Kelvin, EA, Mwai, E, Romo, ML, George, G, Govender, K, Mantell, JE, Strauss, M, Nyaga, EN and Odhiambo, JO, 2017. Evaluating oral HIV self-testing to increase HIV testing uptake among truck drivers in Kenya, 3ie Impact Evaluation Report 64. New Delhi: International Initiative for Impact Evaluation (3ie)
Link to Source
The study uses a randomised controlled design to measure the impact of alternative approaches to HIV testing on the uptake of testing amongst Kenyan truck drivers attending a roadside wellness clinic.
High mobility, particularly along popular transportation routes, is highly correlated with greater HIV transmission. Studies have shown that a large percentage of truck drivers, often away from their main partner for long periods, frequent commercial sex hubs along their routes. While HIV testing and counselling (HTC) efforts have been scaled up in Kenya, challenges remain in reaching high-risk populations. This study assesses whether an alternative approach for HTC, HIV oral self-test kits, can better address individual-level barriers to HTC. The study will address important policy questions surrounding the most effective ways to develop specific recommendations for self-administered oral HIV test.
The primary outcome of interest is the uptake of HIV testing in the clinic or with an HIV self-test within a week of enrolment (HIV testing uptake at baseline). The secondary outcome of interest is the uptake of follow-up HIV testing within six months.
The intervention aims to increase HTC uptake among truck drivers, a population at higher risk of acquiring HIV. Over an eight month study period (two months recruitment, six months follow-up), the intervention offered the choice of a standard HIV blood test or a supervised self-administered oral test, or to take an oral HIV self-test kit to perform later on their own. When the self-administered results were ready, participants were given the choice to view them alone or with the healthcare professional. After three months, those allocated to the intervention arm were sent text messages reminding them to get re-tested and that participants could return to any one of North Star Alliance’s clinics within the next three months to pick up a second self-test kit. Those that picked up a kit were asked to send a text message after performing the test, and the study team then attempted to schedule a phone interview. Participants in both the treatment and control groups were sent text message reminders of the importance of follow-up testing at three months post baseline.
Theory of change
This is based on prospect theory, which posits that the likelihood of an outcome is based on a cost-benefit analysis. Barriers to HIV testing such as time, financial costs, and potential social costs are weighed against the benefits of knowing one’s HIV status and overall health status. The researchers posit that offering an additional option of self-testing or self-administered testing will reduce at least some of these barriers or costs, such as not wanting a blood test, not wanting a health provider perform the test, or not wanting anyone else, including a health provider, to know the results. Thus, the reduction and barriers, and therefore costs, will increase uptake of HIV testing. Uptake of the second test, in six months could also be affected by a perception of convenience—it may be faster to pick up a test and do it on their own than to wait and get the test done at a clinic, which would reduce the cost of waiting.
This study randomly assigned 305 truck drivers who visited North Star Alliance roadside wellness clinics into two study arms (150 in intervention and 155 in control). Eligible truck drivers spoke English or Kiswahili, were 18 years of age or older, primarily residing in Kenya, and self-reported being HIV-negative or were unaware of their status, and had not tested in the last three months. The control group was offered a standard provider administered HIV blood test. Researchers used a difference-in-difference logistic regression to estimate intention-to-treat effects. Bivariate and multi-variable analyses were performed to disaggregate by sub-groups and control for confounders. Authors used in-depth qualitative interviews with 30 drivers to assess HIV risk perceptions and experiences with testing. They used a discrete choice analysis to assess preferences around HIV testing options. Quantitative and qualitative data were collected at enrolment or baseline, and after re-offer, three to six months after enrolment.
The choice of a self-administered oral HIV test had one and a half times higher odds of accepting an HIV compared to the control arm (p=0.189). An additional 11 participants in the choice arm accepted HIV testing when offered a kit to use outside the clinic after refusing testing in the clinic. Including these 11, around 87 per cent participants in the intervention (compared to 73% in the control) accepted HIV testing (OR = 2.8, CI: 1.5 – 5.4, p = 0.002). However, after six months, only 56 per cent in both arms returned for an HIV test.
Cost efficiency analysis
Cost were calculated based on an estimate of 14 per cent increase in testing, and included staff training, counselling time, kits costs, phone service charges, and accounted for no counselling for those who took the kit outside the clinic. Cost-effectiveness is highly dependent upon the cost of the test kits and effectiveness of the intervention, which varied between baseline and re-offer, and did not account for the option of offering the kits for take-away immediately, as opposed to after refusal of in-clinic testing.
Implications for implementers
Rolling out a new intervention requires intensive supervision of staff and continued training. Existing electronic medical records could be used for tracking self-testers if there is consistent and reliable internet connection at the clinics. Truck drivers are ready for HIV self-testing, but careful rollout is recommended to ensure that high quality care and trust of clients is maintained.
Implications for policy and practice
Oral HIV self-testing should be considered a complimentary option and not a replacement for standard testing. It is critical to design appropriate information and counselling resources to accommodate the new model. More research is needed to establish what counselling is needed and desired, and to determine how and when to provide it so that HIV self-testers are more likely to link to care.
Self-administered HIV testing is acceptable and feasible. Initial rollout should provide opportunities for testers to ask questions and ensure adequate staff supervision and training on an ongoing basis. More information is needed about the need for and mechanisms to deliver counselling, and ensure linkage to care