6. HEALTH CONSEQUENCES AND HARMFUL EFFECTS OF NOISE |
6.1 Organization of monitoring activities
Noise as a risk factor differs from other monitored factors in its specific area distribution, especially in considerable changes of noise intensity in individual places under measurement and different times of day. This subsystem has been in operation for the past nine years in 21 sites including three Prague districts. It comprises the physical measurement of noise levels and a health questionnaire survey. In each monitored city two areas are selected: one defined as noisy and the other as quiet. All participating Public Health Service teams are identically equipped and measurements are carried out according to a unified methodology. The noisy and quiet areas were selected to enable total population noise exposure following repeated measurements to an accuracy in excess of 2 dB LAeq. This enables changes in traffic intensity to be recorded because a 3 dB shift occurs (when expressing noise as equivalent acoustic pressure levels) in the event of a 50 % decrease, or doubling, of the amount of noisy events or traffic intensity. Compliance with this criterion of accuracy was tested at several sites in each area. Other criteria for site selection were as follows:
In the basic localities the health effects of noise have been investigated over the past nine years three times (in 1995, 1997, and 2002) through a questionnaire worked out on the basis of one applied in the study of the effects of noise on population health, that has terminated in 1985 (within the framework of a project of the Chief Public Health Officer), as well as on the basis of the CINDI questionnaire.
6.2 Noise measurement
Noise levels are measured for a period of 24 hours to record their development throughout the day. Measurements are taken once per month, alternately in noisy and quiet localities. The required accuracy has been obtained by the application of the same technique in all the localities as well as by precisely observing the measuring procedures in agreement with the unified measuring methodology for Project III in line with the requirements in the Methodological Guideline of the Chief Public Health Officer in 2001.
In the period of 1998 through 2002 the mean annual values of equivalent levels and 90 % levels of noise formed a continuous sequence. The results show that noise in different areas is evenly distributed throughout the whole range of noise levels monitored. Noise levels detected as equivalent acoustic pressure A reach 75 dB by day and 70 dB by night in the most noisy areas, and drop below 45 dB by day and 39 dB by night in the quiet localities. Equivalent levels of acoustic pressure A – LAeq express noise from the technical point of view describing noise as energy levels at different sites (Fig. 6.1a, 6.1b and Fig. 6.2a, 6.2b). The 90%-probability noise levels L90 describe permanent noise at each site (Fig. 6.3a, 6.3b and Fig. 6.4a, 6.4b). As has already become usual, the highest noise levels were found in the noisy localities of Prague 2, 3 and 10, Plzeň, Ostrava and Olomouc. Comparison of these results with the requirements set by Governmental Resolution No. 502/2000 Dig. shows that only a few localities of all 42 (two thirds of quiet localities) are found in the range of noise values below the limits for ambient noise (55 dB in the day and 45 dB at night time in LAeq). These limits correspond with recommendations of the WHO for the ensuring of safe protection of health against noise, when not even under long-term dwelling in environment below that noise level there should not occur any impairment of health.
In the year 2002 there has been carried out a detailed processing of interannual changes of noise levels in the nine-year period of 1994 through 2002 by the summarizing in the form of the linear and quadratic trends for quiet and for noisy localities and for day-time and night-time measurements. A linear decrease has occurred namely in Prague 2, Prague 3, Liberec, and in České Budějovice. An increase has been occurring especially in quiet localities, namely in Prague 10, Liberec, Žďár n. Sázavou, Kladno and Ústí n.Orlicí. An increase in the noise level in noisy localities took place in Ostrava and Olomouc. In the noisy as well as quiet localities in Příbram, Hradec Králové, Brno, and Havlíčkův Brod the interannual noise level has remained constant.
6.3 Health consequences and disturbing effects of noise
The monitoring of effects of noise on the population in the year 2002 in 42 localities was conducted through a questionnaire survey. The response rate to the questionnaire survey was about 53 % (an 18 % increase as against the survey in 1997). In the final processing there have been included 12,123 questionnaires from respondents 30–90 years of age. For the assessment of noise effects on health the data on night-time noise have been used in view of that the population exposure to noise is more homogenous than in day time.
Data on health complaints in the percentage of replies were related to the particular data on the noise level of the locality. Through a thus formed field of points there has been plotted a regression curve expressing the correlation between the noise level in the localities under study and the reply to the given question or phenomenon under follow-up. An evaluation was carried out with the aid of a correlation coefficient and was formulated towards the basic hypothesis of random distribution of the monitored factor in relation to the noise level. Under conditions of refusing that hypothesis of randomness at a corresponding level of significance alpha, one can for simplification speak of “statistical significance” or “significant correlation”. In order to verify the results of the above-mentioned way of statistical processing with the application of a continuous scale of resulting noise level values, there has been applied the STATA program which evaluates the occurrence of health indicators in relation to three categories of night-time noise levels, namely up to 45, from 45 to 62, and over 62 dB(A), simultaneously evaluating the influence of potential confounding factors.
The questionnaire also examines basic demographic and sociological characteristics of the population as well as questions pertaining to lifestyle. This group of questions verified the homogeneity of the population under monitoring with the aim of minimizing situations in which the effects of noise on the population would be simulated by another, so-called “confounding” factor. From the evaluation of the demographic and socio-economic questions it follows, just as in previous periods of monitoring, that there are no significant differences between the individual localities. Data concerning living in the present homes of the respondents likewise do not show any statistically significant differences except for the share of flats in 50–100 years old buildings, the number of which is significantly larger in the noisy localities (city centers) just as in the previous periods of monitoring. However, this factor is not necessarily confounding because one cannot expect significantly worse living conditions on those flats. A marked influence on exposure to noise could have the duration of staying at home, however, no significant differences concerning the duration of exposure to noise or dwelling at home, have been found, similarly as in previous investigations. There have been found no differences between the localities as to the numbers of respondents from noise hazard occupational environments. Neither in the physical activity of the respondents just as the occurrence of health risk factors in lifestyle such as the smoking habit or coffee and alcohol consumption, have not shown any statistically significant differences between the localities.
6.3.1 Results of investigation into the correlation between noise levels and health complaints
To serve as indicators there were selected civilization-linked diseases, a depiction of mental welfare, or information on the occurrence of neurotic symptoms and sleep disorders. With increasing noise levels there is a statistically significant increase in the percentage of respondents annoyed by ambient noise, from 25 % up to 95 %, that being practically the same as in preceding periods of monitoring. Similarly, as in previous investigations, it has been found that there is a statistically significant decrease in the number of subjects annoyed by noise from their neighbors. Noise of that kind, from the mental point of view is namely not as dominant and physically rather blends in with the outdoor noise level.
As against previous questionnaire surveys, the ability to readily relax and rest has been shown as significantly correlated with the ambient noise level. With increasing noise levels in the outdoor milieu the ability of the respondents readily to relax and rest decreases to a statistically significant degree (p < 0.05) from 77 % down to 72 %. The effect of noise on the respondents is manifested by a diminished ability to relax; the closest correlation was manifested at the daily value of L-90.
The occurrence of neurotic symptoms – palpitation of the heart and other unpleasant physical sensations when upset, has shown to be significantly correlated with the noise level in the environment by an increase from 45 % up to 53 % (p < 0.025) just as in the previous questionnaire survey. Statistically significant is also the influence of noise on the occasional use of tranquilizers and sleeping pills, that being in correlation with increasing noise level.
The influence of noise on sleep was followed up by questions on falling asleep, questions on the duration and quality of sleep. In the questionnaire survey conducted there has been for the first time investigated an important factor, namely whether the disturbance of sleep is more correlated with a permanent equivalent level of noise or whether with individual noisy events in a more quiet environment. That can be expressed as L1–L90 = difference between the mean value of the difference of noisy events, that means L1 and the mean value of background noise L90. With an increasing noise level expressed with the aid of an equivalent noise level, in a statistically significant degree rises the number of respondents having problems with falling asleep and having bad sleep. In the case of expressing the noise level by a 90 % noise level or background noise, the percentage of respondents suffering bad sleep again rises with increasing noise levels following a regression curve to a statistically significant degree (p > 0.05) from a former 14 % up to 18 %. Similarly, with increasing noise levels in L90 there rises the number of respondents from 10 % to 16 % who have problems with falling asleep, also to a statistically significant degree. The influence of noise on the mean duration of sleep was not significant.
From the results regarding sleep disturbance it follows that with increasing noise levels there rises the number of respondents who have problems with falling asleep or suffer bad sleep, namely alike and to a statistically significant degree according to all three noise level descriptors. Differences between individual values, however, are apparent from results of more detailed questioning. In the closest significant correlation to the disturbing of falling asleep is the expression of the noise level with the aid of an equivalent level of noise in all the categories of more detailed questions and in the question on occasional sleep disturbance. The dynamic range of noise is not to be necessarily preferred because individual noisy events, when they do not follow closely one by one, disturb the falling asleep to a lesser degree than continuous noise, e.g. of a traffic-loaded traffic lane. In the disturbing of sleep, the dynamic range of noise is as significant as LAeq. It can be stated that expressing the noise level in terms of an equivalent noise level is a suitable descriptor of noise for the follow-up of effects on sleep, and it is not necessary to use any further descriptor, except in extreme cases.
An important indicator of health is occurrence of so-called civilization-linked diseases, i.e. myocardial infarction, stomach and duodenal ulcers, gall and kidney stones, diabetes mellitus, high blood pressure, tumor diseases, and frequent catarrhs of the upper respiratory tract. Statistical significance has now, as in previous surveys, been demonstrated only in hypertension and frequent catarrhs of the upper respiratory tract. The percentage of incidence of hypertension in each locality increases along with the noise level, from 24 % to 34 % according to the regression curve with almost 95 % statistical significance. Thus, noise probably does have an effect on hypertension, however, there is a number of other strong effects because of which the effects of noise cannot always be demonstrated with adequate statistical significance. In the processing of results obtained through the questionnaire with the aid of the STATA program, there have moreover been excluded other potential risk factors, namely age, educational status, frequency of physical activity, BMI, smoking, consumption of coffee and alcohol. In this statistical model it has been estimated that people who live on a long-term basis (at least 5 years) in localities with night-time noise levels exceeding LAeq 62 dB have, even after evaluating the above-mentioned so-called confounding factors, a 1.2 times greater chance of acquiring hypertension (p < 0.01). These conclusions are in line with results published in surveys abroad.
The incidence of falling ill with frequent catarrhs of the upper respiratory tract increases along with rising noise levels with 99 % significance from 15 % to 24 %. People living on a long-term basis in localities with night-time noise levels exceeding LAeq 62 dB are under an up to 1.4 times greater risk of falling ill, again upon the exclusion of other risk factors (age, educational status, physical activity, BMI, smoking, consumption of coffee and alcohol). The increased incidence of catarrhs of the upper respiratory tract can be explained by decreased resistance of the organism exposed to the effects of noise.
In the occurrence of other diseases, i.e. myocardial infarction, stomach and duodenal ulcers, gall and kidney stones, tumor diseases, and diabetes mellitus, no influence of noise has been demonstrated, perhaps also due to the relatively low prevalence of these affections in the population.
There has been demonstrated a correlation between the incidence of the sum of selected civilization-linked diseases and noise levels in the localities. Noise does not cause one or several specific diseases, but there is affected the overall health of the exposed subjects and there is an earlier onset of diseases that would otherwise have probably appeared at a much later age. Noise can also worsen the course of such diseases. In the questionnaire survey in 2002 it has been found that the incidence of the sum of civilization-linked diseases in relation to the noise levels of the localities expressed in night-time LAeq rises from 81 % to 102 % to a statistically significant degree. A statistically more significant correlation between the incidence of the sum of civilization-linked diseases and noise levels has surfaced in the case of expressing day-time noise levels in L90 (p < 0.05). This result is similar to conclusions drawn from previous questionnaire surveys wherein the most significant correlation of the occurrence of civilization-linked diseases was also when expressing day-time noise levels in the L90 value.
6.3.2 Assessment of the health effects of noise in extended localities
The 42 selected basic localities where there is being conducted the monitoring of noise levels and the questionnaire survey, are limited in extent and inhabited by a relatively small number of people in view of the overall population in the cities under monitoring. Therefore, for the assessment of the health effects of noise, there have also been selected so-called adjacent/extended localities. In such extended localities it is not possible, for reasons of working capacity, to conduct complete 24-hour measuring of noise, and therefore an approximation procedure has been chosen for the determination of noise levels. That lies in the conducting of day-time measuring and in an estimate of night-time noise values on the basis of differences between day-time and night-time noise levels established in the respective basic localities. The estimate is based on calculations of noise levels generated by surface traffic and is supplemented with a correlation of noise levels and the sum of health indicators under study found through the questionnaire survey in the basic localities. With the aid of the approximation procedure mentioned, even in these more vast localities there can be assessed the degree of the relative risk of health impairment by noise, but with lesser precision, of course.
6.4 Partial conclusions
In the year 2002, noise levels have been measured regularly in 42 localities. The noise levels in these areas ranged continuously from 47 to 76 dB in day time and from 37 to 69 dB at night time.
From the point of view of temporal trends in noise levels since the year 1994, there has been recorded a significant linear trend in the increase or decrease in noise levels in 12 cities at least in one of the variants (noisy or quiet localities, day-time or night-time noise measuring), in 9 cities there has been found a constant state in the noise levels, i.e. a non-linear course. A linear increase in the noise levels has been recorded mostly in the quiet localities.
The health-oriented section of this subsystem is based on a questionnaire survey of demographic, sociological and medical data of the population in the localities under monitoring. In the so-called civilization-linked diseases there has been found repeatedly a correlation between their occurrence and the degree of night-time noise levels at the 90 % level of statistical significance. However, that correlation is valid only for the sum of all the diseases under monitoring and for hypertension and frequent catarrhs of the upper respiratory tract. Other individual diseases do not reveal such a correlation in the whole group of respondents. There has been found a significant correlation between the percentage of people complaining of being annoyed by noise from the outdoor environment and the noise level measured. There has been found a likewise statistically significant relation between the noise level in the locality and the proportion of persons mentioning problems with falling asleep and with the quality of sleep.