Introduction Providing HIV testing to truck drivers in Africa is crucial but challenging. The introduction of HIV self-testing provides expanded service delivery options for clients, potentially increasing demand for services and expanding coverage – especially important for high-risk and difficult-to-reach populations. This study examines the preferences regarding HIV testing service delivery models, among long distance truck drivers in an effort to identify testing services that would appeal to this population.

Methods Using a discrete choice experiment, this study examines the drivers of choice regarding HIV counselling and testing among 305 truck drivers recruited from two roadside wellness clinics along major trucking routes in Kenya. Participants made trade-offs between characteristics of HIV testing service delivery models by making hypothetical choices in a series of paired HIV testing scenarios. Conditional logit models were used to identify the HIV testing characteristics driving the selection of preferred scenarios, and determine whether preferences interact with individual characteristics – especially HIV testing history.

Results Participants preferred free, provider-administered HIV testing at a roadside clinic, using a finger-prick test, with in-person counselling, undertaken in the shortest possible time. The strongest driver of choice was the cost of the test. Those who had never tested previously preferred oral testing and telephonic counselling, while those who were not regular testers favoured clinic based- over self-testing.

Conclusion The results of this study indicate that for the majority of participants – most of whom had tested before – the existing services offered at roadside clinics were the preferred service delivery model. The introduction of oral self-testing increases the options available to truck drivers and may even improve testing uptake for some, especially among those who have never tested before. However, these findings suggest the impact on HIV testing uptake of introducing oral self-testing may be limited in this population.