Recruitment through social networks is an efficient way to increase demand for HIV testing among high-risk groups

Strategy raises number of people with HIV tested – but in this study, several had been previously diagnosed
This article is more than 11 years old. Click here for more recent articles on this topic

A pilot project which used a peer recruitment strategy, financial incentives and the involvement of people living with HIV has demonstrated that this approach can increase demand for HIV testing services and efficiently identify people with HIV, report Sandra McCoy and colleagues in the Journal of Acquired Immune Deficiency Syndromes.

But the project was not as successful as a Centers for Disease Prevention and Control (CDC) study which demonstrated the viability of this approach four years ago. In particular, a significant number of those testing HIV positive had, in fact, been previously diagnosed – perhaps suggesting that the $10 gift card exerted too powerful a pull in the low-income community where the project was conducted.

Four different community organisations, offering HIV testing at mobile or storefront locations, implemented the pilot in Oakland, California in 2011. Just across the water from San Francisco, Oakland is a much poorer city in which 46% of late HIV diagnoses are in African Americans. As older age groups are more likely to have late diagnosis, the pilot focused on people aged 30 to 60 who identified as African American or Black.

Glossary

pilot study

Small-scale, preliminary study, conducted to evaluate feasibility, time, cost, adverse events, and improve upon the design of a future full-scale research project.

 

stigma

Social attitudes that suggest that having a particular illness or being in a particular situation is something to be ashamed of. Stigma can be questioned and challenged.

anxiety

A feeling of unease, such as worry or fear, which can be mild or severe. Anxiety disorders are conditions in which anxiety dominates a person’s life or is experienced in particular situations.

statistical significance

Statistical tests are used to judge whether the results of a study could be due to chance and would not be confirmed if the study was repeated. If result is probably not due to chance, the results are ‘statistically significant’. 

The pilot began with the recruitment of 48 initial participants. These individuals were provided with HIV testing and counselling themselves (unless they already had diagnosed HIV) and were then asked to recruit up to three other people to the study. These individuals were provided with HIV testing and counselling, then asked to recruit three more people, and so on. This peer recruitment approach is known as respondent driven sampling (RDS).

Through successive waves of recruitment, a further 243 people took part. Only 45% of the participants recruited someone else.

Financial incentives – described below – were offered to all.

Participants were predominantly male, heterosexual and poor – three-quarters earned less than $10,000 a year and half had been homeless in the last year. The average age was 47.

When recruiting, participants were encouraged to approach people who might need to be connected to services. (The researchers deliberately avoided targeting recruitment on the basis of sexual or drug-using behaviours, which may be stigmatised and undisclosed.)

Participants were successful in recruiting from the priority groups that had been specified – 23% of new recruits had never been tested for HIV; 72% had not been tested in the last year; 81% had never used services at the organisation providing testing; and 10% had been in prison in the last year.

A key outcome of interest was the number of participants testing HIV positive. As a benchmark, at other HIV testing services provided by the four community organisations, 16 of 2471 tests were positive (0.6%). This corresponds to needing to test 154 people to find an individual with HIV.

Among those testing within the pilot (excluding the initial participants, some of whom were known to have diagnosed HIV), 9 of 243 individuals were diagnosed with HIV (3.7%). This corresponds to needing to test 27 people to find an individual with HIV.

This would have seemed to be a particularly effective way of identifying people with undiagnosed HIV had the researchers not checked the public health department’s surveillance records. This revealed that seven of the nine who’d tested positive were in fact previously diagnosed, including four who had dropped out of medical care. Project staff ensured they reconnected with care, “an unintended outcome of our study that nevertheless has public health value”, say the authors.

However, with only two people testing positive for the first time, prevalence comes down to 0.8%. The number needed to test in order to find a person with HIV is 122, which is comparable to the figure for standard interventions.

The study had deliberately recruited some initial participants who already had diagnosed HIV (10 of the 48). Other studies have shown that people with HIV are more successful at recruiting people in need of testing, and this appeared to be the case in this study, although the difference was not statistically significant. In networks begun by a person with HIV, 79% of recruits were considered to have high-risk behaviours or were HIV positive, compared to 67% in networks begun by an HIV-negative person. 

Financial incentives

As is already common practice in Oakland, financial incentives were offered. The economic theory behind this is that individuals often have “present-biased preferences” – when deciding whether to do something, the immediate costs and benefits weigh more heavily than the future costs and benefits. Taking an HIV test may have immediate costs (time, transport, anxiety, stigma), while the possible benefits (better long-term health) are delayed.

“The use of financial incentives adds an immediate benefit to counteract present costs and may therefore effectively change behaviour,” say the researchers.

To test two different approaches to providing financial incentives, there was an element of randomisation to the study. In the first arm, there was a relatively simple system of incentives – participants got $20 for each new recruit and $10 for taking an HIV test themselves.

The second arm was more complex, with bonuses that were conditional on recruiting individuals from the priority groups described above. For example, there would be $10 for any new recruit, plus a $10 bonus if he or she had never taken an HIV test before.

As participants were only allowed to recruit up to three other participants, and many did not recruit anyone at all, the average total pay out was low – $17.86 and $14.62 in the first and second arms respectively. The money came in the form of gift cards which could be used at local shops.

Half the participants said that the financial incentives were one of the reasons for taking part.

But there were no differences between the two arms in the number of recruits from a high-priority group, the number who reported high-risk behaviours or who were HIV positive.

The authors therefore concluded that the simpler payment system was as efficient as the more complex system (as well as being easier to implement).

More generally: “Our findings add to the growing evidence base highlighting the value of social network approaches as an efficient and non-stigmatizing strategy to increase demand for HIV testing and counseling, identify people living with HIV infection who are unaware of their status, and as a potential way to recapture PLHIV who are out of care.”

References

McCoy SI et al. Improving the Efficiency of HIV Testing With Peer Recruitment, Financial Incentives, and the Involvement of Persons Living with HIV Infection. Journal of Acquired Immune Deficiency Syndromes, published ahead of print, 2013: DOI: 10.1097/QAI.0b013e31828a7629. (Abstract here)