaños poblacionales de estos grupos, dados el estigma y la discriminación que enfrentan. Hace una década se diseñó un nuevo método de muestreo para acceder a estas poblaciones, basado en lasteorías de Markov y del reclutamiento en cadena. El método se conoce en español como muestreo dirigido por los entrevistados (mde) y se ha implementado en diversos estudios de vih. Este artículo de revisión presenta los antecedentes, los fundamentos teóricos y el método propiamente dicho, y analiza algunos estudios llevados a cabo con esta nueva metodología.

Approximately 30 million people are living with HIV all aroundthe world in 2010. The most vulnerable HIV/aids groups are sex workers, intravenous drug users, transgender people and populationof men who have sex with men (MSM). Up to now, the surveillance and behavioral studies on sexual transmitted diseases (STI) and HIV have been conducted using intentional sampling methods like the snowball methodology, but this kind of sampling does not achieve representation and does not represent a conclusive method for the study of hidden populations in order to extend and apply its results to the general population. Furthermore, it is necessary to add to this limitation the lack of knowledge about the size of these groups that must face situations of stigma and discrimination. A decade ago a new method was designed to sample hidden and hard to reach populations based on Markov theories and on chain recruitment. This new method is known as respondent-driven sampling (RDS) and it has been used in several behavioral and HIV prevalence studies. This review article presents the background, the theoretical support and a description of the method. It also analyzes some studies carried out using this new methodology.

Práticas associadas à infecção pelo HIV na população feminina brasileira, em particular, em mulheres trabalhadoras do sexo / Practices associated with HIV infection among women in Brazil, particularly in female sex workers

O presente trabalho, apresentado sob a forma de três artigos, teve por objetivo investigar as práticas relacionadas à infecção pelo HIV na população feminina brasileira e, em particular, entre as mulheres trabalhadoras do sexo (TS). No primeiro artigo, são analisadas as informações provenientes da Pesquisa de Conhecimentos Atitudes e Práticas (PCAP) de 2008, com o objetivo de investigar as diferenças por sexo nas práticas relacionadas à infecção pelo HIV na população brasileira de 15 a 64 anos. Os resultados mostraram que as mulheres têm menor taxa de atividade sexual, iniciam a vida sexual mais tardiamente e têm menos parceiros casuais do que os homens. No entanto, o uso de preservativo é menor entre as mulheres: enquanto a proporção de uso regular de preservativo no último ano, entre os homens, foi de 51 por cento, entre as mulheres, foi de 34,6 por cento. Da mesma forma, outros indicadores de práticas sexuais de risco indicaram a maior vulnerabilidade das mulheres, em comparação com a população masculina. Por outro lado, a cobertura de teste de HIV é significativamente maior entre as mulheres (45,6 por cento) quando comparadas aos homens (27,2 por cento). No segundo artigo, foi feita revisão bibliográfica com busca sistemática de trabalhos científicos para identificar as metodologias de amostragem usadas em estudos para estimar a taxa de prevalência do HIV entre as trabalhadoras do sexo. Foram identificados 75 estudos, dos quais 63 utilizaram amostras de conveniência, limitando a possibilidade de fazer inferências estatísticas para o grupo das TS. Em 35 estudos analisados, o tamanho de amostra não estava adequado para a estimação da taxa de prevalência do HIV e apenas quatro estudos atenderam aos critérios de qualidade utilizados no trabalho (tamanho de amostra era adequado para estimar da taxa de prevalência do HIV; uso de método probabilístico de amostragem; e consideração do desenho de amostragem na análise)...

This work was presented in the form of three articles aimed at investigating thepractices related to HIV infection among women in Brazil and in particular amongfemale sex workers (FSW). In the first article, we analyzed the information from the Knowledge Attitudes and Practice Survey (PCAP), 2008, in order to examine gender differences in practices related to HIV infection in the Brazilian population aged 15 to 64 years. The resultsshowed that women who have lower rates of sexual activity, have become sexually active later and they have had fewer casual partners than men. However, condom use is lower among women: while the proportion of regular condom use over the past year among men was 51%, among women was 34.6%. Likewise, other indicators of risky sexual practices suggest greater vulnerability of women, compared with the male population. Furthermore, coverage of HIV testing is significantly higher among women (45.6%) when compared with men (27.2%). In the second article we conducted a review with a systematic search of scientific studies to identify the sampling methodology used in studies to estimate the prevalence of HIV among FSW. We identified 75 studies, of which 63 have used convenience sampling method, limiting the possibility of making statistical inferences for the group of FSW. In 35 studies analyzed, the sample size was not adequate toestimate HIV prevalence rate and only four studies met the quality criteria used in the study (sample size was adequate to estimate the HIV prevalence, use of probabilisticsampling, and consideration of sampling design during analysis)...

A respondent-driven sampling survey on HIV and risk factors among men who have sex with men in Chongqing / 中华流行病学杂志

Objective To examine HIV prevalence and related risk factors among men who have sex with men(MSM) in Chongqing, and to explore the feasibility of using respondent-driven sampling (RDS) in the survey. Methods Based on results from formative research, a RDS survey was designed and conducted to collect demographic, behavioral and serologic data. RDSAT was used to calculate point estimation and confidence intervals. SPSS was used for bi-variate analysis using RDSAT exported weighed data. NETDRAW was used to draw network diagram. Results Among 617 subjects recruited, the adjusted HIV and syphilis prevalence rates were 16.8% and 10.9%, respectively. 73.0% of the subjects were 20 to 29 years old and 72.9% were officially registered residents of Chongqing. 83.4% were single, with the proportion of students the highest, accounting for 24.6%. During the last six months, 83.4% of them reported ever having anal sex, and 54.0% reported having unprotected anal sex. Conclusion This survey confirmed that Chongqing had a higher reported HIV prevalence among MSM than from other Chinese cities. Comprehensive intervention services were required to address this alarmingly high prevalence, with focus on intervention through internet and those having syphilis infection. RDS seemed one of the effective ways of recruiting hidden MSM populations in Chongqing which had a large population of active MSM who did not frequently visit MSM venues as compared with snowball method.

Un estudio comparativo de tres métodos de muestreo para poblaciones ocultas ode difícil acceso / A comparison of three methods for sampling hard - to - reach or hidden populations

Los proyectos de investigación que necesitan reclutar participantes pertenecientesa poblaciones “ocultas” o “difíciles de encontrar” requieren de métodos de muestreoque no solo faciliten la recolección de los datos y la confidencialidad, sino tambiénque incrementen la validez externa y permitan hacer inferencias estadísticasapropiadas. Este artículo presenta tres métodos que han demostrado ser útiles enlos trabajos de investigación con estas poblaciones estigmatizadas, clandestinaso de difícil acceso: 1) Muestreo basado en sitios y horarios específicos (VenueBased Time- Location Sampling). 2) Muestreo de áreas en las que se localiza lapoblación blanco o población objeto de estudio (Targeted Sampling). 3) Muestreodirigido por el participante (Respondent Driven Sampling). Luego de analizar lascaracterísticas, ventajas y limitaciones de cada método de muestreo, se presentauna comparación de todos ellos en términos de validez externa, posibilidad deobtener muestras probabilísticas y empleo de investigación etnográfica.

Research projects that need to recruit “hard-to-reach” or “hidden populations”require sampling methods that not only facilitate data collection and confidentialitybut also increase external validity and allow for statistically appropriate inferences.This paper presents three sampling methods useful for researchers who workwith stigmatized hard-to-reach or clandestine populations: 1) Venue Based Time/Location Sampling, 2) Targeted Sampling; and 3) Respondent Driven Sampling.The characteristics, advantages and limitations for each method are analyzed.They are also compared in terms of the possibility of obtaining probabilisticsamples, external validity, and use of ethnographic research.

Projetos de pesquisa que necessitam recrutar participantes pertencentesa povoações “oculta” ou “difíceis de encontrar” requerem de métodos deamostragem que não só facilitem a colheita dos dados e a confidencialidade mastambém que aumentem a validade externa e permitam fazer inferências estatísticasapropriadas. Este artigo teórico apresenta três métodos que demonstraram serúteis nos trabalhos de pesquisa com estas povoações estigmatizadas, clandestinasou de difícil acesso: 1) Amostragem baseado em lugares e horários específicos(Venue Based Time- Location Sampling); 2) Amostragem de áreas nas quais selocaliza a população branco ou população objeto de estudo (Targeted Sampling);e 3) Amostragem dirigido pelo participante (Respondent Driven Sampling).Depois de analisar as características, vantagens e limitações de cada método deamostragem, se apresenta uma comparação de todos eles em termos de validadeexterna, possibilidade de obter amostras probabilísticas e emprego de pesquisaetnográfica.