Robustness of insect repellent efficacy research: exploring the impacts of sampling design and participant heterogeneity

Abstract Humans and mosquitoes have complex interactions that offer numerous opportunities for exploring what attracts mosquitoes to humans, knowledge that can inform mosquito repellent strategies. It is common for the complete protection times (CPTs) of topical mosquito repellents to be evaluated using the arm-in-cage (AIC) test, during which participants insert their forearms into mosquito-infested cages. World Health Organization (WHO), Environmental Protection Agency (EPA), and European Chemical Agency (ECHA) guidelines specify that AIC testing must utilise at least 10 study participants; smaller sample sizes are thought to produce inaccurate CPT estimates. This study compared two sampling designs: a standard AIC test with 10 participants (1 measurement each) and a repeated-measures AIC test with 5 participants (2 measurements each). Both were repeated five times to assess reproducibility and compare CPT estimates. To investigate within-participant variability, we collected additional data from 9 of the 10 participants; overall, we obtained 36 replicates for each of these participants. We observed pronounced variability in CPT both within and among study participants, highlighting that attractiveness to mosquitoes can vary as much within a single individual as among 10 different individuals. The sampling designs produced consistent results and yielded statistically equivalent CPT estimates. This study highlights how CPT estimates may be influenced by within-participant variability, which appears to be comparable in magnitude to among-participant variability. Future research should explore this issue further and focus on the factors affecting the intrinsic variability displayed by a given individual.