Difference between revisions of "Health effects of transport noise in Europe"

Revision as of 08:05, 7 June 2011

About transport noise

Noise from road, rail, and air traffic affects a great number of people. Exposure to transport noise may cause sleep disturbance as well as annoyance, potentially leading to high blood pressure and increased incidence of myocardial infarction (WHO, 2000b; Miedema & Vos 2007; Babisch 2006, 2008). Transport noise exposure as a part of total environmental noise has also been linked to effects on cognition. Transport noise is selected in this study due to its ubiquity and high public health impact. In addition, due to the economic significance of transport, noise levels despite technological progress keep on increasing over time.

Selected health endpoints and exposure-response functions

As health end-points, high sleep disturbance and Ischemic Heart Disease (IHD) were included (Miedema & Vos, 2007; Babisch 2006, 2008). Hypertension and related heart disease due to aircraft noise was not considered because no clear review could be identified. Nevertheless, since causal relationships are very likely and have been reported recently, this health effect may be considered in the near future (Babisch & Kamp, 2009). For railway noise no significant associations with hypertension and IHD could be identified (Barregard et al., 2009).

Effects on cognition were also excluded, as these are difficult to quantify. In addition, severe annoyance, as annoyance was not included as annoyance does not fall weithin our definition of health and not considered a health effect by, amongst others, WHO. Some other studies (e.g. Knol & Staatsen, 2005) have applied a broader definition of health, in which annoyance was included as a health effect because it reduces quality of life. These studies show a substantial burden of disease due to transport noise related annoyance.

The formulas applied to estimate transport noise related high sleep disturbance (HSD) are as follows. The results directly give the number of people severely sleep disturbed at a certain decibel level for each noise source (Miedema et al., 2002, Miedema & Vos, 2007):

Road traffic noise: %HSD = 20.8 - 1.05L_night+ 0.01486(L_night)²

Railway traffic noise: %HSD = 11.3 - 0.55L_night+ 0.00759 (L_night)²

Aircraft noise: %HSD = 18.147 - 0.956L_night+ 0.01482(L_night)²

L_night is a measure of night-time noise, defined as the yearly average of night noise levels (23–7h) at the façade of houses. The formulas can be applied in the range of L_night from 45 to 65 (max. 70) dB(A) (data are regularly available for >50dB(A)).

There is no exclusive causal mechanism postulated specifically to myocardial infarction (MI). Therefore, the OR for MI has been applied to all types of ischemic heart disease (Babisch, 2008) according to the following formula (Babisch 2006, 2008):

OR = 1.63 – 0.000613*(L_day,16h)2 + 0.00000736*(L_day,16h)3

L_day,16h is defined as the yearly average of day and evening noise levels (7-23h) and L_day as the yearly average of day noise levels (7-19h) and L_evening as the yearly average of evening noise levels (19–23h). The exposure-response functions are valid for L_day,16h noise levels ranging from 55 to approximately 80 dB(A).

FIGURE 3-4. The exposure-response functions for high sleep disturbance (HSD; blue axis) and ischeamic heart disease (red axis) presented in graphical form for road, rail and air traffic noise. The OR for Myocardial Infarction, which we applied for all ischeamic heart disease (IHD), is modelled only for road traffic (red curve). (Dotted lines display E-R functions outside the range where they are considered valid).

Exposure data

The exposure metric used is the average 24h-noise level day-evening-night (L_den) and the average 8h-noise level during night-time (L_night) in dB(A), separated for road, aircraft and railway traffic. These indices have been suggested by the European Environmental Noise Directive (2002/49/EC, “END”, (EU, 2002).

From the first phase of END-reporting, carried out in 2007/2008, noise exposure data per 5dB(A) categories are now available for most of the EU-countries. These data concern:

• agglomerations with more than 250 000 inhabitants (separately for road, aircraft and railway traffic);
• roads outside agglomerations with more than 6 000 000 vehicles/passages per year;
• railways outside agglomerations with more than 60 000 trains/passages per year;
• major airports with air traffic higher than 50 000 movements/flights per year (some separated for inside and outside agglomerations and total).

These data have been aggregated recently and still are to some extent being processed. So far, they only cover a relatively small percentage of the EU population. The data are presented in Table 3-11. The population coverage of the data for the countries that have so far been included depends substantially on their urbanity, on the administrative prerequisites (e.g. extent of cities) and on the location of country (central or in periphery) and resulting transit influences. In consequence, the comparability of the countries is limited. Additionally, data for Belgium until April 2010 only included Flanders. For France, only data for major agglomerations were included. Only exposure levels above L_night 50dB (L_den 55 dB) are reported in END data, so there are no data about exposure below those levels in the END database.

For ischemic heart disease, the exposure-response function was calculated for a 16h-daytime level (L_day). As a crude but practical tool for conversion, the following easy formula was used:

L_day,16h = L_den – 2.5 (Babisch, 2008)

Unfortunately, END-reporting of many included countries until April 2010 was only complete for L_den but incomplete considering data about L_night. Therefore adjustments had to be made for calculation of people affected by high/severe sleep disturbance. As a crude but practical tool for conversion, the following easy formula was used:

L_den = L_night + 7.5 (expert judgement after first data-pooling, most valid for road traffic noise in agglomerations)

There are large differences concerning these conversion factors, especially for conversion from L_den to L_night, for which estimates range from +5 (very urban) up to +11 (including rural areas) (WG-AEN 2006). For comparability, 7.5 was chosen as a conservative factor, even though in some cases there might be an underestimation of real exposure.

END-reporting covers 5dB(A)-categories. For modelling purposes, the mid-values of the 5dB(A)-categories (50–54.9 → 52.5) were inserted in the non-linear polynomials as a feasible simplification.