Background A mixed methods study was conducted to look at the magnitude of residual malaria transmission (RMT) and factors contributing to low (< 1% prevalence), but sustained transmission in rural communities on the Thai–Myanmar border. Methods A cross-sectional behaviour and net survey, observational surveys and entomological collections in both villages and forested farm huts frequented by community members for subsistence farming practices were conducted. Results Community members frequently stayed overnight at subsistence farm huts or in the forest. Entomological collections showed higher biting rates of primary vectors in forested farm hut sites and in a more forested village setting compared to a village with clustered housing and better infrastructure. Despite high levels of outdoor biting, biting exposure occurred predominantly indoors, particularly for non-users of long-lasting insecticidal nets (LLINs). Risk of biting exposure was exacerbated by sub-optimal coverage of LLINs, particularly in subsistence farm huts and in the forest. Furthermore, early waking hours when people had left the safety of their nets coincided with peaks in biting in later morning hours. Conclusions Entomological and epidemiological findings suggest drivers and modulators of sustained infection prevalence in the area to be: higher mosquito abundance in forested areas where LLINs were used less frequently or could not be used; late sleeping and waking times coinciding with peak biting hours; feeding preferences of Anopheles taking them away from contact with LLIN and indoor residual spraying (IRS), e.g. exophagy and zoophagy; non-use of LLIN and use of damaged/torn LLIN; high population movement across the border and into forested areas thereby increasing risk of exposure, decreasing use of protection and limiting access to healthcare; and, Plasmodium vivax predominance resulting in relapse(s) of previous infection. The findings highlight gaps in current intervention coverage beyond the village setting. Electronic supplementary material The online version of this article (10.1186/s12936-019-2852-5) contains supplementary material, which is available to authorized users.

No abstract