| [Download PDF]
|Year : 2009 | Volume
| Issue : 44 | Page : 169--175
Reduction of road traffic noise and mental health: An intervention study
Stephen A Stansfeld1, Mary M Haines2, Bernard Berry3, Michael Burr4,
1 Centre for Psychiatry, Wolfson Institute of Preventive Medicine, Barts and the London School of Medicine and Dentistry, London, United Kingdom
2 Centre for Psychiatry, Wolfson Institute of Preventive Medicine, Barts and the London School of Medicine and Dentistry, London , UK and The Sax Institute, Level 8, Building 10, 235 Jones Street, Ultimo 2007, Sydney, Australia
3 Berry Environmental Ltd, Shepperton, Surrey, TW17 0JZ
4 University of Wales College of Medicine, Temple of Peace & Health, Cathays Park, Cardiff, CF1 3NW
Stephen A Stansfeld
Centre for Psychiatry, Queen Mary, University of London, Wolfson Institute of Preventive Medicine, Old Anatomy Building, Charterhouse Square, London EC1M 6BQ.
Road traffic noise exposure leads to annoyance and impairment of quality of life and may impair health. If this association is causal, a reduction in noise exposure should result in a reduction in noise annoyance and improvement in quality of life. This study examines whether the reduction in road traffic noise following the introduction of a bypass leads to reduction in noise annoyance and common mental disorder and an improvement in quality of life. Repeated measures field study with intervention in three small towns in North Wales, UK. Participants were residents 16 to 90 years living in areas of high or low exposure to road traffic noise. At baseline there was no difference in annoyance, quality of life or common mental disorder between traffic noise exposed and quiet areas. There was a small reduction in noise exposure (2-4 dBA) with the opening of the bypass. There was no reduction in noise annoyance and no change in levels of common mental disorder and quality of life following the introduction of the bypass. Traffic noise reduction associated with the introduction of the bypass was not associated with measurable changes in quality of life or common mental disorder. This study suggests that reduction in traffic noise level of 3dB or less is insufficient to influence annoyance or mental health. However, the methodological difficulties of the study limit the conclusions that can be drawn on whether there is a causal effect of noise on common mental disorder.
|How to cite this article:|
Stansfeld SA, Haines MM, Berry B, Burr M. Reduction of road traffic noise and mental health: An intervention study.Noise Health 2009;11:169-175
|How to cite this URL:|
Stansfeld SA, Haines MM, Berry B, Burr M. Reduction of road traffic noise and mental health: An intervention study. Noise Health [serial online] 2009 [cited 2021 Oct 22 ];11:169-175
Available from: https://www.noiseandhealth.org/text.asp?2009/11/44/169/53364
There is little evidence that transport-related environmental noise causes serious mental health problems. Yet noise exposure is consistently associated with annoyance responses, sleep disturbance  and performance deficits and it is difficult to shake the popular opinion that noise 'drives you mad'.
Studies suggest that transport noise exposure may be responsible for psychological symptoms, , and increased consumption of non-prescribed hypnotics  but does not cause significant psychiatric disorder in community surveys  although one recent study linked aircraft noise exposure and anxiety disorders.  These studies have not been accepted as definitive for two reasons. First, there may be selective migration of noise sensitive individuals who might be more prone to noise effects out of highly noise-exposed areas that may bias the results of community surveys. Secondly, the association between noise exposure and health may be confounded by social disadvantage.
One way of overcoming these methodological problems is to carry out an intervention study manipulating noise levels and observing possible changes in health. Preliminary evidence suggests that noise barriers and reductions of road traffic noise level at source by approximately 10dBA may decrease levels of annoyance, improve sleep quality, improve wellbeing and reduce psychosocial symptoms in communities exposed to road traffic. ,,,,,
An opportunity arose to conduct an intervention study in the UK through the Bypass Study in North Wales, a study of air pollution and respiratory health.  In this study health measures were taken from residents living beside a busy main road in three small adjacent towns and in quieter adjacent control areas at baseline. Meanwhile, a bypass road was built diverting road traffic away from the main road. Our study addressed two questions: 1) Do people who are exposed to high levels of road traffic noise experience noise annoyance, lower quality of life, impairment of sleep, and common mental disorder relative to non-noise exposed groups? 2) Does a fall in high levels of road traffic noise result in decreased annoyance, improvement in quality of life and decreased common mental disorder?
Materials and Methods
Sample and intervention
The sample were adults over 16 years living in households, originally recruited for the Respiratory Health Study,  exposed to high levels of road traffic along the main street in three towns in Clwyd, North Wales and adults living in 'control' households exposed to lower levels of road traffic in the same area. Households were divided into two groups: Noise exposed households facing on to the main street and control households on uncongested side streets in the same area. All adults over 16 years from these households were eligible for inclusion in the study. In the Respiratory Health Study 400 residences were identified as eligible on congested streets, 45 of these were empty or for sale, no contact was made with 90 residents, 76 declined to participate and 14 gave little useful information on health. In the uncongested streets there were 505 eligible residences, of which 67 were empty or for sale, no contact was made with 143 residents, 76 declined to participate and 35 gave little useful information on health. In 1996/7 health information was obtained from 175 premises on congested streets (49.3% of eligible households) and 184 premises on uncongested streets (42% of eligible households). The estimate of eligible households may be overestimated because many turned out to be non-residential business premises. Baseline self-report questionnaires were collected during October and November 1997. The bypass was opened in March 1998. Follow-up measurements were carried in October and November 1998. The age range of participants was from 16-90 years.
Noise exposure was measured outdoors in both the high and low road traffic noise exposed streets at baseline in the first two weeks of December 1997; repeat measurements were carried out 12 months later in the first week of December 1998. Noise measurements were made between 10 am and 5 pm on weekdays in both noise exposed and quiet areas with matching of time of day between baseline and follow-up. L 10 and L eq values measurements were made over 3 hours and 15 minutes. Twenty-four hour bi-directional traffic flow counts were obtained from Flintshire County Council at two points on the main street for the first two weeks in 1997 and 1998 (to coincide with the periods during which noise measurements were made).
Health outcome measures
Noise annoyance was measured with three standard questions  that assessed the level of annoyance for three sources of environmental noise at home: (Neighbors, road traffic noise and train noise). The SF-36 General Health Survey  was used to measure quality of life and functioning.  Sleep disturbance was measured by the Jenkins Sleep Scale  The 28-item General Health Questionnaire (GHQ)  was used as a well validated measure of psychological distress.  The Revised Clinical Interview Schedule  assessed the prevalence of common mental disorder in a subsample. Respondents scoring above a threshold of 12 or more were considered to be potential psychiatric cases.
Environmental and household measures included: Neighborhood satisfaction, non-noise annoyance responses, perception of road safety and fear of road traffic accidents and housing problems, hours spent at home, household insulation, number and description of rooms facing the street, number of windows facing the street and how often the windows were open. Sex, age, social class, household deprivation and length of time living in the house were gathered at baseline in 1996/7. Social Class was coded on the basis of the occupation of the main wage earner using the Registrar General's classification. At follow-up, in 1998, additional socio-demographic measures were collected, including crowding, employment status and social class.
The approval of the local research ethics committee was granted to conduct the study. Questionnaire and interview data were collected in the noise exposed and control areas simultaneously. This introductory letter did not refer to the issue of traffic noise, so as not to introduce bias. Noise and road traffic questions were embedded in the environment section of the questionnaire to counter the possibility of 'halo effects' biasing responding. The health data collection was repeated a year later following the same procedure.
Estimation of prevalence of common mental disorder
The revised version of Clinical Interview Schedule was used as a second stage interview to estimate the prevalence of common mental disorder.  The research nurse scored the GHQ dividing participants into low GHQ scorers (range 0-1), medium GHQ scorers (range 2-12) and high GHQ scorers (range 13-28). Seventy-one interviewees were randomly selected, stratified by noise exposure, sex and low, medium and high scorers on the GHQ. The CIS-R interviews were conducted in the participants' homes by MH, who was blind to the GHQ score of the participant, conducted within the same month that the participant completed the GHQ. Cases from the CIS-R were defined as those having a score of 12 or more. The overall prevalence of caseness in the high and low noise groups was determined by weighting the observed strata specific prevalence; using weights equal to the inverse of the sampling fractions used.
Cross-sectional baseline comparisons
In SPSS Version 10, univariate analyses of covariance (ANCOVA) were used to examine the cross-sectional main noise effects. Model 1 was unadjusted and model 2 was adjusted for deprivation assessed with an index based on Townsend's scale incorporating: Car ownership, home ownership, crowding and unemployment.  A small amount of missing values on the sub-scales for deprivation (ranged between 2-7%) were taken into account when calculating the total deprivation score. Deprivation had two levels: Deprived and non-deprived. All statistical tests were two-tailed and alpha was set at 0.05.
Within-subjects follow-up analyses
Univariate analyses of covariance adjusting for baseline health status (ANCOVA) were used to assess the noise effects over time. The independent variable noise exposure was binary (noise exposed or control area). Model 1 was unadjusted, model 2 was adjusted for age, sex and baseline health, and model 3 was adjusted for age, sex, baseline health, and deprivation. In model 3 adjustment for deprivation included both baseline and follow-up scores of household deprivation to account for change in deprivation levels on health status.
The baseline response rate was 70% of the original respiratory survey (63% in high noise, 76% in low noise). At follow-up the response rate was 74%. The household follow-up response rate was higher in the low noise area 78% (n = 161) than in the high noise area 69% (n = 125) [Table 1]. In the high noise area 58 participants were excluded because they worked at businesses and were not resident in the high noise area, thus there were 67 participants in the final high noise sample at follow-up [Table 1]. More participants in the high noise area (16%, n = 28) had moved by the time of follow-up than the participants in the low noise area (4%, n = 9).
Noise exposure results
A summary of these results is presented in [Table 2]. At baseline, noise levels (expressed as L 10 ) in the exposed locations varied from around 75 to 78 dB(A), while in the less exposed locations noise levels were around 55 to 58dB(A) - but were probably influenced to some extent by train noise. The follow-up data showed that most of the exposed locations experienced a small reduction in noise level of around 2 to 4 dB(A).
Values obtained for L eq show similar patterns with pre-bypass levels of around 72 to 75 dB(A) for exposed locations and around 55 to 63 dB(A) for those less exposed. The values at follow-up again indicate a reduction but the range was greater, between 0 and 5 dB(A).
Figures for the percentage of heavy goods vehicles during noise measurements generally fell from around 11-12% in 1997 to around 3-5% in 1998. Traffic flow data showed that the construction of the by-pass reduced the amount of traffic using the main road. The 24 hr traffic count averaged over 5 week days from December 1 st 1997 fell from 25,591 to 22, 523 vehicles for the same period in 1998. At the second site the 24 hr traffic count averaged over 5 week day s from December 1 st 1997 fell from 23,643 to 20,721 vehicles for the same period in 1998. There was also evidence that the percentage of heavy vehicles decreased. Using the CRTN formula  it was estimated that the by-pass had resulted in predicted noise levels falling by around 2.5 dB(A) L 10,18hr ; a large contribution to this was probably due to the decrease in the proportion of heavy vehicles.
Health outcome results
At baseline, the two samples were well matched except for occupation, home ownership, and length of occupancy [Table 3]. At follow-up, the two samples were well matched except for unemployment. At baseline high noise and control samples did not differ on any of the health outcomes in both the unadjusted and deprivation adjusted analyses [Table 4]. At follow-up high noise and control household samples did not differ on any of the health outcomes in the unadjusted cross-sectional comparison at follow-up [Table 5]. The within subjects analyses indicate that change in health status did not differ between the two groups over the one-year period following the introduction of the bypass for any of the health outcomes.
Assessment of prevalence of common mental disorder using the CIS-R
At baseline the prevalence of common mental disorder was higher in the control area than the noise exposed area (22.9% (95% CI 12-41) versus 14.9% (95% CI 8-28)). After the bypass opened, the prevalence rates of common mental disorder fell in both areas, more so in the control area than in the noise exposed area (13.6% (95% CI 6-30) versus 10.0% (95% CI 4-26)). These falls in prevalence are not in keeping with a noise related effect where a larger effect would be expected in the area which had the larger decline in noise exposure.
Mean noise annoyance levels did not differ at baseline or follow-up. We also examined categorical change in annoyance. The low noise sample's level of noise annoyance remained constant at 72% not/a little annoyed, and 28% moderately/very annoyed both at baseline and follow-up. There was a slight decrease in noise annoyance in the high noise sample over the one-year period. At baseline 35% reported being moderately/very annoyed by road traffic noise and this decreased to 27% at follow-up. At baseline 65% reported being not/a little annoyed, this increased to 73% at follow-up. Logistic regression analysis demonstrated that this change in noise annoyance was not statistically significant ( P = 0.76).
Environmental and household factors
At both baseline and follow-up there were no differences between the noise exposed area and the control area on: General neighborhood satisfaction, perception of road safety and fear of road traffic accidents. Household factors were measured because they may influence the association of road traffic exposure and health. There was no difference between the high noise exposed group and control group on hours spent at home and extent of household insulation. People living in the high noise group were less likely to open their windows than people living in the control area (in high noise 11.4% respondents reported that they never opened their windows that face the street compared with 0.5% in the control sample ( P = 0.0001)).
We found no differences between the noise exposed and quiet areas at baseline in annoyance, quality of life or common mental disorder and no change in annoyance, health functioning or common mental disorder related to the introduction of the bypass.
This raises three possibilities: Either (i) the noise exposed and control samples differed because of selective factors at baseline which might differentially affect mental health and obscure any effect of noise on mental health, or (ii) changes in noise exposure with the introduction of the bypass were insufficient to influence health or (iii) changes in noise exposure do not affect mental health.
The noise environment
Overall, the changes in noise exposure were small. This was not expected but may be explained by the local conditions. The towns are situated in a narrow coastal corridor, which contains industrial development and carries both road and rail traffic to North Wales from the North West of England. Nevertheless, although noise levels were generally high there were quieter areas within the control sample. The peripheries of the towns are bordered on one side by the main railway line and on the other by another major road which skirts the edge of the town. Environmental noise exposure from these sources is unlikely to have changed with the introduction of the bypass.
Furthermore, the control sample had been originally chosen on the basis of expectations for air pollution sampling and it is possible that the noise exposure levels did not always coincide with these boundaries resulting in some misclassification of exposure. A further reason for little change in noise exposure related to the bypass was because of the nature of the main road and the traffic which uses it. The main road is a busy shopping street, attracting a lot of local car traffic - this was unlikely to have been affected by the bypass. Moreover, those respondents who live in the quieter control areas are likely to shop on the main road and be exposed to noise for at least that portion of the day. It is also possible that the habits of road users may still change further with greater familiarity with the bypass although measurements taken between 7 and 9 months after the opening of the bypass should be at a sufficient interval to account for this.
Previous studies of the introduction of a bypass have shown reductions in annoyance levels which have been immediate. These studies, however, did show much greater reductions in noise levels than in this study. Other interventions to reduce road traffic noise such as noise barriers and reductions in traffic speed and road traffic at night have yielded reductions in annoyance, improved sleep and well being and reduced 'psychosocial symptoms'. ,,,, However, noise reduction in these studies was 10dBA, 11 dBA and 9-14dBA respectively. Thus it seems quite probable that the lack of effects shown in this study in comparison to previous studies may be related to a lack of decrease in noise exposure sufficient to affect the most sensitive noise effects indicator, annoyance.
Noise and the absence of health effects
Noise levels differed substantially between noise exposed and quiet areas and yet differences in mental health were minimal. Residents along the main road tended to be more transient and may have been healthier. Additionally many residents living on the noisy road may have been away during the day at work exposed to different noise levels. This is in keeping with other community studies which do not show major effects of environmental noise on mental health at currently experienced road transport levels.  Annoyance is likely to be most sensitive to change in noise exposure and the lack of any changes in annoyance means that changes in health indicators are even less likely. In the Caerphilly study, also carried out in Wales, we found a small non-linear association between road traffic noise exposure and anxiety symptoms.  In the bypass study we found no differences in anxiety symptoms on the GHQ-28 between noise exposed and control areas.
Secondly, as questionnaire assessment of common mental disorder has been criticized as being prone to reporting bias, we used the revised Clinical Interview Schedule to derive prevalence of common mental disorder. Change in the prevalence of common mental disorder was not greater in the noise exposed compared to the control area at follow-up. The prevalence of common mental disorder dropped sharply in both areas. We think the fall in prevalence may be a methodological artifact of repeated testing as the GHQ responses did not drop in a similar way making a contextual change such as increased employment a less likely explanation.
This study was funded because it was an innovative and a cost-effective way of conducting an intervention study on two grounds. Firstly this study was funded to test a 'real world' intervention, namely the introduction of a bypass road that had already been funded, received planning approval and building was underway. Secondly our noise study was designed to opportunistically 'piggy-back' onto the respiratory health study. The rationale being that the bypass road would lead to the diversion of road traffic away from the main road that in turn would result in a reduction of both air pollution and noise exposure for the residents.
The results from this study suggest that there are a number of lessons to be learnt from this approach of both testing a 'real world' intervention and combining two studies with the same design. It is important to assess the feasibility to combining a study of two pollutants on the same community sample because the nature of the dispersion and levels of both the pollutants may well be different. The 'control sample' for the respiratory health study was also the control sample for our study. This group lived in residential properties on the uncongested side streets in the same area as the main road, but were also affected by train noise. It is plausible that this group (who indeed lived in significantly quieter road traffic noise exposed homes) were also subject to noise exposure in their daily lives from trains and when they went to the main road to visit the shops, schools and other public amenities.
There are several limitations to this study. Ideally, noise measurements could have been more detailed and lengthier in the control area, especially as the environmental noise climate was complex. Another factor that may have led to bias was non-response, because there was more loss to follow-up in the high noise sample because of differential mobility, although this is less likely to have affected the sickest individuals. Ideally more detailed and sensitive measures of sleep disturbance would have provided a more rigorous test of sleep effects.
This study provided no evidence that road traffic noise exposure has effects on quality of life and common mental disorder in an adult general population sample. The decline in noise exposure in the high noise exposure area with the opening of the bypass was possibly insufficient to test for the effects of road traffic noise on change in mental health but there was no suggestion of an improvement in mental health with a reduction in traffic noise. The methodological problems encountered in the study mean that definitive conclusions on whether noise causes or does not cause common mental disorder cannot be drawn.
|1||Bluhm G, Nordling E, Berglind N. Road traffic noise and annoyance-an increasing environmental health problem. Noise Health 2004;6:43-9.|
|2||Halpern D. More than Bricks and Mortar? Mental Health and the Built Environment . London: Taylor and Francis; 1995.|
|3||Stansfeld S, Gallacher J, Babisch W, Shipley M. Road traffic noise and psychiatric disorder: Prospective findings from the Caerphilly Study. BMJ 1996;313:266-7.|
|4||Franssen EA, van Wiechen CM, Nagelkerke NJ, Lebret E. Aircraft noise around a large international airport and its impact on general health and medication use. Occup Environ Med 2004;61:405-13.|
|5||Tarnopolsky A, Morton Williams J. Aircraft noise and prevalence of psychiatric disorders, Research Report. London: Social and Community Planning Research; 1980.|
|6||Hardoy, M. C, Carta, M. G, Marci, A. R, Carbone, F, Cadeddu, M, Kovess, et al . Exposure to aircraft noise and risk of psychiatric disorders: The Elmas survey--aircraft noise and psychiatric disorders'. Soc Psychiatry Psychiatr Epidemiol 2005;40:24-6.|
|7||Griffiths ID, Raw GJ. Community and individual response to changes in traffic noise exposure. Dev Toxicol Environ Sci 1987;15:333-43.|
|8||Kastka J, Buchta U, Ritterstaedt R, Pausen R, Mau U. The long term effect of noise protection barriers on the annoyance response of residents. Journal of Sound and Vibration 1994;184:823-52.|
|9||Mital A, Ramakrishnan AS. Effectiveness of noise barriers on an interstate highway: A subjective and objective evaluation. J Hum Ergol(Tokyo) 1997;26:31-8.|
|10||Íhrstrφm E, Skφnberg A. Adverse health effects in relation to noise migration - A longitudinal study in the city of Gφteborg. Sweden: Department of Environmental Medicine, University of Gφteborg; 1998.|
|11||Ohrstrom E. Longitudinal surveys on effects of changes in road traffic noise-annoyance, activity disturbances, and psycho-social well-being. J Acoust Soc Am 2004;115:719-29.|
|12||Nilsson ME, Berglund B. Noise annoyance and activity disturbance before and after the erection of a roadside noise barrier. J Acoust Soc Am 2006;119:2178-88.|
|13||Burr ML, Karani G, Davies B, Holmes BA, Williams KL. Effects on respiratory health of a reduction in air pollution from vehicle exhaust emissions. Occup Environ Med 2004;61:212-8.|
|14||Fields J. Effect of personal and situational variables on noise annoyance: With special reference to implications for en route noise. Research report for Federal Aviation Administration Office of Environment and Energy, 800 Independence Avenue SW. Washington, DC 20591 and NASA Langley Research Center Hampton: VA 23665; 1992.|
|15||Ware JE Jr, Sherbourne CD. The MOS 36-item short-form health survey (SF-36). I. Conceptual framework and item selection. Med Care 1992;30:473-83.|
|16||McHorney CA, Ware JE Jr, Raczek AE. The MOS 36-Item Short-Form Health Survey (SF-36): II. Psychometric and clinical tests of validity in measuring physical and mental health constructs. Med Care 1993;31:247-63.|
|17||Jenkins CD, Stanton BA, Niemcryk SJ, Rose RM. A scale for the estimation of sleep problems in clinical research. J Clin Epidemiol 1988;41:313-21.|
|18||Goldberg DP, Hillier VF. A scaled version of the General Health Questionnaire. Psychol Med 1979;9:139-45.|
|19||Goldberg D, Williams PA. Users Guide to the General Health Questionnaire . Windsor:NFER: Nelson; 1998.|
|20||Lewis G, Pelosi AJ, Araya R, Dunn G. Measuring psychiatric disorder in the community: A standardized assessment for use by lay interviewers. Psychol Med 1992;22:465-86.|
|21||Townsend P, Phillimore P, Beattie A. Health and Deprivation . London: Routledge, 1989.|
|22||CRTN. Calculation of road traffic noise . London: HMSO; 1988.|