Given the hidden nature of the IDU population, it is extremely challenging for service providers to identify and reach them for services. Even in Delhi, where a large proportion lives and injects in public, there is still a sizable proportion that is not visible and cannot be reached by traditional outreach methods. PDI approaches have been used to reach members of hidden populations who would otherwise not be reached by traditional outreach methods [10–13]. For this intervention study, we used peer referral along with ORW referral and walk-in method for recruiting IDUs. We ultimately found that for this population of IDUs in Delhi, ORWs were the most successful in recruiting IDUs for the study. However, having a mix of recruitment methods was advantageous in diversifying the sample enrolled into the intervention as there were differences in background and risk profiles of IDUs by recruitment method.
The relative success of ORW recruitment method over the other two recruitment strategies may be, in part, due to the level of training the ORWs received to motivate IDUs to participate in the study. ORWs were likely able to better motivate the IDUs and convince IDUs of the benefits of participation as opposed to peer recruiters who would not have the in-depth understanding of the study and the intervention as the ORWs. The majority of the ORWs in this study had prior experience of working with IDUs, and hence they not only knew the IDU community and hotspots very well, but they had already been sensitized to working with this population and knew the issues that may affect IDUs’ willingness to participate. Lastly, the IDUs that ORWs reach may be considered ‘low hanging fruits’ as they are easily identifiable and congregate openly in public places.
Although the ORW recruitment method was highly successful, it was also more resource intensive than the other two methods. The number of IDUs at a hotspot affected the cost of ORW recruitment since travel costs increased per IDU recruited from small or very small hotspots. On the other hand, in the case of larger hotspots, it proved to be more time-intensive as it was challenging to identify IDUs who had not yet registered in the intervention. The higher cost, however, should not prohibit programmers from using the ORW recruitment method, as outreach is an integral component of and highly effective strategy for harm reduction for IDUs [7–9]. Although in this study, ORWs were specifically instructed not to impart any harm reduction messages to IDUs as our goal was to have a pure baseline behavioral survey, future programs should use ORWs for both recruitment as well as harm reduction counseling.
The three recruitment methods yielded very different types of IDUs, which is useful in expanding coverage of programs. While ORWs were more likely to recruit street-based IDUs, the peer referral method was able to better access IDUs who were home-based and living with family or relatives. Further, we found that walk-in IDUs had higher risky injection behaviors. Walk-ins were also more likely to report previous HIV positive test, which may indicate their desire for services. Additionally, they were more likely to have undiagnosed HIV infection, suggesting a higher perceived risk and a desire for services.
The inclusion of peer referral method assisted in reaching many home-based IDUs who may not have been recruited if there was only ORW referral method [10, 15]. The peer referral approach is more private and anonymous compared to being recruited by an ORW. Residential communities were reluctant to allow ORWs to visit their areas as persons interacting with them would be labeled as IDUs, thus limiting their access to home-based IDUs. Further, we learned that street-based IDUs were not as effective as home-based IDUs in recruiting their peers. Their unstable lifestyle and chaotic living environments (i.e., moving and working at different spots based on drug availability and injecting and sleeping in secluded spots such as near a drain or public toilet) likely make it challenging for them to keep track of their recruitment coupons; many IDUs mentioned that their recruitment coupons had got lost. Given that home-based IDUs were better at recruiting their peers compared to street-based IDUs, this presents an opportunity to take advantage of home-based IDUs to recruit their peers for interventions.
One of the challenges faced in this study with regard to peer referral recruitment was identifying an acceptable level of reimbursement for recruiting their peers. Despite presenting to the ethical committee of NACO of the need and ethical arguments (i.e., justice, beneficence, and respect) for using financial reimbursement for recruitment , the committee ultimately did not allow financial remuneration for recruitment fearing money would further encourage their drug use. Only the modest reimbursement of Rs. 40 (US $0.80) for travel and time irrespective of the mode of recruitment was allowed as financial reimbursement. However, as noted, less than half the recruiters actually claimed their food coupon, which suggests that this was not a sufficiently appealing compensation, which may have resulted in slow and inefficient recruitment. In a study we conducted in 2006–07 with IDUs in Delhi in which only respondent-driven sampling was used, money was given as compensation both for participation in the study as well as for successful recruitment. Unclaimed compensation by recruiters were minimal, and the study successfully recruited 800 IDUs from Delhi within a period of four months . Thus, the provision of sufficiently appealing compensation is crucial to the success of using peer referral recruitment. Financial remuneration for both participation as well as recruitment has been successfully used in many studies that use RDS to recruit participants. Further, there are ethical arguments for providing such remunerations (i.e., respect for the participant’s time and effort) and safeguards can be placed so that remuneration can be provided in an ethically sound manner (such as placing quotas on number of recruits and keeping remuneration modest) .
Some limitations should be kept in mind. The study was not intended for comparing the feasibility and effectiveness of recruiting IDUs through different recruitment methods. So we cannot make conclusions about which recruitment method is more effective than the other methods. The study could only describe the process and the operational details of the three recruitment methods. Second, many IDUs indicated that although a peer had referred them to study, they no longer had the coupon with them. Because IDUs were counted as having been recruited by a peer only if they arrived at the study site with the peer recruitment coupon, some IDUs may have been misclassified as walk-ins and ORW-referred IDUs. Lastly, given the large number of IDUs recruited by ORWs, it is also possible that the IDUs who would have been accessible through the PDI strategy had already been approached or knew that they would eventually be approached by an ORW. Unfortunately, we do not have a record of refusals and acceptances of the ORW approach.