Environmental Research 131 (2014) 17–24

Contents lists available at ScienceDirect

Environmental Research journal homepage: www.elsevier.com/locate/envres

Road traffic noise, sleep and mental health Karin Sygna a,b,1, Gunn Marit Aasvang a, Geir Aamodt b,c, Bente Oftedal a, Norun Hjertager Krog a,n a Department of Air Pollution and Noise, Division of Environmental Medicine, Norwegian Institute of Public Health, P.O. Box 4404 Nydalen, N-0403 Oslo, Norway b Department of Mathematical Sciences and Technology, Norwegian University of Life Sciences, P.O. Box 5003, NO-1432 Aas, Norway c Department of Chronic Diseases, Division of Epidemiology, Norwegian Institute of Public Health, Oslo, Norway

art ic l e i nf o

a b s t r a c t

Article history: Received 15 October 2012 Received in revised form 21 December 2013 Accepted 19 February 2014

This study examines the relationship between road traffic noise, self-reported sleep quality and mental health. The study is cross-sectional and based on data from a survey conducted in Oslo, Norway, in 2000. Psychological distress (Hopkins Symptom Checklist, HSCL-25) was measured along with self-reported somatic health, sleep quality, noise sensitivity and socioeconomic variables. Questionnaire data were combined with modeled estimates of noise exposure. The total study sample consisted of 2898 respondents. After adjustment for potential confounders and stratifying for sleep quality, we found a positive, but not statistically significant association between noise exposure and symptoms of psychological distress among participants with poor sleep quality (slope ¼ 0.06, 95% CI:  0.02 to 0.13, per 10 dB increase in noise exposure). In the same sleep quality group, we found a borderline statistically significant association between noise exposure and a symptom level indicating a probable mental disorder (HSCL Z 1.55) (odds ratio¼1.47, 95% CI: 0.99–1.98, per 10 dB increase in noise exposure). We found no association between road traffic noise and mental health among subjects reporting good and medium sleep quality. The results suggest that road traffic noise may be associated with poorer mental health among subjects with poor sleep. Individuals with poor sleep quality may be more vulnerable to effects of road traffic noise on mental health than individuals with better sleep quality. & 2014 Elsevier Inc. All rights reserved.

Keywords: Road traffic noise Psychological distress Mental disorder Sleep quality Hopkins Symptom Checklist

1. Introduction The World Health Organization (WHO, 2011) considers noise to be an environmental risk factor for poor health and a major environmental issue. In Norway, approximately 1.5 million persons (33%) are exposed to sound levels above 55 dB (dB) outside of their dwellings, which is the highest recommended average noise level. Road traffic is the largest source of noise annoyance in Norway (Englien et al., 2004). It is estimated that 3–6% of the Norwegian population experience severe noise annoyance, and that 2–3% are highly sleep disturbed due to road traffic noise (Aasvang, 2012). It has been postulated that prolonged negative feelings towards noise may increase the risk of more severe psychological problems (Cohen and Weinstein, 1981). Mental health is of global concern, and it is estimated that one in every four worldwide will be affected by a mental disorder at n

Corresponding author. Fax: þ 47 21 07 66 86. E-mail address: [email protected] (N.H. Krog). 1 Present address: Oslo University Hospital HF, P.O. Box 4950 Nydalen, 0424 Oslo, Norway. http://dx.doi.org/10.1016/j.envres.2014.02.010 0013-9351 & 2014 Elsevier Inc. All rights reserved.

some stage of life (WHO, 2001). The World Health Organization (WHO, 2001) has characterized mental and behavioral disorders as combinations of abnormal thoughts, emotions, behavior and relationships with others. A Norwegian review found that the lifetime prevalence of any mental disorder was around 40%, while the 12 months' prevalence ranged from approximately 10–33%, depending on the disorder (Mykletun et al., 2009). Annoyance and sleep disturbances are the most widespread and well-documented subjectively reported effects of environmental noise (WHO, 2011), but morning tiredness, headaches and milder psychological conditions have also been reported to be associated with noise in adult populations (Kluizenaar et al., 2011; Tarnopolsky et al., 1980; Öhrström et al., 1988; Stansfeld et al., 1996). Both aircraft and road traffic noise have been linked to psychological symptoms, but not to definable mental disorders (Stansfeld and Matheson, 2003; Tarnopolsky et al., 1978, 1980). However, an Italian study found a significant association between aircraft noise and an anxiety diagnosis (Hardoy et al., 2004). A survey by Rocha et al. (2012) found that noise perceived as an environmental problem was associated with the prevalence of common mental disorders as assessed with the General Health

18

K. Sygna et al. / Environmental Research 131 (2014) 17–24

Questionnaire (GHQ-12). Most previous research on noise and mental health has examined aircraft noise (Abey-Wickrama et al., 1969; Meecham and Smith, 1977; Hardoy et al., 2004; Schreckenberg et al., 2010; Tarnopolsky et al., 1978, 1980; Watkins et al., 1981), and only a few studies have addressed road traffic noise and mental health (Kishikawa et al., 2009; Stansfeld et al., 1993, 1996). Various methods of measuring mental health as well as noise exposure may have contributed to inconsistent findings. Furthermore, a potential causal association between noise and mental health problems may be limited to certain vulnerable and noise sensitive groups. Dratva et al. (2011) implied that vulnerable groups, such as people with preexisting diseases, are at greater risk than others to experience more severe health effects by transportation noise. A study from Japan reported a positive association between road traffic noise and “anxiety and insomnia”, but only among noise sensitive subjects (Kishikawa et al., 2009). Noise can affect sleep by increasing the time to fall asleep, induce awakenings and sleep stage changes, and thus reduce the total amount and quality of sleep (Aasvang et al., 2011; Basner et al., 2008; Brink, 2011; Griefahn et al., 2006; Öhrström et al., 1988). Furthermore, the close interrelationship between poor sleep quality and mental health problems is well acknowledged (Breslau et al., 1996; Ford and Kamerow, 1989; Neckelmann et al., 2007; Peterson and Benca, 2006; Sloan, 2011; Tsuno et al., 2005). Thus, there are reasons to believe that sleep quality may be an important factor in a potential association between traffic noise and mental health problems (Evans and Lepore, 2008; Pirrera et al., 2010), but the knowledge is still poor regarding the role of sleep in the noise–health relationship (Pirrera et al., 2010). The main aim of our study was to examine the relationship between road traffic noise and mental health. We wanted to examine whether road traffic noise contributes to increased levels of psychological distress in general, and whether it increases the risk of a mental disorder. Furthermore, we wanted to test the hypothesis that sleep quality modifies the noise–mental health relationship.

2. Materials and methods 2.1. Study sample This study is cross-sectional, using data from a survey carried out in Oslo, Norway, during autumn 2000. Data on residential addresses ( E21,000) were obtained from the Norwegian Public Roads Administration and the City of Oslo in connection with their ongoing work on residential noise mapping. Using the Norwegian National Population Register, we sampled from the residential addresses provided by the authorities 51% males and 49% females from different age strata (18–37, 38–57, 58–77, and 4 78), altogether 5390 persons. A total of 3262 persons (60.5%) returned the questionnaire. 364 individuals were excluded, either because they had recently moved (107), or they had not answered the questions about psychological distress (Hopkins Symptom Checklist) satisfactorily (257), leaving 2898 individuals in the study population. Informed consent was obtained from the respondents, and the survey was approved by the Regional Committee for Medical and Health Research Ethics in Norway. 2.2. Variables 2.2.1. Outcome variables Psychological distress was measured by Hopkins Symptom Checklist-25 (HSCL25). HSCL-25 is widely used in population surveys, and has proved to have satisfactory validity and reliability as a measure of psychological distress (Derogatis et al., 1974; Strand et al., 2003). The inventory consists of 25 items tapping symptoms of anxiety and depression, such as “suddenly scared for no reason”, “feeling fearful”, “trembling”, “poor appetite”, “feeling lonely”, or “crying easily”. The respondents were asked to rate how they were affected by each symptom: “not at all bothered”, “somewhat bothered”, “rather bothered”, or “very bothered”. In the present study, only 22 items from the original HSCL-25 were used. To avoid repetition, two items from the original HSCL-25 about sleeping problems were not included in our questionnaire, since this topic was covered elsewhere. In addition, one question about sexual activity was excluded, and an item about tinnitus, which was not in the original HSCL-25, was included. The mean score of

the HSCL items (ranging from 1 to 4) was calculated, to indicate the level of psychological distress of each respondent. The association between road traffic noise and mental health was examined in two ways. First, we wanted to investigate the association between road traffic noise exposure and any increase in symptoms of psychological distress, taking into account the whole range of symptom levels. For this purpose, the mean score of HSCL was used as a continuous variable. This approach was chosen since psychological distress exists on a continuous scale, and most previous studies have found an association only with milder degrees of psychological distress (Stansfeld et al., 1993; Stansfeld and Matheson, 2003). Furthermore, HSCL was dichotomized, to examine the association between road traffic noise and more severe levels of psychological distress, i.e. mental disorder with a potential need of being treated. A cut-off value of 1.55 was chosen, as this value has been proven to be appropriate when the purpose is to screen for probable psychiatric cases (Veijola et al., 2003). 2.2.2. Noise exposure assessment Road traffic noise was assessed at the most exposed façade of the home address of each participant in the study. Digital maps and geographical coordinates of the home addresses were used as basis for the noise exposure assessment. The road traffic noise exposure was calculated according to the Nordic prediction method for road traffic noise (Jonasson et al., 1996) for the year 2000, using the software program CadnaA (DataKustik, 2004) to calculate acoustic propagation and noise levels. The Nordic prediction method for road traffic noise calculates noise exposure at the most exposed façade with an accuracy of 73–5 dB, depending on the distance from the noise source (Jonasson et al., 1996). This method is designed to be valid up to a distance of 300 m from the road (Bendtsen, 1999). The development of the prediction models is based on standard noise emissions from road traffic. The standard noise emissions are based on a large number of sound measurements of passing cars under well-defined conditions, along with measurements of speed. Input data to CadnaA were digitalized terrain data, ground types, buildings and noise screens in 3D, in addition to road traffic data such as traffic counts, percentages of heavy vehicles, speed limits and diurnal distributions from the Norwegian Public Roads Administration and the City of Oslo. Lden at the most exposed façade, as defined according to the European Environmental Noise Directive (Directive 2002/49/EC, European Parliament and Council, 2002) was used as the noise metric. Lden is the A-weighted average sound pressure level over a 24 h period, in which levels during the evening (19.00–23.00) and night (23.00–07.00) are increased by 5 dB and 10 dB, respectively. Road traffic noise was used as a continuous variable in the analyses. 2.2.3. Sleep quality Self-reported sleep quality was measured by one question from the Basic Nordic Sleep Questionnaire (Partinen and Gislason, 1995): “How well do you usually sleep?” The original five categories were reduced to three: good sleep (well, rather well), medium sleep (neither well nor badly), and poor sleep (rather badly, badly). This single-item assessment of sleep was used to obtain a summary measure of the overall sleep quality. 2.2.4. Potential confounders Age, gender, socioeconomic status, somatic diseases and noise sensitivity are variables found to be associated with mental health (Prince et al., 2007; Rocha et al., 2012; Stansfeld et al., 1993), and were included as potential confounders in the analyses. Age was used as a continuous variable. We used income, education and employment status to measure socioeconomic status (Kristenson et al., 2003). The total income of the household was reported as o 200,000 NOK; 200,000– 400,000 NOK; 400,000–600,000 NOK; 600,000–800,000 NOK; 4800,000 NOK, and dichotomized into the categories “ o400,000 NOK” and “ 4400,000 NOK” for our analyses. The questionnaire contained two items on education. The first item asked for the highest level of education completed, with the following response categories: “did not complete primary school”, “primary school (6–7 years)”, “secondary school (8–10 years)”, or “high school/college”. The second item asked what kind of further training/education had been completed: “practical training”, “up to 1 year (same subject)”, “1–2 years (technical college/commercial school)”, “1–2 years (high school/university)”, “3–4 years (high school/university)”, or “more than 4 years of higher education”. These two variables were combined into three categories: “o 12 years of education”, “Z 12 years and o 15 years of education” and “Z 15 years of education”. Employment status was categorized as follows: “working outside home”, “working at home”, “student”, “retired”, and “unemployed or disabled”. This variable was dichotomized into two groups: “unemployed” (unemployed/disabled) and “employed/others” (working outside home/ working at home/student/retired). The variable somatic diseases was constructed as follows: the respondents who answered “yes” to at least one of the eight diseases included in the questionnaire (myocardial infarction, angina pectoris, stroke, high blood pressure, diabetes, frequent infectious diseases, metabolic disorder, and hearing loss) were included as having a somatic disease. To measure noise sensitivity, one six-point scale item from Weinstein's Noise Sensitivity Scale (Weinstein, 1978) was employed, in which the participants were asked to respond to the statement “I am sensitive to noise”. The six response categories were merged into three in the following way: “low sensitivity” (disagree strongly, disagree fairly),

K. Sygna et al. / Environmental Research 131 (2014) 17–24 “medium sensitivity” (disagree slightly, agree slightly), and “high sensitivity” (agree fairly, agree strongly).

19

Table 1 Sample characteristics. N

%

Total

2898

100.0

Gender Female Male Missing

1456 1442 0

50.2 49.8 0.0

Age r 25 26–35 36–50 51–65 Z 66 Missing

210 469 924 822 472 1

7.2 16.2 31.9 28.4 16.3 0.0

Income o 400,000 NOK 4 400,000 NOK Missing

1158 1666 74

40.0 57.5 2.6

Education Primary Secondary University Missing

827 677 1370 24

28.5 23.4 47.3 0.8

Residential history o 1 year 1–2 years 2–5 years 4 5 years Missing

238 207 499 1943 11

8.2 7.1 17.2 67.1 0.4

Employment Employed Unemployed Missing

1912 986 0

66.0 34.0 0.0

Somatic diseases No Yes Missing

1996 902 0

68.9 31.1 0.0

Noise sensitivity Low Medium High Missing

954 994 893 57

32.9 34.3 30.8 2.0

Noise o 50 50–54 55–59 60–64 Z 65 Missing

285 721 774 610 474 34

9.8 24.9 26.7 21.0 16.4 1.2

Sleep quality Good Medium Poor Missing

2140 451 277 30

73.8 15.6 9.6 1.0

Hopkins Symptom Checklist Mean (SD)

1.36 (0.38)

Hopkins Symptom Checklist o 1.55 Z 1.55 Missing

2124 528 246

2.3. Statistical analyses Data were analyzed using R version 2.15.0. In the main analyses we used both a linear regression model and a logistic regression model to investigate the association between noise exposure and psychological distress. Symptoms of psychological distress measured as a continuous variable (mean score of HSCL) was the outcome in the linear model, whereas probable mental disorder (cut-off value of Z 1.55 on HSCL) was applied as a dichotomous outcome variable in the logistic regression model. Road traffic noise exposure was entered as a continuous variable in both regression analyses. For the result presentation (Tables 4 and 5), the effect estimates were recalculated to show the effect per 10 dB increase in noise exposure. The analyses were performed both with crude data and with adjustment for the potential confounders described in Section 2.2.4. In a separate model, we included an interaction term between sleep quality and noise exposure to test whether sleep quality was an effect modifier for noise in association with psychological distress. As a consequence of investigating interactions between sleep quality and noise, we stratified on sleep quality in all main analyses. To better depict the associations between noise and psychological distress, we estimated a smooth function using non-parametric regression with cubic spline as a smoother. Such models are named Generalized Additive Models (GAM), and we used the GAM function in the R library to produce the figures. The figures are presented with 95% confidence limits. In addition, we performed a missing data analysis using multiple imputation techniques. We used the functions in the mice library (R), where five full datasets were produced. Each of the five datasets was analyzed, and the results combined into pooled effect measures. Information in the observed variables was used to fill in for the missing values. The filled-in values in the five sets of data were sampled from distributions for the missing variables. The functions in the mice library imputed incomplete multivariate data by chained equations (Van Buuren and Groothius-Oudshoorn, 2011). An important assumption for these methods is missing at random (MAR), which means that the structure of the missing data for the different variables should be independent of the variables themselves. However, we observed that the mean HSCL score was negatively correlated with the number of HSCL items answered (Spearman's correlation coefficient¼  0.31), which could indicate a violation of the MAR assumption. Thus, we excluded all participants who had answered less than 10 HSCL items (8.0%), reducing the study population to 2898. 2652 of these had answered all questions. For the complete case analyses, we included participants who had answered more than 17 of the 22 HSCL items. In the missing data analysis, we performed imputations also for the HSCL items, recalculating mean HSCL score and the dichotomous variable based on the 1.55 cut-off.

3. Results 3.1. Descriptive statistics Table 1 presents the characteristics of the study sample including missing values. We observed that most of the missing values were linked to the HSCL items, 246 in all; however, a relatively large fraction (94%) answered more than 17 items (defined as the complete cases). Table 1 also shows that the majority of the participants (91.4%) had lived in their residence for more than 1 year, which ensures that most respondents had been exposed to the estimated noise levels for a longer period of time. Table 2 shows the distribution of the variables divided into the three sleep quality categories. Low household income, being unemployed, having somatic diseases, and being noise sensitive were more frequently reported among participants with poor sleep quality than among participants with good sleep quality. Psychological distress was higher in the poor sleep quality group compared to the medium and good sleep quality groups. Concerning noise exposure, there was a slightly higher percentage of people with high noise exposure in the group with poor sleep quality compared to the group with good sleep quality. Table 3 presents the distribution of noise exposure for participants above and below the cut-off value of 1.55 on HSCL. The table shows that there was no systematic difference in noise exposure between the groups scoring below and above the cut-off 1.55 on HSCL. Further, the table presents the distribution of sleep quality in the dichotomized

73.3 18.2 8.5

groups of HSCL. There was a higher proportion of people with medium and poor sleep quality above the cut-off value of 1.55 on HSCL. In the separate model with noise and sleep as an interaction term, the interaction term was significantly associated with psychological distress (p o 0.001 for both continuous and dichotomous outcomes), showing a significant effect-modification of sleep quality in the association between noise and mental health.

20

K. Sygna et al. / Environmental Research 131 (2014) 17–24

Our a priori hypothesis concerning the role of sleep was therefore statistically justified. Due to the modifying role of sleep in the relation between noise and psychological distress, all further analyses were stratified by sleep quality.

3.2. Regression analyses We performed multiple imputations for both linear and logistic regression models in addition to the complete case analyses. For the multiple imputations, we fitted models where all explanatory variables were included (adjusted analysis). Results from the complete case (unadjusted and adjusted) and multiple imputations are reported.

3.2.1. Symptoms of psychological distress The results from the linear regression analyses are presented in Table 4, where the crude and adjusted effect measures are presented with 95% confidence intervals. The unadjusted results showed a statistically significant association between noise exposure and symptoms of psychological distress only among subjects with poor sleep quality. In the poor sleep quality group, the mean score of psychological distress increased by 0.08 (95% CI: 0.00– 0.15) per 10 dB increase in noise exposure. When adjusting for the potential confounders, the association attenuated slightly and was not statistically significant (slope ¼0.06, 95% CI:  0.02 to 0.13). The results from the multiple imputations showed a further attenuated effect in the poor sleep quality group. The results for the other sleep quality groups were unchanged. Fig. 1 presents the smoothed relationship between noise and symptoms of psychological distress (upper row). In the groups with medium and good sleep quality, the smoothed curves were quite flat (left and middle panel), in contrast to the slightly increasing tendency with increasing noise levels among those with poor sleep (right panel).

3.2.2. Probable mental disorder Table 5 shows the results of the logistic regression analysis. Estimated odds ratios (OR) are presented together with their 95% confidence intervals. The unadjusted results showed a significant association between noise exposure and probable mental disorder in the group with poor sleep quality, where the odds increased 1.47 (95% CI: 1.06–2.04) times per 10 dB increase in road traffic noise. As for the analyses of symptoms of psychological distress (linear regression model), there was no association between noise exposure and probable mental disorder among individuals with good and medium sleep quality. Among the individuals reporting poor sleep quality, the results of the adjusted model showed a borderline statistically significant association (OR¼ 1.40, 95% CI: 0.99–1.98) per 10 dB increase in road traffic noise exposure. However, the results from the multiple imputations were attenuated. Fig. 1 (bottom row) shows the associations between road traffic noise and risk of probable mental disorder parallel to the results for the continuous outcome variable (upper row). In the poor sleep quality group (right panel), a linear relationship was observed with an intensified risk of probable mental disorder for those exposed to the highest noise levels.

4. Discussion The results of this study showed a slight tendency of increased symptoms of psychological distress with increasing levels of traffic noise exposure among poor sleepers only. There was a similar weak association between road traffic noise exposure and probable mental disorder. An interaction term of noise and sleep was statistically significant, indicating sleep quality to be an effect modifier in the relationship between noise and mental health. To the best of our knowledge, only a few previous studies have reported on the association between objectively measured road traffic noise and mental health in adults (Kishikawa et al., 2009; Stansfeld et al., 1993, 1996). In accordance with Kishikawa et al. (2009), we found a weak positive association in a sub-group of potentially susceptible individuals, in our case those experiencing

Fig. 1. Functions of noise and Hopkins Symptom Checklist for different groups of sleep quality.

K. Sygna et al. / Environmental Research 131 (2014) 17–24

21

Table 2 Sample characteristics for the different sleep quality groups (excluding missing values).

Gender Female Male Age r 25 26–35 36–50 51–65 Z 66

Good sleep quality

Medium sleep quality

Poor sleep quality

1046 (48.9%) 1094 (51.1%)

231 (51.2%) 220 (48.8%)

164 (59.2%) 113 (40.8%)

171 373 693 571 331

(8.0%) (17.4%) (32.4%) (26.7%) (15.5%)

23 55 138 149 86

(5.1%) (12.2%) (30.6%) (33.0%) (19.1%)

16 34 86 94 47

(5.8%) (12.3%) (31.0%) (33.9%) (17.0%)

Income o 400,000 NOK 4400,000 NOK

804 (38.5%) 1284 (61.5%)

207 (47.3%) 231 (52.7%)

136 (50.7%) 132 (49.3%)

Education Primary Secondary University

572 (26.9%) 509 (23.9%) 1047 (49.2%)

153 (34.5%) 105 (23.7%) 185 (41.8%)

93 (33.9%) 56 (20.4%) 125 (45.6%)

Residential history o 1 year 1–2 years 2–5 years 45 years

187 155 363 1432

33 25 74 319

17 25 59 176

Employment Employed Unemployed

1462 (68.3%) 678 (31.7%)

261 (57.9%) 190 (42.1%)

172 (62.1%) 105 (37.9%)

Somatic diseases No Yes

1531 (71.5%) 609 (28.5%)

283 (62.7%) 168 (37.3%)

160 (57.8%) 117 (42.2%)

Noise sensitivity Low Medium High

761 (36.2%) 721 (34.3%) 619 (29.5%)

122 (27.6%) 169 (38.2%) 151 (34.2%)

64 (23.7%) 91 (33.7%) 115 (42.6%)

203 561 598 432 320

55 92 104 107 89

Noise r 50 50–54 55–59 60–64 Z 65 Hopkins Symptom Checklist Mean (SD) Hopkins Symptom Checklist o 1.55 Z 1.55

(8.8%) (7.3%) (17.0%) (67.0%)

(9.6%) (26.5%) (28.3%) (20.4%) (15.1%)

1.28 (0.30) 1830 (85.5%) 310 (14.5%)

Table 3 Distribution of road traffic noise, sleep quality and Hopkins Symptom Checklist ( o 1.55 or Z 1.55). Hopkins Symptom Checklist o 1.55 Traffic noise o50 dB 50–54 dB 55–59 dB 60–65 dB Z65 dB Sleep quality Good Medium Poor

208 553 567 443 327

(9.9%) (26.4%) (27.0%) (21.1%) (15.6%)

1728 (82.0%) 270 (12.8%) 109 (5.2%)

Z 1.55

52 131 134 112 93

(10.0%) (25.1%) (25.7%) (21.5%) (17.8%)

268 (51.2%) 130 (24.9%) 125 (23.9%)

poor sleep quality. Kishikawa et al. (2009) reported a positive association between road traffic noise and “anxiety and insomnia”, but only among noise sensitive subjects. A study by Stansfeld et al. (1993) showed an increasing trend of psychiatric caseness with increasing noise in the lowest and middle tertiles of noise

(7.3%) (5.5%) (16.4%) (70.7%)

(12.3%) (20.6%) (23.3%) (23.9%) (19.9%)

23 61 67 63 59

(6.1%) (9.0%) (21.3%) (63.5%)

(8.4%) (22.3%) (24.5%) (23.1%) (21.6%)

1.48 (0.41)

1.76 (0.54)

300 (66.5%) 151 (33.5%)

121 (43.7%) 156 (56.3%)

sensitivity using Weinstein's noise sensitivity scale. A similar trend was not demonstrated among the most noise sensitive persons, which contradicted the hypothesis that noise sensitive subjects are more susceptible to the effects of noise on mental health. Noise sensitivity was also included among the covariates in our analyses. However, while previous studies focused especially on the role of noise sensitivity, we, in our study, specifically investigated the possible modifying role of sleep quality on the relationship between road traffic noise and psychological distress. As far as we know, few studies have examined the role of sleep disturbances in this relationship (Pirrera et al., 2010). Knowing that sleep is important for physiological and mental restoration (Berglund et al., 2000; Breslau et al., 1996; Sloan, 2011), we suggest that people with poor sleep quality may have reduced coping abilities and may thus be more susceptible to psychological distress. Stansfeld and Clark (2008) proposed that noise might be more harmful to health in situations where other stressors interact. According to their model, sleep disturbance might act as a stressor, and affect the level of psychological distress. In a longitudinal study, Neckelmann et al. (2007) found that poor sleep might be a risk factor for later development of mental health problems. The effects of noise on sleep are well documented (Aasvang et al.,

22

K. Sygna et al. / Environmental Research 131 (2014) 17–24

Table 4 Association between road traffic noise exposure and Hopkins Symptom Checklist for each sleep quality group. Linear regression analysis. All (N ¼2774)

Good sleep quality (n ¼2050)

Medium sleep quality (n ¼450)

Poor sleep quality (n¼ 274)

Slope

95% CI

Slope

95% CI

Slope

95% CI

Slope

95% CI

Unadjusted per 10 dB increase in noise exposure

0.03

(0.01–0.05)

0.01

(  0.01 to 0.03)

 0.01

(  0.06 to 0.05)

0.08

(0.00–0.15)

Adjusteda per 10 dB increase in noise exposure

0.02

(0.00–0.04)

0.01

(  0.01 to 0.03)

 0.01

(  0.06 to 0.04)

0.06

(  0.02 to 0.13)

0.02

(0.00–0.03)

0.01

(  0.01 to 0.03)

 0.01

(  0.06 to 0.04)

0.02

(  0.06 to 0.1)

b

Multiple imputation per 10 dB increase in noise exposure a b

Adjusted for sex, age, education, employment, noise sensitivity, and somatic diseases. Participants who answered more than 17 items were included in the complete-case analysis.

Table 5 Association between road traffic noise exposure and Hopkins Symptom Checklist Z 1.55 for each sleep quality group. Logistic regression analysis. All (N ¼ 2774)

Good sleep quality (n¼2050)

Medium sleep quality (n¼ 450)

Poor sleep quality (n¼ 274)

OR

95% CI

OR

95% CI

OR

95% CI

OR

95% CI

Unadjusted per 10 dB increase in noise exposure

1.10

(0.97–1.26)

0.97

(0.82–1.16)

1.01

(0.77–1.33)

1.47

(1.06–2.04)

Adjusteda per 10 dB increase in noise exposure

1.05

(0.92–1.21)

0.94

(0.79–1.13)

0.98

(0.73–1.31)

1.40

(0.99–1.98)

Multiple imputationb per 10 dB increase in noise exposure

1.01

(0.99–1.03)

1.00

(0.98–1.02)

1.00

(0.94–1.06)

1.06

(0.98–1.13)

a b

Adjusted for sex, age, education, employment, noise sensitivity, and somatic diseases. Participants who answered more than 17 items were included in the complete-case analysis.

2008, 2011; Basner et al., 2008; Griefahn et al., 2006), and poor sleep quality may thus be an intermediate factor in the association between noise and mental health. However, from our cross sectional study we cannot decide whether poor sleep was independent of noise or was an intermediate factor in the noise– mental health relationship. Prospective findings from the Caerphilly study (Stansfeld et al., 1996) revealed no association between road traffic noise and overall psychiatric disorder, but there was some evidence for an association with anxiety. Thus, our overall results are in line with most previous studies, suggesting only a weak association between road traffic noise and psychological distress. In addition to looking at symptoms of psychological distress by using mean score of HSCL, we defined another mental health outcome using a cut-off value of HSCL (Z1.55). Both the 1.55 and 1.75 cut-off values have been used in previous studies of mental health status measured by HSCL (Gamperiene et al., 2008; Veijola et al., 2003). Although the specificity of the cut-off value of 1.75 has been shown to be somewhat better than the specificity of the lower cut-off, the cut-off of 1.55 has been recommended due to higher sensitivity (Veijola et al., 2003). The cut-off value of 1.55 is a rather low value meant to screen for possible psychiatric disorder. It could be argued that this is not very different from using a mean score of HSCL, and rather similar results were obtained for the two outcomes of HSCL used in our study. The majority of studies exploring the association between environmental noise and mental health with adequate control for potential confounders have reported weak positive associations with milder psychological symptoms, but not with manifest psychiatric disorders (Stansfeld et al., 1993, Stansfeld and Matheson, 2003), although some studies have reported an association between noise and more severe mental health problems (Hardoy et al., 2004; Rocha et al., 2012). However, the study by Hardoy et al. (2004) examined effects of aircraft noise, and was based on a very small sample (N¼71). In addition, the noise exposure was

defined solely on the basis of distance to the airport. Nor in the study by Rocha et al. (2012) was the noise exposure objectively assessed, but the results demonstrated an increasing gradient in the prevalence of common mental disorders with increasing numbers of self-reported environmental problems, including noise. Thus, the evidence for an association between traffic noise and more severe mental health problems is still poor. Our findings provide some indication of an association between road traffic noise and more severe mental health problems, when adjusting for other risk factors. There were some limitations in our study that should be considered. Initially, the large number of missing values was a limitation. Most of the missing values were linked to the psychological distress variable, and we observed that individuals with high scores on a few HSCL items were more likely not to answer the full set of HSCL items. This is a violation of the MAR assumption, and the results from the missing data analysis should be interpreted with care. We excluded participants who had not answered the HSCL items satisfactorily. For these participants we have no knowledge about the association between noise and mental health, and we do not know why they answered only a few of the HSCL items. They could represent a vulnerable group of individuals, or they may have restricted their answers to the few items which seemed relevant. Another weakness in the present study was the cross-sectional design, which precluded causal inferences. We have shown a borderline significant association between traffic noise and mental health among individuals with poor sleep quality, but we were not able to investigate whether poor sleep quality was caused by noise or by other factors. Individuals with poor mental health might evaluate their environment negatively, and thus experience higher levels of noise annoyance than others (Clark and Stansfeld, 2007; Tarnopolsky et al., 1978). However, in our study we employed objectively estimated noise exposure and not subjectively measured noise annoyance. Cross-sectional evidence should be treated

K. Sygna et al. / Environmental Research 131 (2014) 17–24

with caution; still, we consider it unlikely that psychological distress should lead to higher objectively estimated road traffic noise exposure. There is a possibility, though, that psychological distress might affect individual noise exposure through socioeconomic mechanisms. First and foremost this might be true for severe and disabling mental disorders, affecting the ability to work, and thus the economic resources governing the overall opportunity to freely choose where to live. It has been assumed that a process of self-selection might take place, i.e. noise sensitive persons would either not move into, or they would leave, areas with high noise levels. However, a Dutch study did not find evidence of a process of self-selection based on noise sensitivity (Nijland et al., 2007). Furthermore, our study population in the group with poor sleep quality was rather small (n ¼274). Thus, our findings need to be further investigated in a larger population. Another limitation in our study was the use of a single question to assess sleep quality. However, single item assessments of overall sleep quality have previously been shown to provide a useful summary measure of sleep quality (Cappelleri et al., 2009). While rankings and experiences will vary among individuals, a common set of key components that constitute overall sleep quality is likely (Cappelleri et al., 2009). A study of the subjective meaning of sleep quality among individuals with insomnia and normal sleepers found that both groups defined sleep quality by tiredness on waking and throughout the day, feeling rested and restored on waking, and the number of awakenings experienced during the night (Harvey et al., 2008). Moreover, there is a general risk of misclassification in self-reported questionnaires. Since both sleep quality and psychological distress were measured by questionnaire, there was some risk of shared response bias. Still, the possible misclassification of health problems was assumed to be independent of noise (non-differential), and the effect measures would therefore be attenuated. When the data were collected, the overall purpose to examine effects of noise was not revealed to the participants, to avoid selection and information bias. Instead, the survey was introduced as a general study of health and wellbeing. One of the strengths in the present study was that noise exposure was objectively assessed, using the Nordic prediction method for road traffic noise that has been widely used. Also, our study population was exposed to a wide range of road traffic noise levels. However, only noise exposure assessed for the most exposed façade was used in the present study. How well this reflected individual noise exposure depends on a number of factors including the layout of the residence and hours spent at home (outdoors and indoors), factors which were not included in the present study. Thus, the possibility of misclassification due to underestimation or overestimation of true individual noise exposure cannot be excluded. A critical point for the accuracy of the noise exposure assessment is the traffic flow. However, the equivalent noise level is quite robust against changes in traffic flow; a 26% change in traffic density equals a change of 71 dB. At lower levels of traffic noise, i.e. those modeled for dwellings near roads with low traffic or more distant to a major road, the modeled noise levels are in general more uncertain than for roads with heavy traffic (Alberola et al., 2005). In the present study, we used the best available input data collected in a database regarding traffic flow, diurnal distribution and speed limits. We cannot exclude impact from other noise sources, as noise from several sources (railway, noise from neighbors, aircraft, etc.) are common in urban settings. However, we excluded all addresses that were simultaneously included by the authorities in the railway noise mapping program. Thus, we can safely assume that none of the participants in this study population were exposed to high levels of railway noise (24-h average Z35 dB indoors, inside of the most exposed façade). The possible influence of exposure to aircraft

23

noise is considered negligible since the nearest airport (Oslo Airport, Gardermoen) is located 47 km north of Oslo. The prevalence of respondents exposed to road traffic noise 455 dB (64.1%) was higher in our study population than the prevalence (33%) reported by Englien et al. (2004) for the total Norwegian population. This is reasonable, since our study population was drawn from the largest city in Norway with a higher noise exposure compared to smaller agglomerations. Another strength in our study was that we employed a widely used and validated instrument for assessing psychological distress, the Hopkins Symptom Checklist (Derogatis et al., 1974; Strand et al., 2003). Finally, we included a rich set of potential confounders in the analyses of the association between noise and mental health. Although we cannot exclude the possibility of residual confounding, this reduces the possibility of spurious effects.

5. Conclusions We found only weak indications of an association between road traffic noise exposure and mental health, and only among individuals with poor sleep quality. The number of individuals with poor sleep quality was, however, rather modest in our sample, and more research is needed to gain further understanding of the potential long-term effects of road traffic noise on mental health. Both longitudinal studies as well as studies of potential mechanisms and possible vulnerable groups are needed.

Conflict of interest The authors declare that they have no competing financial interests or other conflict of interest.

Ethics The survey was approved by the Regional Committee for Medical and Health Research Ethics in Norway. Informed consent was obtained from all the respondents.

Financial support The data collection was financially supported by the Norwegian Research Council and the Norwegian Public Roads Administration. The present work was given financial support by the Norwegian Public Roads Administration. The funding sources had no involvement in the scientific work or the decision to submit the article for publication.

Acknowledgments The authors wish to thank the study participants. Further, we thank Karin Melsom for language revision. References Aasvang, G.M., 2012. Helsebelastning som skyldes veitrafikkstøy i Norge. The Norwegian Institute of Public Health, Oslo (in Norwegian). Aasvang, G.M., Moum, T., Engdahl, B., 2008. Self-reported sleep disturbance due to railway noise: exposure–response relationships for nighttime equivalent and maximum noise level. J. Acoust. Soc. Am. 124 (1), 257–268. Aasvang, G.M., Øverland, B., Moum, T., Ursin, R., 2011. A field study of effects of road traffic and railway noise onpolysomnographic sleep parameters. J. Acoust. Soc. Am. 129 (6), 3716–3726. Abey-Wickrama, I., A'Brook, M.F., Gattoni, F.E., Herridge, C.F., 1969. Mental-hospital admissions and aircraft noise. Lancet 2 (7633), 1275–1277.

24

K. Sygna et al. / Environmental Research 131 (2014) 17–24

Alberola, J., Flindell, I.H., Bullmore, A.J., 2005. Variability in road traffic noise levels. Appl. Acoust. 66, 1180–1195. Basner, M., Glatz, C., Griefahn, B., Penzel, T., Samel, A., 2008. Aircraft noise: effects on macro- and microstructure of sleep. Sleep Med. 9, 382–387, http://dx.doi. org/10.1016/j.sleep.2007.07.002. Bendtsen, H., 1999. The Nordic prediction method for road traffic noise. Sci. Total Environ. 235 (1), 331–338. Berglund, B., Lindvall, T., Schwela, D., Goh, K.T., 2000. Guidelines for Community Noise. World Health Organization, Geneve. Breslau, N., Roth, T., Rosenthal, T., Andreski, P., 1996. Sleep disturbance and psychiatric disorders: a longitudinal epidemiological study of young adults. Biol. Psychiatry 39, 411–418. Brink, M., 2011. Parameters of well-being and subjective health and their relationship with residential traffic noise exposure – a representative evaluation in Switzerland. Environ. Int. 37 (4), 723–733, http://dx.doi.org/10.1016/j. envint.2011.02.011. Cappelleri, J.C., Bushmakin, A.G., McDermott, A.M., Sadosky, A.B., Petrie, C.D., Martin, S., 2009. Psychometric properties of a single-item scale to assess sleep quality among individuals with fibromyalgia. Health Qual. Life Out. 7 (art no. 54). Clark, C., Stansfeld, S.A., 2007. The effect of transportation noise on health andcognitive development: a review of recent evidence. Int. J. Comp. Psychol. 20 (2), 145–158. Cohen, S., Weinstein, N., 1981. Nonauditory effects of noise on behavior and health. J. Soc Issues 37 (1), 36–70. DataKustik, 2004. CadnaA, Computer Aided Noise Abatement, Manual ver. 3.4. 2004 Computer Program. Derogatis, L., Lipman, R.S., Rickels, K., et al., 1974. The Hopkins Symptom Checklist (HSCL): a self-report symptom inventory. Behav. Sci. 19 (1), 1–15. Dratva, J., Phuleria, H.C., Foraster, M., Gaspoz, J.M., Keidel, D., Künzli, N., et al., 2011. Transportation noise and blood pressure in a population-based sample of adults. Environ. Health Perspect. 120, 50–55. Englien, E., Haakonsen, G., Steinnes, M., 2004. Støyplage i Norge. Resultater fra førstegenerasjons modell for beregning av antall støyutsatte og SPI. Notater 2004/43. Statistisk Sentralbyrå, Oslo (in Norwegian). European Parliament and Council, 2002. EU Directive 2002/49/EC of the European Parliament and of the Council of 25 June 2002 Relating to the Assessment and Management of Environmental Noise. Evans, G.W., Lepore, S.J., 2008. Psychosocial processes linking the environment and mental health. In: Freeman, H., Stansfeld, S. (Eds.), The Impact of the Environment on Psychiatric Disorder. Routledge/Taylor & Francis Group, New York, pp. 127–157. Ford, D., Kamerow, D., 1989. Epidemiologic study of sleep disturbances and psychiatric disorders. An opportunity for prevention? J. Am. Med. Assoc. 262 (11), 1479–1484, http://dx.doi.org/10.1001/jama.1989.03430110069030. Gamperiene, M., Nygård, J.F., Sandanger, I., Lau, B., Bruusgaard, D., 2008. Selfreported work ability of Norwegian women in relation to physical and mental health, and to the work environment. J. Occup. Med. Toxicol. 3 (8), http://dx.doi. org/10.1186/1745-6673-3-8. Griefahn, B., Marks, A., Robens, S., 2006. Noise emitted from road, rail and air traffic and their effects on sleep. J. Sound Vib. 295, 129–140, http://dx.doi.org/10.1016/ j.jsv.2005.12.052. Hardoy, M.C., Carta, M.G., Marci, A.R., Carbone, F., Cadeddu, M., Kovess, V., Dell'Osso, L., Carpiniello, B., 2004. Exposure to aircraft noise and risk of psychiatric disorders: the Elmas survey. Soc. Psychiatr. Epidemiol. 40, 24–26, http://dx.doi. org/10.1007/s00127-005-0837-x. Harvey, A.G., Stinson, K., Whitaker, K.L., Moskovitz, D., Virk, H., 2008. The subjective meaning of sleep quality: a comparison of individuals with and without insomnia. Sleep 31, 383–393. Jonasson, H., Kragh, J., Parmanen, J., Sandberg, U., Storeheier, S., 1996. Road Traffic Noise: Nordic Prediction Method, Nordic Council of Ministers, TemaNord, Report 525. Kishikawa, H., Matsui, T., Uchiyama, I., Miyakawa, M., Hiramatsu, K., Stansfeld, S.A., 2009. Noise sensitivity and subjective health: questionnaire study conducted along trunk roads in Kusatsu, Japan. Noise Health 11, 111–117. Kluizenaar, Y., Salomons, E.M., Janssen, S.A., van Lenthe, F.J., Vos, H., Zhou, H., Miedema, H.M.E., Mackendach, J.P., 2011. Urban road traffic noise and annoyance: the effect of a quiet façade. J. Acoust. Soc. Am. 130 (4), 1936–1942. Kristenson, M., Eriksen, H.R., Sluiter, J.K., Starke, D., Ursin, H., 2003. Psychobiological mechanisms of socioeconomic differences in health. Soc. Sci. Med. 58 (8), 1511–1522, http://dx.doi.org/10.1016/S0277-9532(03)00536-8.

Meecham, W.C., Smith, H.G., 1977. Effects of jet aircraft noise on mental hospital admissions. Br. J. Audiol. 11 (3), 81–85. Mykletun, A., Knudsen, A.K., Mathiesen, K.S., 2009. Psykiske lidelser i Norge: Et folkehelseperspektiv. Rapport 2009:8. The Norwegian Institute of Public Health, Oslo (in Norwegian). Neckelmann, D., Mykletun, A., Dahl, A., 2007. Chronic insomnia as a risk factor for developing anxiety and depression. Sleep 30 (7), 873–880. Nijland, H.A., Hartemink, S., Van Kamp, I., Van Wee, B., 2007. The influence of sensitivity for road traffic noise on residential location: does it trigger a process of spatial selection? J. Acoust. Soc. Am. 122 (3), 1595–1601. Öhrström, E., Rylander, R., Björkman, M., 1988. Effects of night time road traffic noise – an overview of laboratory andfield studies on noise dose and subjective noise sensitivity. J. Sound Vib. 127 (3), 441–448. Partinen, M., Gislason, T., 1995. Basic Nordic Sleep Questionnaire (BNSQ): a quantitated measure of subjective sleep complaints. J. Sleep Res. 4, 150–155. Peterson, M., Benca, R., 2006. Sleep in mood disorders. Psychiatr. Clin. N. Am. 29, 1009–1032, http://dx.doi.org/10.1016/j.psc.2006.09.003. Pirrera, S., De Valck, E., Cluydts, R., 2010. Nocturnal road traffic noise: a review on its assessment and consequences on sleep and health. Environ. Int. 36 (5), 492–498, http://dx.doi.org/10.1016/j.envint.2010.03.007. Prince, M., Patel, V., Saxena, S., Maj, M., Maselko, J., Philips, M., Rahman, A., 2007. No health without mental health. Lancet 370, 859–877, http://dx.doi.org/10.1016/ S0140-6736(07)61238-0. Rocha, K., Pèrez, K., Rodriguez-Sanz, M., Obiols, J., Borrell, C., 2012. Perception of environmental problems and common mental disorders (CMD). Soc. Psychiatr. Epidemiol. 47 (10), 1675–1684, http://dx.doi.org/10.1007/s00127-012-0474-0. Schreckenberg, D., Griefahn, B., Meis, M., 2010. The association between noise sensitivity, reported physical and mental health, perceived environmental quality, and noise annoyance. Noise Health 12 (46), 7–16, http://dx.doi.org/ 10.4103/1463-1741.59995. Sloan, E., 2011. Sleep deprivation and postpartum mental health. Case report. Arch. Womens Ment. Health 14, 509–511, http://dx.doi.org/10.1007/s00737-0110247-x. Strand, B.H., Dalgard, O.S., Tambs, K., Rognerud, M., 2003. Measuring the mental health status of the Norwegian population: a comparison of the instruments SCL-25, SCL-10, SCL-5 and MHI-5 (SF-36). Nord. J. Psychiatry 57, 113–118, http: //dx.doi.org/10.1080/08039480310000932. Stansfeld, S., Sharp, D.S., Gallacher, J., Babisch, W., 1993. Road traffic noise, noise sensitivity and psychological disorder. Psychol. Med. 23, 977–985. Stansfeld, S., Gallacher, J., Babisch, W., Shipley, M., 1996. Road traffic noise and psychiatric disorder: prospective findings from the Caerphilly Study. Br. Med. J. 313 (7052), 266–267. Stansfeld, S.A., Matheson, M.P., 2003. Noise pollution: non-auditory effects on health. Br. Med. Bull. 68, 243–257. Stansfeld, S., Clark, C., 2008. Noise and psychiatric disorder. In: Freeman, H., Stansfeld, S. (Eds.), The Impact of the Environment on Psychiatric Disorder. Routledge/Taylor & Francis Group, New York, pp. 242–267. Tarnopolsky, A., Barker, S.M., Wiggins, R.D., McLean, E.K., 1978. The effect of aircraft noise on the mental health of a community sample: a pilot study. Psychol. Med. 8, 219–233. Tarnopolsky, A., Watkins, G., Hand, D.J., 1980. Aircraft noise and mental health: I. Prevalence of individual symptoms. Psychol. Med. 10, 683–698. Tsuno, N., Besset, A., Ritchie, K., 2005. Sleep and depression. J. Clin. Psychiatry 66 (10), 1254–1269. Van Buuren, S., Groothius-Oudshoorn, K., 2011. Mice: multiple imputation by chained equations in R. J. Stat. Softw. 45 (3) (Available at)〈http://www.jstat. org/〉. Veijola, J., Jokelainen, J., Läksy, K., Kantojärvi, L., Kokkonen, P., Järvelin, M.R., Joukamaa, M., 2003. The Hopkins Symptom Checklist-25 in screening DSMIII-R axis-I disorders. Nord. J. Psychiatry 57, 119–123. Watkins, G., Tarnopolsky, A., Jenkins, L.M., 1981. Aircraft noise and mental health: II. Use of medicines and health care services. Psychol. Med., 11; , pp. 155–168. Weinstein, N.D., 1978. Individual differences in reaction to noise: a longitudinal study in a college dormitory. J. Appl. Psychol. 63 (4), 458–466. WHO, 2001. The World Health Report 2001: Mental Health: New Understanding, New Hope. World Health Organization, Switzerland. WHO, 2011. Burden of Disease from Environmental Noise. Quantification of Healthy Life Years Lost in Europe. World Health Organization, Denmark.