Behavioural Science Interventions for HIV Prevention and Treatment

Image: Logo of the PROTECT study. Designed by Leah de Jager.

Human behaviour is an important determinant of health outcomes, and poses a ‘last mile’ challenge to ensuring the effectiveness of existing health services and maximizing the impact of available resources. In Africa, approximately 25% of people living with HIV are either unaware of their status or not taking treatment regularly, and high rates of new HIV infections persist in Eastern and Southern Africa: approximately 500,000 every year. How people behave matters, and innovative, behaviourally-informed programmes and interventions are needed. In 2023, two new grant awards provide SALDRU researcher Brendan Maughan-Brown the opportunity to expand the science on low-cost, light-touch behavioural solutions (known as ‘nudges’) that increase the uptake of health services and promote behaviour change more generally.

The National Institutes of Health (NIH) funded PROTECT Study: Protecting against HIV vaccine misinformation in South Africa, is co-led by  Prof. Alison Buttenheim (University of Pennsylvania), Prof. Harsha Thirumurthy (University of Pennsylvania), together with Brendan Maughan-Brown at SALDRU. In this study, we will test messages that work like an ‘information vaccine’, building a resistance to misinformation that people may hear about the HIV vaccine. Results from this study will help inform how to communicate about a future HIV vaccine in South Africa to ensure that groups at high risk of HIV infection can benefit from the vaccine’s protection.

The study is a randomized controlled trial of inoculation messages with adolescent girls and young women (AGYW) in South Africa, a population highly vulnerable to both HIV infection and to vaccine misinformation. Behavioural economic insights will be applied to ‘boost’ the efficacy of inoculation messages, both at the time of message exposure and for a more durable effect. For example, priming an identity as a misinformation-spotting ‘Vaccine Truth Champion’ may increase engagement with specific inoculation messages and may increase generalized skills in recognising misinformation in the future. Offering tokens (e.g., stickers, badges, or bracelets) related to the identity and inviting a commitment to be a ‘Vaccine Truth Champion’ in the future may also increase the effectiveness of inoculation messages.

Mirroring a vaccine trial, this trial will evaluate the efficacy, safety, durability, and generalised ‘immunity’ effects of inoculation messages that are boosted with behavioural approaches. We will also evaluate differential responses to the messages by important subgroups of AGYW, like individuals with general vaccine hesitancy, to establish how to tailor-generate messages for key populations in the future. This proof-of-concept project has the potential to identify innovative communication strategies to build resistance to emerging and evolving HIV vaccine misinformation. Results from the study will advance the science of HIV vaccine demand creation and inoculation-theory based approaches to vaccine communication.

The second grant, awarded by the Bill & Melinda Gates Foundation, sees the continuation of Indlela: Behavioural Insights for Better Health for another four years. Indlela (“the path” in Zulu) is a behavioural nudge unit based at the Health Economics and Epidemiology Research Office (HE2RO) at the University of The Witwatersrand, in collaboration with the University of Pennsylvania, Boston University, and the University of Cape Town (SALDRU). It was established in 2020 to build behavioural science capacity in South Africa and design behavioural interventions that improve health outcomes, with a particular focus on HIV.

Between 2020-2023, the Indlela team worked closely with many organisations implementing HIV prevention and treatment services in South Africa to co-develop nudges that address behavioural challenges to the success of HIV programmes, such as low adherence to HIV treatment. Wherever possible we concurrently developed rapid evaluations on the impact of these interventions, leveraging administrative data and behavioural outcomes (rather than biological outcomes that can take longer to collect) to enable fieldwork and analysis to be completed in approximately 4-8 months.

In one study, we partnered with Right to Care, to design recruitment forms for voluntary medical male circumcision, an effective intervention to reduce the risk of HIV infection amongst men. Using behavioural economics principles we adapted the visual and text content of the standard of care form, which provided general information on the health benefits of circumcision. We found a 1.3 percentage point increase in contact details being submitted for a follow-up call when the form was based on the foot-in-the-door technique: a strategy based on creating the opportunity for smaller, more manageable steps to be taken before the larger decision of committing to something (medical circumcision in this case). While the absolute percentage-point change was small, the relative change (±20%) is equivalent to thousands of men per year expressing interest in medical male circumcision who would otherwise not have done so. Read this policy brief for further details.

In another study, we designed an intervention to reduce HIV treatment interruptions, which are a contributing factor to treatment resistance, onward HIV transmission, opportunistic infections, and mortality. To prompt HIV care recipients to return to the clinic we designed text messages leveraging the ‘fresh start effect’, a behavioural principle that uses temporal landmarks, such as the start of the year, a holiday, or a birthday, to motivate aspirational behaviours. The temporal landmark signifies a new time period and a clean slate going forward. Results showed that low-cost text messages sent around a ‘fresh start’ date increased the likelihood that patients who missed appointments returned to care. Read this policy brief for further details.

The value of nudges for health behaviour change is an ongoing question for policy makers. This is partly because the absolute effect size from such interventions is typically small – there are few big shiny results. This is not surprising given that human behaviour is complex and influenced by multiple forces, and the behaviours that we want to nudge are generally behaviours that people want to enact, but for a myriad of reasons have not. The size of the impact of the interventions that we are designing and testing generally align with those of other nudge units globally. The power of such interventions is at scale. When carefully designed to be low-cost and implemented at scale, these interventions can be highly cost effective. The impact of our voluntary medical male circumcision intervention (above) is a good example, with the potential for thousands more men in South Africa engaging with counsellors on circumcision due to an intervention of negligible cost.

Given small effect sizes from successful nudges, we need to design for success. Nudges need to be precisely targeted based on a thorough understanding of the behavioural challenges – without first identifying the reasons why people do not engage in a specific health behaviour, the risk of failure can be high. We also need to design for implementation and engage stakeholders up front. With careful design, future nudges will continue to build on the array of low-cost interventions that have successfully resulted in individual behaviour change and improved health care delivery overall.

As Indlela moves into its next phase, a question that we are asking ourselves is whether we can move the application of behavioural economics upstream from individuals so that policies, structures and systems are designed to reduce behavioural barriers and create contexts in which the actions that people want to enact are easier to make happen. For example, simplifying or removing steps that a policy requires for product/service uptake, could reduce the friction that creates the behavioural barrier that individual nudges are subsequently designed to overcome. We look forward to expanding our work with policy makers and implementing partners to co-design and test both individual-level interventions and system-level interventions that ultimately contribute to improved well-being.