OP XI – 6 Evaluation and prediction of indoor and outdoor noise differences in residential dwellings using statistical model. (18th March 2018)

• Record Type: Journal Article
• Authors: Naim, Faridah; Fecht, Daniela; Hansell, Anna; Gulliver, John
• Is Part Of: Occupational and environmental medicine. Volume 75(2018)Supplement 1
• Page Start: A22
• Page End: A23
• Publication Date: 2018-03-18
• Subjects: Medicine, Industrial -- Periodicals; Environmental health -- Periodicals; 616.980305
• DOI: 10.1136/oemed-2018-ISEEabstracts.54
• Languages: English
• ISSNs: 1351-0711

Abstract

Abstract : Background/aim: Most epidemiological studies investigating the association between noise exposure and health use modelled outdoor noise estimates as exposure proxy. This study aims to explore the relationship between indoor and outdoor noise at residential dwellings in London and evaluates potential parameters of dwelling ad sound insulation characteristics to predict the indoor exposure. Methods: Measurements were done at 49 homes mostly in winter and spring seasons when windows are closed. Continuous noise measurement were made inside and outside each home for three consecutive days using an Optimus CR:171B sound level metre. Selected homes were located close to major roads, railway, under an aircraft flight path or any combination of them. Building survey and time activity diaries of the occupants were also collected. Results: From the unadjusted linear regression model, significant associations were found between indoor and outdoor noise for LAeq, 16h (β=0.413, 95% Confidence Interval (CI): 0.081–0.745, p=0.016) and Lnight (β=0.332, 95% CI: 0.039 to 0.624, p=0.027). After adjusting for room volume, window size, source of noise, window type and number of occupants, the association remained significant in the adjusted model for LAeq, 16h (β=0.378, 95% CI: 0.071 to 0.685, p=0.018) and Lnight (β=0.297, 95% CI: 0.013 to 0.580, p=0.041) with only a small reduced of the β coefficient from the unadjusted model. These linear models yield an explained variance of 64% (adjusted R 2 ) for LAeq, 16h and 51% for Lnight . Based on the preliminary analysis, it is predicted that the indoor level could be 60%–70% lower than the outdoor. Conclusion: The wide CI suggests misclassification bias of exposure when conducting epidemiologiocal studies using the outdoor noise estimates. Further analysis will be done to improve the model by adding some more related sound insulation factors.