Design matters: structural variation in experimental huts significantly alters entomological endpoints in insecticide treated net (ITN) evaluation –a comparative experimental hut trial using a latin square design in Tanzania and Côte d’Ivoire

Experimental huts are the gold standard for the semi-field evaluation of insecticide-treated nets (ITNs) but the design of the huts used varies markedly between sites. Structural differences may influence mosquito behaviour – and, consequently, the measured efficacy of ITNs. This raises concerns about the comparability, validity, and policy relevance of trial outcomes. We assessed the influence of four commonly used hut designs on entomological endpoints central to ITN evaluations. A 126-night comparative trial was conducted in Tanzania, using 18 experimental huts in two contiguous 9 × 9 Latin squares. Four hut types–East African, West African, Ifakara, and Rapley–were tested with PRONet Duo, Interceptor ® G2, MAGNet ® , and untreated nets. Sleepers and treatments were rotated throughout the trial. The primary outcome was 72-h mortality (M72); secondary outcomes were mosquito density, blood-feeding success, and exiting. Endpoint estimates for each hut type were compared using East African huts as the reference. Non-inferiority analysis of PRONet Duo in each hut type were compared with findings from a sister trial conducted exclusively in West African huts in West Africa. A total of 22611 Anopheles gambiae s.l. (99.9% An. arabiensis ) were collected. Rapley and Ifakara huts captured over six times more mosquitoes than East African huts (rate ratio [RR] = 6.50 and 6.22, respectively; both p < 0.0001), whereas West African huts caught 88% fewer mosquitoes (RR = 0.12, p < 0.0001). West African huts recorded the highest mortality (70%) and blood feeding (15%) whilst the other hut types recorded 40–46% mortality and 3–8% blood feeding. All hut designs and the West African sister trial detected the same direction of effect between PRONet Duo and Interceptor ® G2 and predicted non-inferiority of PRONet Duo, although smaller sample sizes limited statistical certainty in East African and West African huts in Tanzania. Experimental hut design substantially affected key entomological outcomes, influencing both the magnitude of measured efficacy of ITNs and the statistical power of trials. Whilst relative product performance was broadly consistent across designs, absolute values varied widely, underscoring the need for greater standardisation of hut architecture in ITN evaluations. Harmonising experimental hut designs and reporting standards would strengthen cross-site comparability, improving the reliability of vector control product assessments, and enhancing the evidence base for global malaria control policy.