Lessons learned from respondent-driven sampling recruitment in Nairobi: experiences from the field

Smartphone-based Respondent Driven Sampling (RDS): A methodological advance in surveying small or ‘hard-to-reach’ populations

Producing statistically robust profiles of small or ‘hard-to-reach’ populations has always been a challenge for researchers. Since surveying the wider population in order to capture a large enough sample of cases is usually too costly or impractical, researchers have been opting for ‘snowballing’ or ‘time-location sampling’. The former does not allow for claims to representativeness, and the latter struggles with under-coverage and estimating confidence intervals. Respondent Driven Sampling (RDS) is a method that combines snowballing sampling with an analytical algorithm that corrects for biases that arise in snowballing. For all its advantages, a major weakness of RDS has been around data collection. Traditionally done on-site, the process is costly and lengthy. When done online, it is cheaper and faster but under a serious threat from fraud, compromising data quality and validity of findings. This paper describes a real-life application of a RDS data collection system that maximizes fraud prevention while still benefiting from low cost and speedy data collection.

"My Friends Would Believe My Word": Appropriateness and Acceptability of Respondent-Driven Sampling in Recruiting Young Tertiary Student Men Who Have Sex with Men for HIV/STI Research in Nairobi, Kenya

(1) Background: We conducted formative research to assess the appropriateness and acceptability of respondent-driven sampling (RDS) in recruiting tertiary student men who have sex with men (TSMSM) into a prospective human immunodeficiency virus/sexually transmitted infection (HIV/STI) biobehavioral survey in Nairobi, Kenya. (2) Methods: Between September and October 2020, semi-structured qualitative interviews were held with service providers from organizations that serve MSM (n = 3), and TSMSM (n = 13). Interviews were conducted in English, audio-recorded and transcribed, then thematically analyzed using NVivo version 11. (3) Results: Service providers reflected that RDS was appropriate due to the large though concealed networks of TSMSM. TSMSM perceived RDS to be acceptable based on their large social network sizes and the trust that existed amongst themselves. TSMSM were concerned about participating due to the risk of being outed as MSM and hence emphasized that researchers needed to assure them of their confidentiality and include MSM as part of the study team to encourage participation. (4) Conclusions: RDS was perceived as both an appropriate and acceptable sampling method. Use of RDS should be considered as a strategy for recruiting young, marginalized populations for HIV/STI research.

Implementation of respondent driven sampling in Nairobi, Kenya, for tracking key family planning indicators among adolescents and youth: lessons learnt

OBJECTIVE

Adolescents and youth constitute a significant proportion of the population in developing nations. Conventional survey methods risk missing adolescents/youth because their family planning/contraception (FP/C) behavior is hidden. Respondent-driven sampling (RDS), a modified chain-referral recruitment sampling approach, was used to reach unmarried adolescents/youth aged 15-24 in Nairobi, Kenya to measure key FP/C indicators. Seeds were selected and issued with three coupons which they used to invite their peers, male or female, to participate in the study. Referred participants were also given coupons to invite others till sample size was achieved. We report on key implementation parameters following standard RDS reporting recommendations.

RESULTS

A total of 1674 coupons were issued to generate a sample size of 1354. Coupon return rate was 82.7%. Study participants self-administered most survey questions and missing data was low. Differential enrolment by gender was seen with 56.0% of females recruiting females while 44.0% of males recruited males. In about two months, it was possible to reach the desired sample size using RDS methodology. Implementation challenges included presentation of expired coupons, recruitment of ineligible participants and difficulty recruiting seeds and recruits from affluent neighborhoods. Challenges were consistent with RDS implementation in other settings and populations. RDS can complement standard surveillance/survey approaches, particularly for mobile populations like adolescents/youth.

Employing Respondent Driven Sampling (RDS) to recruit people who inject drugs (PWID) and other hard-to-reach populations during COVID-19: Lessons learned

BACKGROUND

Respondent Driven Sampling (RDS) is an effective sampling strategy to recruit hard-to-reach populations but the impact of the COVID-19 pandemic on the use of this strategy in the collection of data involving human subjects, particularly among marginalized and vulnerable populations, is not known. Based on an ongoing study using RDS to recruit and study the interactions between HIV infection, injection drug use, and the microbiome in Puerto Rico, this paper explores the effectiveness of RDS during the pandemic and provided potential strategies that could improve recruitment and data collection.

RESULTS

RDS was employed to evaluate its effectiveness in recruiting a group of people who inject drugs (PWID) and controls (N = 127) into a study in the midst of the COVID-19 pandemic. The participants were distributed among three subsets: 15 were HIV+ and PWID, 58 were HIV- PWID, and 54 were HIV+ and not PWID.

FINDINGS

Results show that recruitment through peer networks using RDS was possible across all sub-groups. Yet, while those in the HIV+ PWID sub-group managed to recruit from other-sub groups of HIV- PWID and HIV+, this occurred at a lower frequency.

CONCLUSION

Despite the barriers introduced by COVID-19, it is clear that even in this environment, RDS continues to play a powerful role in recruiting hard-to-reach populations. Yet, more attention should be paid at how future pandemics, natural disasters, and other big events might affect RDS recruitment of vulnerable and hard-to-reach populations.

Project BESPOKE (Integrated Bio-Behavioral Assessment of HIV and STI Among Young Tertiary Student Men Who Have Sex With Men in Nairobi, Kenya): A Respondent-Driven Sampling Survey Protocol

Introduction: Globally, men who have sex with men (MSM) are a key population for the human immunodeficiency virus (HIV) epidemic. Among MSM, young men who have sex with men (YMSM) are disproportionately affected by HIV and other sexually transmitted infections (STI). However, there is a dearth of research and interventions targeting HIV/STI prevention among YMSM. In Kenya, there is paucity of knowledge on the burden of HIV/STI and related factors among YMSM, including tertiary student men who have sex with men (TSMSM). The barriers TSMSM experience in accessing and utilizing health services in their learning institutions have seldom been explored. In the context of healthcare providers (HCP) working in tertiary institutions, little is known about their knowledge, attitudes, and practices toward providing services to TSMSM. Methods: The aims of the study are to: estimate prevalence and correlates of HIV/STI among TSMSM; estimate population size of TSMSM; explore experiences of TSMSM with access and utilization of health services; and assess HCP knowledge of, attitudes toward, and practices in provision of services to TSMSM. A mixed-methods approach will be used in three phases: Phase I-formative qualitative research will be conducted to understand TSMSM social networks, select "seeds", and explore strategies for implementing a respondent-driven sampling (RDS) survey. Interviews will be conducted with at least three staff who work in community based/non-governmental organizations (CBO/NGO) that serve MSM and at least 10 TSMSM. Phase II-an integrated bio-behavioral assessment (IBBA) will be conducted, where 200 TSMSM recruited by RDS will be offered HIV/STI testing, complete a behavioral survey, and provide information for population size estimation (PSE). Phase III-in-depth interviews will be held with 20 TSMSM selected from 200 TSMSM in phase II, to explore their experiences with access and utilization of healthcare services. Focus group discussions (FGD) will be conducted with HCP working in tertiary institutions to assess their knowledge of, attitudes toward, and practices in providing services to TSMSM. Data collection started in September 2020 and is expected to end by September 2021. Discussion: Findings from this study will be useful in informing HIV/STI prevention programming for TSMSM, by policy makers such the Kenyan ministries of health and education, tertiary education institutions, service providers, advocacy groups, and other interested stakeholders.

Predictors of Depressive Symptoms and Post Traumatic Stress Disorder Among Women Engaged in Commercial Sex Work in Southern Uganda

This study examined the factors associated with depressive symptoms and post traumatic depressive disorder (PTSD) among economically vulnerable women engaged in commercial sex work (WESW) in southern Uganda. Baseline data from a longitudinal cluster randomized study involving 542 self-identified WESW (18-55 years), recruited from 19 HIV hotspots were analyzed. Hierarchical linear regression modelling was utilized to estimate individual, family-level and economic-level predictors of depressive symptoms and PTSD. Family cohesion, sex work stigma, HIV status, financial distress, household assets, number of children and number of household income earners, were associated with PTSD. Similarly, family cohesion, number of people in the household, HIV status, sex work stigma, financial distress, and household assets, were associated with depressive symptoms. Women engaged in commercial sex work are at a higher risk of HIV and poor mental health outcomes. Sex work stigma and financial distress elevate levels of depressive symptoms and PTSD, over and above an individual's HIV status. Family and economic-level factors have the potential to mitigate the risk of poor mental health outcomes. As such, integrating stigma reduction and economic strengthening components in the programming targeting WESW-a key population, may be critical to address their mental health outcomes.

Telephone peer recruitment and interviewing during a respondent-driven sampling (RDS) survey: feasibility and field experience from the first phone-based RDS survey among men who have sex with men in Côte d'Ivoire

Background

Many respondent-driven sampling (RDS) methodologies have been employed to investigate hard-to-reach populations; however, these methodologies present some limits. We describe a minimally investigated RDS methodology in which peer recruitment and interviewing are phone-based. The feasibility of the methodology, field experiences, validity of RDS assumptions and characteristics of the sample obtained are discussed.

Methods

We conducted a phone-based RDS survey among men who have sex with men (MSM) aged 18 or above and living in Côte d’Ivoire. Eight initial MSM across Côte d’Ivoire were selected. Participants were asked to call a hotline to be registered and interviewed by phone. After the participants completed the questionnaire, they were asked to recruit a maximum of 3 MSM from their acquaintances.

Results

During the 9 months of the survey, 576 individuals called the hotline, and 518 MSM completed the questionnaire. The median delay between the invitation to participate and the completion of the questionnaire by peer-recruited MSM was 4 days [IQR: 1–12]. The recruitment process was not constant, with high variation in the number of people who called the hotline during the survey period.

RDS chain convergence to equilibrium was reached within 6 waves for most of the selected variables. For the network size estimation assumption, participants who incorrectly estimated their network size were observed.

Regarding the sample obtained, MSM were recruited from all the regions of Côte d’Ivoire with frequent interregional recruitment; 23.5% of MSM were recruited by someone who does not live in the same region. Compared to the MSM who participated in two other surveys in Côte d’Ivoire, the MSM in our sample were less likely to know about an MSM non-governmental organisation. However, MSM aged 30 years old and above and those with a low level of education were underrepresented in our sample.

Conclusion

We show that phone-based RDS surveys among MSM are feasible in the context of sub-Saharan Africa. Compared to other classical RDS survey methodologies, the phone-based RDS methodology seems to reduce selection bias based on geography and proximity with the MSM community. However, similar to other methodologies, phone-based RDS fails to reach older and less-educated MSM.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12874-021-01208-x.

Mental Health and Health Risk Behaviors of Active Duty Sexual Minority and Transgender Service Members in the United States Military

Purpose: The aim of this study was to examine health risk behaviors and mental health outcomes among sexual minority and transgender active duty military service members and their heterosexual and cisgender counterparts. Methods: Participants (N = 544) were recruited by using respondent-driven sampling between August 2017 and March 2018 and completed an online survey by using validated measures of cigarette smoking, alcohol use, anxiety, depression, post-traumatic stress disorder (PTSD), and suicidality. Bayesian random intercept multiple logistic regressions were used to understand differences between sexual minority participants and heterosexual participants as well as between transgender participants and both their cisgender sexual minority and cisgender heterosexual peers. Results: Cisgender sexual minority women service members were more likely to meet criteria for problematic alcohol use (adjusted odds ratio [aOR] = 10.11) and cigarette smoking (aOR = 7.12) than cisgender heterosexual women. Cisgender sexual minority men had greater odds of suicidality (aOR = 4.73) than their cisgender heterosexual counterparts. Transgender service members had greater odds of anxiety, PTSD, depression, and suicidality than their cisgender peers. Conclusion: Military researchers and policymakers who seek to improve the overall health and well-being of sexual minority and transgender service members should consider programs and policies that are tailored to specific health outcomes and unique sexual minority and transgender subgroups.

Recognizing the hidden: strengthening the HIV surveillance system among key and priority populations in Mozambique

High quality, representative data from HIV surveillance systems that have country ownership and commitment are critical for guiding national HIV responses, especially among key and priority populations given their disproportionate role in the transmission of the virus. Between 2011 to 2013, the Mozambique Ministry of Health has conducted five Biobehavioral Surveillance Surveys among key populations (female sex workers, men who has sex with men and people who inject drugs) and priority populations (long distance truck drives and miners) as part of the national HIV surveillance system. We describe the experience of strengthening the HIV surveillance system among those populations through the implementation of these surveys in Mozambique. We document the lessons learned through the impact on coordination and collaboration; workforce development and institutional capacity building; data use and dissemination; advocacy and policy impact; financial sustainability and community impact. Key lessons learned include the importance of multisectoral collaboration, vital role of data to support key populations visibility and advocacy efforts, and institutional capacity building of government agencies and key populations organizations. Given that traditional surveillance methodologies from routine data often do not capture these hidden populations, it will be important to ensure that Biobehavioral Surveillance Surveys are an integral part of ongoing HIV surveillance activities in Mozambique.

Respondent Driven Sampling for Immigrant Populations: A Health Survey of Foreign-Born Korean Americans

This study examined feasibility and methodological utilities of respondent driven sampling (RDS) for Korean immigrants. We conducted the Health and Life Study of Koreans (HLSK), a Web-based RDS study targeting foreign-born Korean Americans. Through chain referrals, n = 638 participated. Geographic coverage and estimates of HLSK were compared to foreignborn Korean samples in the American Community Survey and the California Health Interview Survey as benchmarks. Compared to the benchmarks, HLSK fared well on the geographic coverage, household type and size, employment and health insurance but over-captured those who were younger, more recent immigrants, with higher education and with disability. Existing RDS-specific estimators were largely ineffective. Conclusions. RDS may serve as a cost-effective tool for recruiting recent immigrants, a harder-to-recruit subgroup within minorities. However, recruitment noncooperation posed operational challenges, a critical gap in the literature. This leaves RDS yet to be a reliable methodology.

Evolving HIV epidemics: the urgent need to refocus on populations with risk

Purpose of review

To explore the comparative importance of HIV infections among key populations and their intimate partners as HIV epidemics evolve, and to review implications for guiding responses.

Recent findings

Even as concentrated epidemics evolve, new infections among current and former key population members and their intimate partners dominate new infections. Prevalent infections in the general population grow primarily because of key population turnover and infections among their intimate partners. In generalized epidemic settings, data and analysis on key populations are often inadequate to assess the impact of key population-focused responses, so they remain limited in coverage and under resourced. Models must incorporate downstream infections in comparing impacts of alternative responses.

Summary

Recognize that every epidemic is unique, moving beyond the overly simplistic concentrated/generalized epidemic paradigm that can misdirect resources. Guide HIV responses by gathering and using locally relevant data, understanding risk heterogeneity, and applying modeling at both national and sub-national levels to optimize resource allocations among different populations for greatest impact. Translate this improved understanding into clear, unequivocal advice for policymakers on where to focus for impact, breaking them free of the generalized/concentrated paradigm limiting their thinking and affecting their decisions.

Preferences for long-acting Pre-Exposure Prophylaxis (PrEP) for HIV prevention among South African youth: results of a discrete choice experiment

INTRODUCTION

Existing biomedical HIV prevention options, though highly effective, present substantial adherence challenges. End-user input on early-stage design of new HIV prevention approaches is critical to yielding products that achieve high uptake and adherence. The iPrevent Study examined youths' preferences for key attributes of long-acting Pre-Exposure Prophylaxis (PrEP), with a focus on characteristics pertinent to product delivery alongside key modifiable product attributes.

METHODS

A discrete choice experiment was conducted with female and male youth aged 18 to 24 in two high-density communities in Cape Town, South Africa during the period July 2017 to January 2019. Sexually active, PrEP-naïve youth were recruited using population-based sampling; targeted sampling was used to enrol men who have sex with men (MSM). In a series of nine questions, participants were asked to choose between two hypothetical products composed of five attributes (form, dosing frequency, access, pain, insertion site). We used a random-parameters logit model to estimate preference weights and trade-offs among product alternatives. We examined differences across three subgroups: females, men who have sex with only women (MSW) and MSM.

RESULTS

A total of 807 participants (401 female) were enrolled with a median age of 21 years. Males included 190 MSM. Most youth had tested for HIV (95%) and reported being HIV-negative (91%). Across all groups, duration of effectiveness was the most important attribute, with strong preference for less frequent dosing. Injections were favoured over implants, though these preferences were strongest for females and MSM. Females preferred a product offered at a health clinic and disliked pharmacy access; all groups preferred the arm as the insertion site. Youth were willing to trade their preferred product form for longer duration.

CONCLUSIONS

Youth indicated strong preferences for longer duration products. Each attribute nonetheless influenced preferences, offering insight into trade-offs that inform long-acting PrEP development.

Lessons learned from using respondent-driven sampling (RDS) to assess sexual risk behaviors among Kenyan young adults living in urban slum settlements: A process evaluation

Background

Respondent-driven sampling (RDS) is a peer-referral sampling methodology used to estimate characteristics of underserved groups that cannot be randomly sampled. RDS has been implemented in several settings to identify hidden populations at risk for HIV, but few studies have reported the methodological lessons learned on RDS design and implementation for assessing sexual risk behaviors in marginalized youth.

Methods

We used RDS to recruit N = 350 young adults, aged 18 to 22, who were living in urban slum settlements in Nairobi, Kenya. A structured survey was used to assess sexual risk behaviors. Twenty seeds were selected and asked to recruit up to three eligible peers. We used small monetary incentives and a three-day recruitment coupon with sequential numbers linking recruiters to their recruits.

Results

Data collection was completed in 8 days with a maximum chain length of 6 waves. Each seed yielded 16 to 21 eligible recruits. Three (15%) seeds were unproductive and were replaced. RDS benefits were high identification rates (90% coupons returned per coupons given), high eligibility rates (100% eligible recruits per coupons returned), and high efficiency (~39 eligible recruits per day). 44% of the sample was female. Most recruits (74%) reported being “friends” for 7+ years with their recruiter. RDS overcame feasibility concerns of household-, clinic-, and school-based sampling methodologies in that underserved youth who were unemployed (68%), out of school (48%), ethnic minorities (26%), and having prior residential instability (≥2 moves in the past year) (20%) were successfully recruited, based on weighted analyses. Youth reporting HIV risk behaviors, including unprotected sex (38%), sex while high/drunk (35%), and sex exchange for pay (14%), were also enrolled. However, 28% were not sexually active within the last 6 months. Challenges included managing wait times during peaks and participant referral expectations. Community engagement, use of study-stamped coupons, broad inclusion criteria, incentives, and study sites within walking distances all contributed to the successful implementation of the sampling methodology.

Conclusion

RDS is an important tool in reaching a diverse sample of underserved and at-risk young adults for study participation. Implications for optimizing RDS for behavioral studies in this population are discussed.

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.

Human immunodeficiency virus (HIV) among men who have sex with men: results of the first integrated biological and behavioral survey in Burkina Faso, West Africa

BACKGROUND

Many men who have sex with men (MSM) are at significant risk for HIV infection. The objective of this study was to determine the prevalence and correlates of HIV infection among MSM in Burkina Faso.

METHODS

A cross-sectional biological and behavioral survey was conducted from January to August 2013 among MSM in Ouagadougou and Bobo-Dioulasso. MSM 18 years old and above were recruited using respondent driven sampling (RDS). A survey was administered to study participants followed by HIV testing. Population prevalence estimates and 95% confidence intervals (CI) adjusted for the RDS design were produced using the RDS Analysis Tool version 6.0.1 (RDS, Inc., Ithaca, NY).

RESULTS

A total of 662 MSM were enrolled in Ouagadougou (n = 333) and Bobo-Dioulasso (n = 329). The majority were unmarried, with an average age of 22.1 ± 4.4 years old in Ouagadougou and 23.1 ± 4.7 years old in Bobo-Dioulasso. RDS-adjusted HIV prevalence was 1.7% (95% CI: 0.9-3.1) in Ouagadougou and 2.7% (95% CI: 1.6-4.6) in Bobo-Dioulasso. HIV prevalence among MSM under 25 years old was 1.3% (95% CI: 0.6-2.8) and 0.9% (95% CI: 0.4-2.5) respectively in Ouagadougou and Bobo-Dioulasso, compared to 5.4% (95% CI: 2.2-12.5) and 6.6% (95% CI: 3.4-12.3) among those 25 years old or older in these cities (p = 0.010 and p < 0.001).

CONCLUSIONS

Results from this first biological and behavioral survey among MSM in Burkina Faso suggest a need for programs to raise awareness among MSM and promote safer sex, particularly for young MSM to prevent HIV transmission. These programs would need support from donors for innovative actions such as promoting and providing pre-exposure prophylaxis, condoms and water-based lubricants, HIV counseling, testing, early treatment initiation and effective involvement of the MSM communities.

Using respondent-driven sampling (RDS) to recruit illegal poly-substance users in Cape Town, South Africa: implications and future directions

BACKGROUND

South Africa continues to witness an increase in illicit poly-substance use, although a precise measurement continues to be compounded by difficulties in accessing users. In a pilot attempt to use respondent-driven sampling (RDS)-a chain referral sampling method used to access populations of individuals who are 'hard-to-reach'-this article documents the feasibility of the method as recorded in a simultaneously run, multisite, poly-substance study in Cape Town. Here we aim to a) document the piloting of RDS among poly-substance users in the three socio-economic disparate communities targeted; b) briefly document the results; and c) review the utility of RDS as a research tool.

METHODS

Three cross-sectional surveys using standard RDS procedures were used to recruit active poly-substance users and were concurrently deployed in three sites. Formative research was initially conducted to assess the feasibility of the survey. To determine whether RDS could be used to successfully recruit poly-substance users, social network characteristics, such as network size was determined.

RESULTS

A 42.5 % coupon return rate was recorded in total from 12 initial seeds. There were vast differences in the recruitment chains of individual seeds-two generated more than 90 recruits, and 2 of the 10 recruitment chains showing a length of more than 10 waves. Findings include evidence of the use of 3 or more substances in all three sites, high levels of unemployment among users, with more than a third of participants in two sites reporting arrest for drug use in the past 12 months.

CONCLUSIONS

Our results indicate that RDS was a feasible and acceptable sampling method for recruiting participants who may not otherwise be accessible. Future studies can use RDS to recruit such cohorts, and the method could form part of broader efforts to document vulnerable populations.