ABSTRACT BACKGROUND፡ The ability of mercury to deposit throughout the body and alter a wide range of molecular and cellular pathways results in a polymorphic and complex clinical phenotype with over 250 possible symptoms. However, some of them are recurrently cited as evoking chronic mercury poisoning. In this light, dentists users of dental amalgams are chronically exposed to mercury so that in-depth epidemiological investigations and adapted statistical methods are required to highlight adverse effects of this exposure.METHODS: In order to study the health impact of the occupational mercury exposure in a population of liberal dentists practicing in two Moroccan regions, a list of eighteen subjective symptoms commonly associated with micro-hydrargyrism was drawn up. Then,  seven statisctical models adapted to count data were fitted. Finally, three methods were used to compare their relative performance in order to choose the most appropriate one.RESULTS: The adopted logical path, from the response variable selection till models’ comparison, led us to lean towards quasi-Poisson regression as the best way to predict the number of symptoms declared by dentists according to mercury exposure.CONCLUSIONS: Interpretation of the selected model allowed us to conclude that the reduction of dental amalgam use allows the reduction of subjective symptoms related to mercury exposure.KEYWORDS: Mercury exposure, Dental amalgam, Dentist, count data, Generalized linear  models, Over-dispersion, Models selection