Lisa g. Johnston , lsjohnston.global@gmail.Com

Sampling Adolescents/Young Key Populations (A/YKP) at Risk of HIV Exposure Using Respondent Driven Sampling (RDS)L ISA G . JOHNSTONWWW.L ISAGJOHNSTON.COM, LS JOHNSTON.GLOBAL@GMAIL .COMUNIVERSITY OF CAL IFORNIA , SAN FRANCISCOTUL ANE UN IVERS ITY, NEW ORLEANS

OutlineoWhy is RDS superior for sampling A/YKP?oRDS OverviewoRDS Challenges oRDS Resources

Why is RDS superior for sampling A/YKP?

◦ Useful for gathering large samples of quantitative data from hard-to-reach and hidden populations◦ No sampling frame◦ Allows participants to remain anonymous◦ Peers recruit their peers (stigma, discrimination and lack of trust)

◦ Has already been used successfully used in several countries to sample A/YKP◦ When analyzed correctly, provides representative findings about the network of the

population sampled◦ Manuals and tools available to help you◦ Free computer software available

Malekinejad M, et al. Using respondent-driven sampling methodology for HIV biological and behavioral surveillance in international settings: A systematic review. AIDS and Behavior. 2008.12(Suppl 1): 105-130. Johnston LG, Malekinejad M, Rifkin MR, et al. Implementation challenges to using respondent-driven sampling methodology for HIV biological and behavioral surveillance: Field experiences in international settings. AIDS and Behavior. 2008.12(Suppl 1): 131-141.

RDS OVERVIEW Basics and Challenges

Questions to ask yourself before initiating RDS

◦ Is the population of interest socially networked (i.e., they know each other as part of that population)?

◦ Can you find a small number of persons who are part of the population?

RDS Recruitment ChainSeed

RDS Recruitment ChainSeed Wave 1

RDS Recruitment ChainSeed Wave 1 Wave 2

RDS Recruitment ChainSeed Wave 1 Wave 2 Wave 3

RDS Recruitment ChainSeed Wave 1 Wave 2 Wave 3 Wave 4

RDS Recruitment ChainSeed Wave 1 Wave 2 Wave 3 Wave 4 Wave 5

RDS Recruitment ChainSeed Wave 1 Wave 2 Wave 3 Wave 4 Wave 5 Wave 6

RDS Recruitment ChainSeed Wave 1 Wave 2 Wave 3 Wave 4 Wave 5 Wave 6 Wave 7

RDS Recruitment ChainSeed Wave 1 Wave 2 Wave 3 Wave 4 Wave 5 Wave 6 Wave 7 Wave 8

RDS Recruitment ChainSeed Wave 1 Wave 2 Wave 3 Wave 4 Wave 5 Wave 6 Wave 7 Wave 8 Wave 9

RDS Recruitment ChainSeed Wave 1 Wave 2 Wave 3 Wave 4 Wave 5 Wave 6 Wave 7 Wave 8 Wave 9 Wave 10

RDS Recruitment ChainSeed Wave 1 Wave 2 Wave 3 Wave 4 Wave 5 Wave 6 Wave 7 Wave 8 Wave 9 Wave 10 Wave 11

RDS Recruitment Chain

Seed Wave 1 Wave 2 Wave 3 Wave 4 Wave 5 Wave 6 Wave 7 Wave 8 Wave 9 Wave 10 Wave 11 Wave 12

RDS Recruitment ChainSeed Wave 1 Wave 2 Wave 3 Wave 4 Wave 5 Wave 6 Wave 7 Wave 8 Wave 9 Wave 10 Wave 11 Wave 12 Wave 13

RDS Basicso Is both a sampling and analysis method

A big challenge for those developing RDS is to match the estimators to the sampling design assumptions

o Type of chain referral sampling method to reach hidden populationso Begin with a set of ‘seeds’o Seeds recruit peers, who recruit peers, etc.o Recruits are linked by coupons with unique identifying numberso Recruitment is limited through a minimum set of couponso Incentives provided for completed survey and for each successful recruit

Typical steps in an RDS survey

Enrollment (valid coupon,

screening, consent)

Interview

Optional: Biological testing for

HIV and STI (pre and post

test counseling)

Coupon explanation, provision of

coupons, incentive

Enrollment

Interview

Biological Testing

Coupon Explanation

Incentive Payment

Recruitment Management

Recruitment Management (2)

RDS Assumptionso For most estimators, RDS is heavily assumption driven:

First order Markov Process (i.e., sampling occurs with replacement, sample becomes independent from the seeds)

Respondents know one another as members of the target population (and these are reciprocal)

Respondents are linked by a network composed of a single component Respondents can accurately report their personal network size Peer recruitment is a random selection from the recruiter’s network Each respondent recruits a single peer

RDS required datao Must collect each participant’s social network sizeo Must collect information about who recruited whom

The importance of social networks in RDSo A participant’s social network size sets up the probability of someone being selected into the studyo Must include all of the criteria used in eligibilityo Only members of a participant’s social network should be recruited

into studyo Analysis is based on each social network member having an equal

probability of being included in the study.oAdapt the question for A/KYP

Where to get answers during formative assessment?

◦ Go to A/YKP◦ Find leaders within the A/YKP target population ◦ Talk to organizations working with A/YKP◦ Talk to gate keepers who may be involved with A/YKP

What kinds of information should you get during formative assessment?

o Network structure, characteristics of the populationo Logistics, acceptability, identification of seeds

Selecting ‘good’ seeds to start your study◦ Seeds are non-randomly selected members of the target

population who◦ Are well regarded by their peers◦ Have large social networks and know diverse types of people◦ Can communicate well ◦ Have some interest in the purpose of the survey

Incentives◦ Find out the legal restrictions are about giving incentives to minors

during research◦ Incentives are usually monetary but other types of incentives are fine◦ Find out what the population wants

CouponsoProvide information about the survey

(hours, location, benefits, etc.) oTrack recruitment progress oManage incentive payments oLink information about recruits and their

recruiters through the ID numberingoAllows participants to be anonymous

AnalysisSoftware programs are available for free

Available at www.hpmrg.org Available at www.respondentdrivensampling.org

RDS strengths and challenges

StrengthsRepresentativeness Efficient for hard to reach populationsManuals and tools available Peer recruit peers No mapping neededAllows for anonymityFree computer software available

Challenges Can be difficult to verify group

membership Uses incentives Difficult to deal with selective non-

response bias Analysis is challenged and

challenging Relies on numerous assumptions

Extra formative research Assess social networksRDS logistics and acceptanceIdentify seedsMobility of A/YKP

Simple coupon designs (pictures)Modified network size questionA/YKP appropriate incentivesToys and activities in waiting room (one staff member)What to do with family members/friends who want to go to interview site

Special Issues: RDS among A/YKP

Include social worker with specialization in A/YKP at interview site Conduct community meetings and involvement of organizations, concerned adults who may or may not be parents or guardiansQuestionnaire design-PILOT A/YKP of 14 years interpret questions differently than A/YKP of 21

yearsThink about communications with policeEthical considerations

Special Issues for A/YKP in any sampling method

Resources Johnston, L. G. Introduction to Respondent Driven Sampling. 2013. Geneva, Switzerland. http://applications.emro.who.int/dsaf/EMRPUB_2013_EN_1539.pdf

NetDraw and RDS Analyst Software Manual. http://applications.emro.who.int/dsaf/EMRPUB_2014_EN_1686.pdf?ua=1

www.hpmrg.org (RDS Analyst Software)

www.respondentdrivensampling.org (RDSAT, references, RDS materials)

www.lisagjohnston.com (resource materials, journal articles, etc.)

Young Key Populations at Higher Risk of HIV in Asia and the Pacific: Making the case with Strategic Information. http://www.unicef.org/eapro/Young_key_populations_at_high_risk_of_HIV_in_Asia_Pacific.pdf

Thank You