Land Use Regression Modelling of traffic-related noise in Athens, Greece for use in epidemiological studies

Abstract

BACKGROUND AND AIM: Even though exposure to noise has been associated with higher blood pressure and sleep disturbance, a very limited number of studies have developed Land Use Regression (LUR) models to estimate exposure to transportation noise. Our aim was to develop a LUR model that estimates long-term exposure to road-traffic noise, within the greater Athens area in Greece, for subsequent use in epidemiological studies. METHODS: We developed a LUR model for road-traffic noise exposure to Lday-evening-night (Lden) using noise measurements conducted at 241 points in the study area, during 2014-2015. We applied a semiparametric approach including linear and smooth functions of spatial covariates and a bivariate smooth thin plate function for the geographical coordinates of the monitoring sites. The final set of explanatory variables was selected based on the adjusted-R2. We tested the final model’s performance using 10-fold cross-validation (CV). Subsequently, we applied the developed model to estimate exposure to road-traffic noise of 1,321 participants in the National Survey of Morbidity and Risk Factors (EMENO), at their residential addresses. Logistic and linear regression models were used to evaluate the association between exposure to road-traffic noise and the risk of hypertension and mean arterial blood pressure (MAP). RESULTS:The final set of predictor variables included traffic load variables, number and surrounding area of bus lanes, open space areas and nearest building height. The R2 of the developed model was 0.51 (CV-R2:0.39). Mean exposure (SD) of participants to Lden was 56.8 (2.54)dB. A 10dB increase in Lden was associated with an odds ratio (OR) for hypertension of 1.69 (95% confidence interval (CI):0.84, 3.41) and an increase in MAP of 1.3mmHg (95% CI: -2.00, 4.60 mmHg). CONCLUSIONS:LUR modeling provides a useful tool for estimating exposure to transportation noise for subsequent use in epidemiological studies, but further improvements of exposure are needed. KEYWORDS: Environmental epidemiology, Exposure assessment, Noise, Long-term exposure, Cardiovascular diseases