Abstract
Everybody is exposed to radiofrequency electromagnetic fields (RF-EMF) in daily life, either by their own communications device(s) or by mobile phone base stations or radio transmitters in their living environment. There are still concerns about the potential health effects of RF-EMF among scientists and in society. This thesis addresses some
... read more
of these uncertainties in RF-EMF epidemiology such as investigating the repeatability of self-reported use of mobile communication devices and improvement of exposure assessment of mobile phone use by combining self-reported and operator-recorded data. Furthermore we examined the role of perception besides modeled exposure for health symptoms.
The first part of my thesis was about improving RF-EMF epidemiology by harmonizing and combining data from multiple prospective cohort studies and by evaluating and determining exposure assessment of RF-EMF by improving the accuracy of (self-reported) mobile phone use within the COSMOS study. We showed that combing different sources (self-reported and operator recorded) by regression calibration resulted in limited bias and outperformed other statistical methods such as complete-case and multiple imputation methods.
Furthermore a case ascertainment method was developed to identify Parkinson’s Disease within two large prospective cohort studies (AMIGO and EPIC-NL). We included multiple information sources like registry data and self-reported data. We evaluated our developed likelihood score, existing of four categories, with known etiological risk factors of Parkinson’s Disease. For epidemiological studies for which no specific registry is available, as is the case for Parkinson’s Disease, it is important to combine multiple information sources to maximize sensitivity and specificity in the disease outcome measure.
In the last part of this thesis we expanded the knowledge on possible health effects of RF-EMF. We found no effects with birth weight or fetal growth and maternal mobile phone use during pregnancy. However we found effects of shorter pregnancy duration and increased risk of preterm birth. We could not adjust for maternal stress in this study, although it has been shown to be an important confounder. Furthermore we investigated the impact of both modeled and perceived exposure to RF-EMF from mobile phone base stations, road traffic noise and air pollution on self-reported health outcomes. We used longitudinal data from the AMIGO cohort. Perceived exposures showed overall stronger associations with symptoms than modeled exposures. When both modeled and perceived exposure was included as predictors in a single model, the impact of modeled exposure on self-reported health outcomes disappeared or strongly diminished. This shows the importance of perceived exposures in certain health effects.
This thesis showed that using multiple information sources for assessment of both exposure and health outcomes within epidemiological studies is feasible. The findings presented in this thesis will be input for future epidemiological analyses within COSMOS and for studies on Parkinson’s Disease and environmental exposures within the AMIGO and EPIC-NL cohorts. The methods proposed in this thesis could also be applied in other (environmental) epidemiological studies. Further integration of multiple sources of RF-EMF by integrative exposure assessment models in epidemiological studies is essential to get more definitive answers on potential effects of RF-EMF exposure.
show less
