Background - In 2003, the DGS requested the InVS to assess the health impact of environmental exposure to asbestos fibers among populations living near former industrial sites where asbestos was mined or processed, as well as near natural outcrops. To respond to this request, the InVS proposes to conduct two studies: a case-control study to compare mesothelioma risks based on past environmental exposure among populations; and a study to estimate current exposure among residents living near former industrial sites where asbestos was mined or processed and natural asbestos-bearing sites, in order to predict future consequences. Objective - The year 2005 was devoted to determining methodological choices, analyzing the quality of available data, and studying data collection methods. It also allowed for an assessment of technical feasibility, necessary resources, and the difficulties inherent in each stage. Method - Methodologically, the key issue concerns the estimation of population exposure. Data collection involves the inventory of industrial sites and the mapping of natural geological outcrops. Regarding exposure, for the case-control study, the goal is to estimate the retrospective environmental exposure of cases and controls. Drawing on information available from the National Mesothelioma Surveillance Program, the developed method relies on the construction of an exposure score that combines the distance between individuals’ various residences and the site’s location; a Geographic Information System (GIS) is used to objectively measure these distances. In a second step, using site characteristics, an assessment of exposure intensity will be integrated into the previous score. For populations living near natural and industrial sites, the goal is to estimate their current exposure based on data describing the characteristics of the sites, their environment, and the populations present. This data enables an assessment of the exposure level of the populations living near each site. The values provided are established through an assessment, which involves compiling an information grid and questionnaires; this must be carried out by experts from BRGM and Lepi for each site. A decision tree will integrate all the acquired information to enable the categorization of sites according to their exposure risk to the population. Comparing these initial estimates with the results of measurements from a metrological campaign conducted by Lepi will confirm or refute the validity of the method. Result—The inventory of industrial sites lists 1,119 shipyards and 553 other sites; the survey of natural sites identified 19 sites of former operations and confirmed asbestos outcrops. For the control study, the addresses of PNSM subjects were retrieved for schools and workplaces, then geocoded. Non-occupational exposures recorded by the PNSM will be re-evaluated for women, as the data is not reported consistently. The investigation of industrial site characteristics yielded information only on the type of manufacturing, the site’s period of operation, and the number of employees, and this data was inconsistent. The estimate of exposure intensity was initially based on whether or not the sites appeared on the list of DRT decrees. Regarding the availability of data for assessing current exposure, the diagnostic tool used to develop the grid and questionnaires was tested on natural sites. These tools enabled the collection of necessary data on site characteristics, their environment, and the local populations living near the sites. However, regarding industrial sites, access difficulties have prevented the same approach from being carried out at this time. Ultimately, the information gathered through this diagnostic process will be integrated into a decision tree that will allow for the assessment of exposure levels and risk, and thus the health implications for nearby populations. This decision tree will address the challenges encountered regarding the heterogeneity of parameters related to the assessment. Discussion and conclusion - The case-control study is feasible and can be extended to mainland France. The results of the case-control analysis will be available by the end of 2006. The exposure score based on distance from the site and duration of residence near the site is an appropriate method even when data are missing. However, the difficulties encountered in determining the historical characteristics of former industrial sites necessitate further refinement of the construction of exposure intensity to better define the strength of the tested epidemiological relationship. Regarding other domestic and occupational exposures, it is difficult to account for them because there is no established list of buildings containing asbestos. For women, it is necessary to reassess their non-occupational exposure. Finally, since the PNSM is not present in all departments—particularly those where natural deposits exist—only populations living near industrial sites will be analyzed. It would be desirable to extend the PNSM to departments where the presence of natural outcrops of asbestos-bearing rock is confirmed, particularly in Corsica. Regarding the study to estimate current exposures, most of the data evaluated during the assessment, although easy to obtain, exhibit significant spatial and/or temporal variability. The extent of this variability could not be assessed through field expertise. The analysis will initially focus solely on natural sites due to difficulties in accessing industrial sites. The method currently being developed, based on a decision tree that integrates all the information obtained to estimate population exposures, is rarely used and still needs to be adapted for the assessment of asbestos fiber exposure. This method should lead to estimates of current exposure levels. The metrological surveys, conducted during the summer of 2006 at two sites, are intended to quantitatively assess population exposure levels at the two sites most at risk and to confirm or refute the qualitative approach. The results of this study on natural sites will be available in early 2007.