peer reviewed ; A human biomonitoring study was carried out in 2015 within an adult population living in Liege (Belgium). Some phthalate metabolites and parabens were measured in the urine of 252 participants, and information were collected about their food habits, life styles and home environment to identify some predictors of exposure. Concomitantly, an awareness campaign was initiated by the Provincial Authorities of Liege and spread over 2 years. Three years later (2018), 92 of the initial participants provided again urine samples, and the levels of phthalate metabolites, phthalate substitute (DINCH), parabens, bisphenol-A and bisphenol alternatives (bisphenol-S, -F, -Z, -P) were determined and compared to those obtained in 2015 to assess time trends. In 2015, methyl- and ethylparaben were the most abundant parabens (P50 = 9.12 μg/L and 1.1 μg/L respectively), while propyl- and butylparaben were sparsely detected. Except for mono-2-ethylhexyl phthalate and 6-OH-mono-propyl-heptyl phthalate, all other targeted phthalate metabolites were positively quantified in most of the urine samples (between 89 and 98%) with median concentrations ranging between 2.7 μg/L and 21.3 μg/L depending on the metabolite. The multivariate regression models highlighted some significant associations between urinary phthalate metabolite or paraben levels and age, rural or urban character of the residence place, and the use of some personal care products. However, all determination coefficients were weak meaning that the usual covariates included in the models only explained a small part of the variance. Between 2015 and 2018, levels of parabens and phthalate metabolites significantly decreased (from 1.3 to 2.5 fold) except for monoethyl phthalate which seemed to remain quite constant. Contrariwise, all bisphenol alternatives and DINCH metabolites were measured in higher concentrations in 2018 vs 2015 while BPA levels did not differ significantly. However, it was not feasible to unequivocally highlight an impact of the awareness ...