Delphine Jezewski-Serra

O48-2 Exposure assessment to engineered nanomaterials within the french epinano program: inter-method reliability study

Inter-method reliability is a measure of the ability of two different methods which evaluate the same underlying exposure to yield similar results on the same subjects. In this study we compared an observational “EpiNano-method” developed for assessing exposure to nano-objects and their agglomerates and aggregates (NOAA) in workplaces within the framework of the French epidemiological surveillance program EpiNano with a more accurate but more expensive and time-consuming “Reference-method”. Methods EpiNano-method consists in an onsite visit of facilities and an observation of workstations where NOAA are present using a standardised tool, the Onsite technical logbook. Reference-method involved an onsite visit, interviews with operators and supervisors, analysis of the available exposure measurements data, a debriefing meeting and an expert’s report to validate results through consensus. Seven workstations which synthesise and functionalize carbon nanotubes were assessed by Reference-method and by EpiNano-method implemented by three evaluators. All statistics were calculated based on pairwise comparisons of the evaluator’s respective results with the results from the Reference-method by using mainly Spearman correlation coefficients (ρ) and weighted Fleiss kappa (kF). Results The prediction of exposure to NOAA using EpiNano-method exhibited substantial agreement with that of the reference method ((kF = 0.70, sensitivity = 0.88, specificity = 0.92). The best reliability of quantitative exposure scores was found for the EpiNano-method results of Evaluator#1, when comparing with the Reference-method (ρ = 0.75). The correlation was positive but of moderate strength (ρ = 0.57) for Evaluator#2 and negative for Evaluator#3. When comparing the semi-quantitative exposure potential variable almost perfect correlation was found; all but one workstation were classified in the first category corresponding to “Negligible to low” exposure potential. Conclusion This study allowed a first validation of the EpiNano-method. However, it should be confirmed by further comparison with more robust quantitative exposure measurement data, available for a larger workstation number, with more contrasting exposure conditions and various types of NOAA.