• Model-Then-Add : Usual intake modelling of multimodal intake distributions

      van der Voet H; Kruisselbrink J; Boer WJ; Boon PE; VVH; V&Z (Rijksinstituut voor Volksgezondheid en Milieu RIVMBiometrisDLOWageningen University and Research centre, 2014-03-20)
      The National Institute for Public Health and the Environment (RIVM) and Wageningen University and Research Centre (WUR) have jointly developed software to estimate the amount of chemicals ingested via the diet (Monte Carlo Risk Assessment, MCRA). Examples of such chemicals include contaminants (e.g. acrylamide, dioxins, lead) and micronutrients. To estimate the intake of these chemicals in the long run, a module called Model-Then-Add has been added to the software. The long-term intake is relevant for chemicals that exert their beneficial or adverse health effect over a long period of ingestion. The Model-Then-Add module can be used when the distribution of individual mean intakes of the chemical in a population does not display a normal statistical distribution after a logarithmic transformation. This may, for example, be the case when the chemical is present in only a limited number of foods. In such cases, the module can be used to obtain a more realistic estimation of the longterm intake. A case study was performed to assess the long-term intake of smoke aromas, a group of chemicals that is potentially adverse at high intakes, using the Model- Then-Add module and the presently used methodology, which is known to overestimate the long-term exposure. The Model-Then-Add module resulted in lower intakes. The use of this module may thus result in less risk mitigation or environmental policy measures that need to be taken to reduce possible health risks. To estimate the intake of chemicals via the diet using MCRA, concentrations of chemicals in foods and beverages are linked to information on the consumption of these foods during a limited number of days. In the Netherlands, food consumption data are typically obtained from the Dutch National Food Consumption Surveys (Voedselconsumptiepeiling, VCP), in which information on food consumption is collected during two days. Statistical models are necessary to assess the long-term intake of chemicals based on these data.