Evaluation of the TNO "Stoffenmanager" with MEGA exposure data

Status:

completed 12/2009

Aims:

The "Stoffenmanager" is an internet-based tool for the evaluation of risks arising from working with hazardous substances. The program is available in Dutch and also in English (since 2007). Both language versions can be used by companies on the Internet for free. The program contains tools for determining a risk category ("control banding") and for quantitative estimation of exposure. The model, which was validated during the project, is aimed at quantitative estimation of the level of exposure. The project examined firstly how the data from the BGIA's MEGA database of measured data relating to exposure to hazardous substances at the workplace could be used for validation, and secondly, in what working areas/during what tasks deviations occurred between the modelled exposure height and the measurement values.

Activities/Methods:

Data from the BGIA's MEGA exposure database were used in order to put the validation of the "Stoffenmanager" exposure model on a broader basis following an initial review with TNO's own data. For this purpose, the variables from MEGA were assigned to the model parameters of the "Stoffenmanager". For the purpose of translation, the individual values from the coding lists in MEGA (for example for the information on general ventilation or local exhaust ventilation) were assigned to the values in the coding lists of the "Stoffenmanager". For a second part of the variables, it was necessary to link several variables from MEGA in accordance with fixed rules in order to obtain a model parameter, or to evaluate free-text fields. Following the translation of the variables, the dust exposure (inhalable fraction) was modelled for individual groups of working areas (such as mixing, sawing, and grinding). The exposure height which was estimated by the model was compared with the measured value for assessment of the quality of the model.

Results:

In the scenarios involving the filling and mixing of dust-raising products and cleaning work, the exposure is underestimated by the model, with close correlation between measured and modelled exposure levels (n=15; r=0.89; relative bias -33%). High correlation coefficients are also observed in the scenarios for the machining of wood and stone; in this case, however, the model yields an overestimation of the exposure (n=22; r=0.84; relative bias 68%). These results of the validation project were presented to the scientific community at the "X2009 - Sixth International Conference on Innovations in Exposure Assessment" in Boston, where they met with a positive reception. Validation also serves as preparation for the creation of a German version of the "Stoffenmanager", the GESTIS-Stoffmanager. The German version shall be placed on the Internet in the course of 2010, where it will be freely available for small and medium-sized enterprises as a tool for risk assessment.

Last Update: