A comparison of occupational and nonoccupational noise exposures in Sweden Richard L. Neitzel1, Eva B. Svensson2, Stephanie K. Sayler3, Johnson Ann-Christin2 Department of Environmental Health Sciences and Risk Science Center, University of Michigan, Ann Arbor, MI, USA, 2Department of Clinical Science Intervention and Technology, Karolinska Institutet, Stockholm, Sweden, 3Department of Environmental Health Sciences, University of Michigan, Ann Arbor, MI, USA 1
Abstract This study was conducted to evaluate noise exposures and the contributions of occupational and nonoccupational activities among three groups of Swedish workers (office workers, day care workers, and military flight technicians), and to evaluate risk factors for elevated hearing threshold levels. Forty-five subjects were recruited across the three groups. Each subject completed a risk factor questionnaire along with Békésy audiometry at frequencies between 125 and 8000 Hz. Subjects also wore a noise dosimeter continuously for 1 week, and documented their occupational and nonoccupational activities using a time-activity log. Subjects in all groups completed >7400 h of dosimetry, and had weekly exposures between 76 and 81 dBA. Day care workers had the highest daily exposures, and flight technicians had the highest weekly exposures. Most daily and weekly exposures exceeded the 70 dBA exposure limit recommended for prevention of any hearing loss. Subjects’ perceptions of their exposures generally agreed well with measured noise levels. Among office workers, exposures were predominately nonoccupational, while among flight technicians nonoccupational and occupational activities contributed roughly equally, and among day care workers occupational exposures were dominant. Extreme exposures and cumulative noise exposure were associated with an increased risk of hearing threshold levels >10 dB hearing level. Effective hearing loss prevention programs may be needed in occupations not historically considered to be at high risk of noise-induced hearing loss (e.g., day care workers). Prevention efforts need to address nonoccupational exposures as well as occupational exposures, as nonoccupational activities may present the dominant risk of noise-induced hearing loss for some workers. Keywords: Exposure assessment, hearing loss risk, noise, nonoccupational, occupational, source apportionment
Introduction Noise exposure is a ubiquitous but often overlooked occupational and environmental hazard. While the causal relationship between noise and noise-induced hearing loss (NIHL) has been known for hundreds of years, there is increasing evidence that a host of other health effects are associated with exposure to lower levels of noise. These effects include coronary heart disease,[1] hypertension,[2] stress,[3] and sleep disturbance.[4] The personal, social, and economic impacts of NIHL are substantial, and these additional health effects may present an even larger public health burden. Access this article online Quick Response Code:
Website: www.noiseandhealth.org DOI: 10.4103/1463-1741.140503 PubMed ID: ***
Noise & Health, September-October 2014, Volume 16:72, 270-278
One factor that hampers our understanding of these health effects is inadequate noise exposure assessment. A number of European studies have developed and utilized sophisticated models designed to evaluate community-level exposures to specific sources of noise, and especially road[5] and airport[6] noise. The key shortcoming of these studies is their focus on single sources of noise. Without holistic assessment of exposure from all occupational and nonoccupational activities, these studies cannot evaluate total noise exposure, and this limitation may inadvertently limit our understanding of noiserelated health effects. Model-based assessments are certainly useful for determining the fraction of a population at risk of exposure over a given level — for example, the 24-h limit of 70 A-weighted decibels (dBA) recommended by the World Health Organization (WHO) and the US Environmental Protection Agency (EPA)[7,8] — from a particular source. However, such studies cannot determine the contribution of specific sources of noise to an individual’s total exposure. Individual-level dosimetry measurements with simultaneous assessment of activities are needed to adequately assess the contribution of different exposure sources to health risk. Unfortunately, only a few studies[9-11] have conducted such 270
Neitzel, et al.: Occupational and nonoccupational noise exposures in Sweden
assessments, and due to study design issues many of these dosimetry-based studies have not been able to explore the contribution of specific occupational and nonoccupational activities to total noise exposure and subsequent health effects.
Noise dosimetry
To better characterize the importance of occupational and nonoccupational noise exposures to risk of NIHL, we evaluated noise exposures in several groups of Swedish workers. The goals of our study were threefold. The first was to assess, in a holistic fashion, total noise exposures over a short period (1 week). The second was to evaluate the contributions of occupational and nonoccupational activities to total noise exposures. The third was to explore the relationship between observed hearing threshold levels (HTLs) and noise exposures and other risk factors.
To accomplish our first study goal, evaluation of noise exposures, subjects were issued a Larson Davis 706RC (Larson Davis, USA) dosimeter at the outset of the 1-week study period. Dosimeters data logged average noise levels at 1-min intervals, and were configured with a 70 dBA criterion level, 24 h criterion time, 3 dB time-intensity exchange rate, and 40-110 dB measurement range. Previous assessments have had limited ability to evaluate low-level nonoccupational exposures due to use of a 70-140 dB occupational measurement range.[22] Dosimeters were worn during waking hours, and kept nearby when sleeping or bathing. Dosimeters were recovered by research staff at the end of the 1-week measurement period, and were calibrated pre- and post-measurement.
Methods
Time-activity log
We selected three occupations to evaluate. The first group was office workers, which we expected to have occupational noise exposures of ~60 dBA.[12] The second group was day care workers, with expected exposures of ~70 dBA.[13,14] The final group was flight technicians employed by the Swedish Air Force, anticipated to have exposures of ~90 dBA[15,16] Our target sample was 15 workers per group (n = 45). Office workers employed by the Karolinska Institutet near Stockholm, Sweden, were approached individually during normal work hours. Day care workers employed at 5 day care centers within a single community near Stockholm were approached at scheduled center meetings during normal work hours. Military flight technicians at two Swedish airbases in southern Sweden were approached at scheduled meetings during normal work hours. All potential subjects were given an overview of the study, and interested individuals completed informed consent forms. Study procedures were reviewed and approved by the Regional Ethical Review Board in Stockholm. Each subject completed a range of procedures. These included a one-time questionnaire; noise dosimetry over a 1-week period; completion of a time-activity log during dosimetry; and otoscopy and audiometry sometime during the 1-week period. Questionnaire Subjects completed a 31-item questionnaire, written in Swedish, which collected information concerning demographics; hearing health; perceived hearing ability; job history and work experience; “typical” perceived occupational noise exposure and use of hearing protection devices (HPDs); and noisy leisure time activities and use of HPDs. The questionnaire, as a whole or in parts, has been used previously[17-19] An English version of the perceived noise item has been used[20] and validated[21] in US workers. 271
Subjects were given a time-activity log, written in Swedish. The log, similar to one we have used previously,[22] allowed subjects to report the timing and duration of their activities during their dosimetry measurement. The log also allowed subjects to report their use of headphones (which would add additional exposure not detected by the dosimeter), their use of HPDs, and their perceived noise levels (via the same item used in the questionnaire). Time-activity log information, combined with dosimetry-derived noise levels, was used to achieve our second study goal, evaluation of the contributions of occupational and nonoccupational activities to total noise exposures. Otoscopy and audiometry Otoscopy was performed prior to audiometric testing. Tests on office workers were conducted in an audiometric test booth at the Karolinska Institutet. Tests on day care workers and flight technicians were conducted in quiet rooms at each day care center and airbase, respectively. Background noise levels were not measured but rooms were judged to be very quiet by research staff. Békésy audiometry was conducted using a laptop PC running a program called “Fixfrekvens Békésy Audiogram” (Technical Audiology Department, Karolinska Institutet) an external sound card (UGM96, ESI, Germany), and Sennheiser superaural headphones (HAD200, Sennheiser, Germany). HTLs in dB hearing level (HL) were measured at frequencies between 125 and 8000 Hz. These audiometric data, when combined with dosimetry and questionnaire information, were used to address our third study goal, exploring the relationship between increased HTLs and noise exposures and other risk factors. Analyses Descriptive analyses were conducted for survey responses and for audiometric test results at individual frequencies and arithmetically averaged across 3, 4, and 6 kHz. Age was analyzed in four categories (10 dB HL 4 kHz Left Right 6 kHz Left Right Average 3, 4, 6 kHz Left Right Subjects>20 dB HL 4 kHz Left Right 6 kHz Left Right Average 3, 4, 6 kHz Left Right
Overall (n = 45) n Mean SD
Hearing threshold level (dB HL) Office workers (n = 15) Day care workers (n = 16) n Mean SD n Mean SD
Flight technicians (n = 14) n Mean SD
43 43
7.1 7.9
9.5 10.3
14 14
10.6 6.8
12.5 8.4
15 15
6.4 10.5
8.1 12.2
14 14
4.5 3.6
6.5 9.3
43 43
10.1 8.4
9.8 9.6
14 14
11.3 7.5
10.4 7.1
15 15
10.3 10.8
8.2 9.2
14 14
8.8 6.7
11.3 12.2
43 43
8.0 7.1
8.1 8.5
14 14
10.3 6.9
9.8 6.3
15 15
8.4 10.0
6.7 8.8
14 14
5.2 4.3
7.2 9.6
13 14
18.1 18.0
9.2 9.3
5 4
23.4 17.8
12.7 5.0
5 7
15.9 19.5
5.4 12.2
3 3
12.9 14.8
0.7 7.0
20 14
18.8 18.6
7.1 9.4
7 5
19.5 14.2
7.8 4.4
8 6
16.7 18.9
5.3 9.1
5 3
21.1 25.1
9.2 14.8
16 11
16.1 17.4
6.3 8.0
4 4
23.1 14.2
8.4 4.8
8 5
13.8 18.4
3.9 10.7
4 2
13.7 21.1
1.5 5.7
3 4
32.9 29.1
6.7 10.1
2 1
36.7 23.7
2.0 —
1 2
25.3 34.9
— 13.0
0 1
— 22.9
— —
4 5
29.7 29.0
7.5 8.0
1 1
35.4 20.2
— —
1 2
26.6 29.8
— 6.8
2 2
28.4 32.6
11.0 10.1
3 4
26.7 26.0
7.4 6.6
2 1
28.7 21.4
9.2 —
1 2
22.6 28.8
— 9.7
0 1
— 25.1
— —
*Difference between groups P < 0.05, Chi-square. SD = Standard deviation, dB HL = Decibels hearing level
Unlike the daily exposures, flight technicians had the highest weekly LEQ(168) exposures [mean 83.5 dBA, Figure 1c], and office workers the lowest (mean 75.2 dBA). Weekly LEQ(168) exposures had a mean level of 78.6 ± 8.9 dBA. Variability in LEQ(168) exposures was smallest among day care workers and greatest among flight technicians. Exceedance fractions were generally high [Table 3]. Nearly, 87% of measurements exceeded 70 dBA. Exceedance fractions were highest among day care workers and lowest for office workers, and were lower on weekends than weekdays. All weekly LEQ(168) values for day care workers exceeded 70 dBA, compared to 87% for flight technicians and 73% for office workers. At the weekly level, about 43% of total exposure came from occupational activities [Table 3]. This percentage ranged from 24% (office workers) to 55% (day care workers). For day care workers and flight technicians, as well as overall, exposure during weekdays was greater for occupational than nonoccupational activities. Among office workers, occupational activities never contributed more than nonoccupational activities on weekdays. Occupational Noise & Health, September-October 2014, Volume 16
activities during weekend days contributed only a small fraction of exposure. The relationship between measured mean 1-min LEQ levels and categories of perceived noise from time-activity logs was monotonic and strong for the four lowest categories for office workers and flight technicians (data not shown). The trend was somewhat weaker among day care workers, and failed for all groups at the fifth category. Trends were statistically significant for all three groups and overall. Office workers had a mean cumulative LEQ, Cumi exposure of 77.5 ± 12.5 dBA, lower than day care workers (79.6 ± 4.7 dBA) and flight technicians (82.5 ± 7.5 dBA) [Figure 1d]. As with LEQ(168) exposures, variability in LEQ, Cumi was smallest among day care workers and greatest among flight technicians. When continuous exposures were collapsed into four categories, 13 of 15 (87%) office workers were in categories 1 (70 P (% dose) n % >70 P (% dose) dBA Work Nonwork dBA Work Nonwork dBA Work Nonwork Work Nonwork 86.7 42.5 57.5 15 73.3 24.3 75.7 16 100 55.4 44.6 14 85.7 48.1 51.9
44 44 45 45 45
68.9 80.0 76.3 82.2 71.1
57.2 59.9 62.4 60.8 46.0
42.8 40.1 37.6 39.2 54
14 14 15 15 15
46.7 66.7 66.7 46.7 40
35.1 27.2 25.0 37.5 37.2
64.9 72.8 75 62.5 62.8
16 16 16 16 16
100 100 93.8 100 100
73.9 83.0 83.3 78.7 68.7
26.1 17 16.7 21.3 31.3
14 14 14 14 14
57.1 71.4 64.3 100 71.4
60.5 66.9 80.0 65.6 34.9
39.5 33.1 20 34.4 65.1
44 44
60.0 36.0
5.0 6.4
95 93.6
14 14
53.3 26.7
0 8.1
100 91.9
16 16
68.8 50.0
0 0.2
100 99.8
14 14
57.1 28.6
16.4 12.3
83.6 87.7
n Weekly LEQ(168) Weekday LEQ(24) Monday Tuesday Wednesday Thursday Friday Weekend LEQ(24) Saturday Sunday
% >70 dBA
dBA = A-weighted decibels
a
c
b
d
Figure 1: Measured and estimated noise exposures. Red reference lines indicated recommended 24-h exposure limit. (a) Measured Activity-specific LEQ (n = 45 subjects). (b) Measured LEQ(24) (n = 311 days). (c) Measured LEQ(168) (n = 45 subjects). (d) Estimated LEQ, (n = 45 subjects) Cum
Risk factors for elevated hearing thresholds Risk factors were not varied across frequencies or between ears within frequency in our logistic regression analyses [Table 4]. Only two factors, extreme noise exposure and categorized LEQ,Cumi exposure, appeared in models at all frequencies. Extreme noise exposure was protective in the left ear, but was a significant risk factor in the right ear. Cumulative LEQ,Cumi was a significant risk factor for the 4 kHz and average 3, 4, and 6 kHz models. Categorized age was a significant risk factor at 6 kHz and average 3, 4, and 6 kHz. Job tenure was also a significant factor at 275
4 and 6 kHz. Overall, the models indicated that extreme noise exposure (an indicator of possible acoustic trauma) and chronic noise exposures were consistent risk factors for elevated HTLs.
Discussion This study provides useful insights into noise exposures, the contributions of occupational and nonoccupational activities towards total exposure, and risk of NIHL for three diverse groups of Swedish workers (office workers, day care workers, Noise & Health, September-October 2014, Volume 16
Neitzel, et al.: Occupational and nonoccupational noise exposures in Sweden Table 4: Logistic regression models for risk of HTL >10 dB HL Frequency/variable 4 kHz Extreme noise exposure Cumulative LEQ category Painkillers Job tenure Tinnitus 6 kHz Extreme noise exposure Age category Cumulative LEQ category Ear infections Family hearing loss HPD user Painkillers Job tenure Tinnitus Average 3, 4, and 6 kHz Extreme noise exposure Age category Cumulative LEQ category
Model pseudo R2 0.23
Left ear OR
SE
P value
0.17
0.17
0.07
6.31 1.14 0.05
6.92 0.08 0.08
0.09 0.04 0.06
0.26 6.04 3.01
0.20 3.80 1.78
0.08 0.004 0.06
0.38
Model pseudo R2 0.26
Right ear OR
SE
P value
1.16
0.09
0.02
1.17 0.04
0.08 0.07
0.01 0.04
4.31
2.89
0.03
4.36 0.09 0.03
3.73 0.13 0.05
0.09 0.09 0.03
1.26 0.07
0.11 0.11
0.007 0.09
4.41 1.25
2.81 0.07
0.02 0.04
0.45
5.05
4.92
0.10
0.29
0.24 0.37 3.27
0.22 1.61
0.09 0.02
HTL = Hearing threshold level, dB HL = Decibels hearing level, OR = Odds ratio, HPD = Hearing protection device, SE = Standard error
and flight technicians). Our analyses suggest that most workers in these groups are exposed above limits recommended to prevent NIHL in the public, that nonoccupational exposures can contribute a substantial fraction of total exposure, that workers’ perceived exposures are highly correlated with measured exposures, and that workers with even moderate exposures may be at risk of elevated hearing thresholds from noise. The first aim of our study was to characterize noise exposures among the three groups of workers assessed. The occupational exposures of the groups did not meet our expectations. While we anticipated work exposures of approximately 60, 70 and 85 dBA for the office workers, day care workers, and flight technicians, respectively, observed exposures were around 70, 82, and 84 dBA. Group exposure levels were ordered as expected but showed smaller-than-expected separation. The unexpectedly high occupational exposures among day care workers compared to the expected levels is likely a reflection of noisy activities associated with children (e.g., noisy play, singing or playing of musical instruments, shouting, speech, etc.,), whereas the unexpectedly high occupational exposures among office workers may be due to the social and meal gatherings which are common in Swedish office environments. The vast majority (about 88%) of weekly average exposuresand a slightly lower fraction of LEQ(24) exposures - exceeded the WHO/EPA 24-h exposure limit of 70 dBA. This finding is remarkably consistent with other dosimetry studies on urban dwellers in the United States (70% exceedance),[10] Spain (84% exceedance),[9] and China (85% exceedance).[11] Noise & Health, September-October 2014, Volume 16
LEQ(24) exposures among these various studies have ranged from 74.9[9] to 79 dBA,[10] again quite consistent with our mean LEQ(24) of 73.6 dBA. This consistency increases our confidence in the generalizability of our findings, and also suggests similarities in total noise exposures in diverse urban environments around the globe. The noise levels and contribution of exposures associated with various occupational and nonoccupational activities generally agree with previous dosimetry-based studies. The levels measured by Diaz and Pedrero[9] on workers in Madrid are similar to those measured here for home, shopping, sleep, and commuting, though weekend days were noisier than weekdays, in contrast to our findings. Nonoccupational activities assessed by Diaz and Pedrero contributed about 65% of total noise dose, which compares well to the nearly 58% nonoccupational contribution measured here. Zheng et al.[11] found that, on average, work contributed about 70% of total noise dose for office workers and teachers in Beijing, compared to 24 and 55%, respectively, in the current study. Levels for sleep are lower here than previously reported in the US[22] and China.[11] This is due at least in part to our low measurement range (40-110 dB) and absence of a 70 or 80 dBA threshold level, which more accurately captured low-level exposures, but which may have resulted in censoring of very high exposures. Self-assessments of “typical” occupational noise exposure levels generally showed good agreement with 1-min LEQ and LEQ(24) exposure levels. These findings are generally consistent with the performance of this self-report noise scale in US workers,[20,21] as well as musicians[23] and others,[24] and 276
Neitzel, et al.: Occupational and nonoccupational noise exposures in Sweden
suggest that self-report can be a useful measure of noise levels over short (i.e., days to weeks) and possibly longer periods. Our second aim was to evaluate the contributions of occupational and nonoccupational activities to total exposure. The average individual received 57.5% of their weekly exposure from nonoccupational activities. The distribution of exposure differed between groups: Office workers received about 76% of their exposure from nonoccupational activities, flight technicians about 52%, and day care workers about 45%. On weekend days the vast majority of exposure was from nonoccupational activities, and weekend days had lower LEQ(24) noise exposures than weekdays. These results suggest that workers with low occupational exposure, such as office workers, may receive the majority of their exposure from nonoccupational activities, but also suggest that workers with somewhat higher occupational exposures can still receive a majority of their exposure from nonoccupational activities. This finding is consistent with our recent study of noise in New York City.[25] We noted statistically significant (though generally small) differences in mean nonoccupational LEQ levels between groups, suggesting that workers may have systematic differences in the way in which they are exposed during nonoccupational activities. Nevertheless, certain aspects of nonoccupational exposure were shared among all groups: For example, sleep and time at home had the lowest mean LEQ levels, while social activities had the highest, consistent with other dosimetry-based studies.[11,22] Our third and final aim was to evaluate the association between noise exposure and other risk factors for NIHL and elevated hearing thresholds. We noted that HTLs were generally worse at 6 kHz than at 4 kHz, a finding that highlights the need to include 8 kHz HTLs in any audiometric analysis in order to confirm that the audiometric configuration is consistent with NIHL, e.g., that a “noise notch” is present. Risk factors that were consistently associated with elevations in HTLs included extreme noise exposures, categorized cumulative LEQC exposure, age category, and job tenure. Our model results were likely influenced by two factors. First, most subjects had normal HTLs — that is, thresholds
Abstract This study was conducted to evaluate noise exposures and the contributions of occupational and nonoccupational activities among three groups of Swedish workers (office workers, day care workers, and military flight technicians), and to evaluate risk factors for elevated hearing threshold levels. Forty-five subjects were recruited across the three groups. Each subject completed a risk factor questionnaire along with Békésy audiometry at frequencies between 125 and 8000 Hz. Subjects also wore a noise dosimeter continuously for 1 week, and documented their occupational and nonoccupational activities using a time-activity log. Subjects in all groups completed >7400 h of dosimetry, and had weekly exposures between 76 and 81 dBA. Day care workers had the highest daily exposures, and flight technicians had the highest weekly exposures. Most daily and weekly exposures exceeded the 70 dBA exposure limit recommended for prevention of any hearing loss. Subjects’ perceptions of their exposures generally agreed well with measured noise levels. Among office workers, exposures were predominately nonoccupational, while among flight technicians nonoccupational and occupational activities contributed roughly equally, and among day care workers occupational exposures were dominant. Extreme exposures and cumulative noise exposure were associated with an increased risk of hearing threshold levels >10 dB hearing level. Effective hearing loss prevention programs may be needed in occupations not historically considered to be at high risk of noise-induced hearing loss (e.g., day care workers). Prevention efforts need to address nonoccupational exposures as well as occupational exposures, as nonoccupational activities may present the dominant risk of noise-induced hearing loss for some workers. Keywords: Exposure assessment, hearing loss risk, noise, nonoccupational, occupational, source apportionment
Introduction Noise exposure is a ubiquitous but often overlooked occupational and environmental hazard. While the causal relationship between noise and noise-induced hearing loss (NIHL) has been known for hundreds of years, there is increasing evidence that a host of other health effects are associated with exposure to lower levels of noise. These effects include coronary heart disease,[1] hypertension,[2] stress,[3] and sleep disturbance.[4] The personal, social, and economic impacts of NIHL are substantial, and these additional health effects may present an even larger public health burden. Access this article online Quick Response Code:
Website: www.noiseandhealth.org DOI: 10.4103/1463-1741.140503 PubMed ID: ***
Noise & Health, September-October 2014, Volume 16:72, 270-278
One factor that hampers our understanding of these health effects is inadequate noise exposure assessment. A number of European studies have developed and utilized sophisticated models designed to evaluate community-level exposures to specific sources of noise, and especially road[5] and airport[6] noise. The key shortcoming of these studies is their focus on single sources of noise. Without holistic assessment of exposure from all occupational and nonoccupational activities, these studies cannot evaluate total noise exposure, and this limitation may inadvertently limit our understanding of noiserelated health effects. Model-based assessments are certainly useful for determining the fraction of a population at risk of exposure over a given level — for example, the 24-h limit of 70 A-weighted decibels (dBA) recommended by the World Health Organization (WHO) and the US Environmental Protection Agency (EPA)[7,8] — from a particular source. However, such studies cannot determine the contribution of specific sources of noise to an individual’s total exposure. Individual-level dosimetry measurements with simultaneous assessment of activities are needed to adequately assess the contribution of different exposure sources to health risk. Unfortunately, only a few studies[9-11] have conducted such 270
Neitzel, et al.: Occupational and nonoccupational noise exposures in Sweden
assessments, and due to study design issues many of these dosimetry-based studies have not been able to explore the contribution of specific occupational and nonoccupational activities to total noise exposure and subsequent health effects.
Noise dosimetry
To better characterize the importance of occupational and nonoccupational noise exposures to risk of NIHL, we evaluated noise exposures in several groups of Swedish workers. The goals of our study were threefold. The first was to assess, in a holistic fashion, total noise exposures over a short period (1 week). The second was to evaluate the contributions of occupational and nonoccupational activities to total noise exposures. The third was to explore the relationship between observed hearing threshold levels (HTLs) and noise exposures and other risk factors.
To accomplish our first study goal, evaluation of noise exposures, subjects were issued a Larson Davis 706RC (Larson Davis, USA) dosimeter at the outset of the 1-week study period. Dosimeters data logged average noise levels at 1-min intervals, and were configured with a 70 dBA criterion level, 24 h criterion time, 3 dB time-intensity exchange rate, and 40-110 dB measurement range. Previous assessments have had limited ability to evaluate low-level nonoccupational exposures due to use of a 70-140 dB occupational measurement range.[22] Dosimeters were worn during waking hours, and kept nearby when sleeping or bathing. Dosimeters were recovered by research staff at the end of the 1-week measurement period, and were calibrated pre- and post-measurement.
Methods
Time-activity log
We selected three occupations to evaluate. The first group was office workers, which we expected to have occupational noise exposures of ~60 dBA.[12] The second group was day care workers, with expected exposures of ~70 dBA.[13,14] The final group was flight technicians employed by the Swedish Air Force, anticipated to have exposures of ~90 dBA[15,16] Our target sample was 15 workers per group (n = 45). Office workers employed by the Karolinska Institutet near Stockholm, Sweden, were approached individually during normal work hours. Day care workers employed at 5 day care centers within a single community near Stockholm were approached at scheduled center meetings during normal work hours. Military flight technicians at two Swedish airbases in southern Sweden were approached at scheduled meetings during normal work hours. All potential subjects were given an overview of the study, and interested individuals completed informed consent forms. Study procedures were reviewed and approved by the Regional Ethical Review Board in Stockholm. Each subject completed a range of procedures. These included a one-time questionnaire; noise dosimetry over a 1-week period; completion of a time-activity log during dosimetry; and otoscopy and audiometry sometime during the 1-week period. Questionnaire Subjects completed a 31-item questionnaire, written in Swedish, which collected information concerning demographics; hearing health; perceived hearing ability; job history and work experience; “typical” perceived occupational noise exposure and use of hearing protection devices (HPDs); and noisy leisure time activities and use of HPDs. The questionnaire, as a whole or in parts, has been used previously[17-19] An English version of the perceived noise item has been used[20] and validated[21] in US workers. 271
Subjects were given a time-activity log, written in Swedish. The log, similar to one we have used previously,[22] allowed subjects to report the timing and duration of their activities during their dosimetry measurement. The log also allowed subjects to report their use of headphones (which would add additional exposure not detected by the dosimeter), their use of HPDs, and their perceived noise levels (via the same item used in the questionnaire). Time-activity log information, combined with dosimetry-derived noise levels, was used to achieve our second study goal, evaluation of the contributions of occupational and nonoccupational activities to total noise exposures. Otoscopy and audiometry Otoscopy was performed prior to audiometric testing. Tests on office workers were conducted in an audiometric test booth at the Karolinska Institutet. Tests on day care workers and flight technicians were conducted in quiet rooms at each day care center and airbase, respectively. Background noise levels were not measured but rooms were judged to be very quiet by research staff. Békésy audiometry was conducted using a laptop PC running a program called “Fixfrekvens Békésy Audiogram” (Technical Audiology Department, Karolinska Institutet) an external sound card (UGM96, ESI, Germany), and Sennheiser superaural headphones (HAD200, Sennheiser, Germany). HTLs in dB hearing level (HL) were measured at frequencies between 125 and 8000 Hz. These audiometric data, when combined with dosimetry and questionnaire information, were used to address our third study goal, exploring the relationship between increased HTLs and noise exposures and other risk factors. Analyses Descriptive analyses were conducted for survey responses and for audiometric test results at individual frequencies and arithmetically averaged across 3, 4, and 6 kHz. Age was analyzed in four categories (10 dB HL 4 kHz Left Right 6 kHz Left Right Average 3, 4, 6 kHz Left Right Subjects>20 dB HL 4 kHz Left Right 6 kHz Left Right Average 3, 4, 6 kHz Left Right
Overall (n = 45) n Mean SD
Hearing threshold level (dB HL) Office workers (n = 15) Day care workers (n = 16) n Mean SD n Mean SD
Flight technicians (n = 14) n Mean SD
43 43
7.1 7.9
9.5 10.3
14 14
10.6 6.8
12.5 8.4
15 15
6.4 10.5
8.1 12.2
14 14
4.5 3.6
6.5 9.3
43 43
10.1 8.4
9.8 9.6
14 14
11.3 7.5
10.4 7.1
15 15
10.3 10.8
8.2 9.2
14 14
8.8 6.7
11.3 12.2
43 43
8.0 7.1
8.1 8.5
14 14
10.3 6.9
9.8 6.3
15 15
8.4 10.0
6.7 8.8
14 14
5.2 4.3
7.2 9.6
13 14
18.1 18.0
9.2 9.3
5 4
23.4 17.8
12.7 5.0
5 7
15.9 19.5
5.4 12.2
3 3
12.9 14.8
0.7 7.0
20 14
18.8 18.6
7.1 9.4
7 5
19.5 14.2
7.8 4.4
8 6
16.7 18.9
5.3 9.1
5 3
21.1 25.1
9.2 14.8
16 11
16.1 17.4
6.3 8.0
4 4
23.1 14.2
8.4 4.8
8 5
13.8 18.4
3.9 10.7
4 2
13.7 21.1
1.5 5.7
3 4
32.9 29.1
6.7 10.1
2 1
36.7 23.7
2.0 —
1 2
25.3 34.9
— 13.0
0 1
— 22.9
— —
4 5
29.7 29.0
7.5 8.0
1 1
35.4 20.2
— —
1 2
26.6 29.8
— 6.8
2 2
28.4 32.6
11.0 10.1
3 4
26.7 26.0
7.4 6.6
2 1
28.7 21.4
9.2 —
1 2
22.6 28.8
— 9.7
0 1
— 25.1
— —
*Difference between groups P < 0.05, Chi-square. SD = Standard deviation, dB HL = Decibels hearing level
Unlike the daily exposures, flight technicians had the highest weekly LEQ(168) exposures [mean 83.5 dBA, Figure 1c], and office workers the lowest (mean 75.2 dBA). Weekly LEQ(168) exposures had a mean level of 78.6 ± 8.9 dBA. Variability in LEQ(168) exposures was smallest among day care workers and greatest among flight technicians. Exceedance fractions were generally high [Table 3]. Nearly, 87% of measurements exceeded 70 dBA. Exceedance fractions were highest among day care workers and lowest for office workers, and were lower on weekends than weekdays. All weekly LEQ(168) values for day care workers exceeded 70 dBA, compared to 87% for flight technicians and 73% for office workers. At the weekly level, about 43% of total exposure came from occupational activities [Table 3]. This percentage ranged from 24% (office workers) to 55% (day care workers). For day care workers and flight technicians, as well as overall, exposure during weekdays was greater for occupational than nonoccupational activities. Among office workers, occupational activities never contributed more than nonoccupational activities on weekdays. Occupational Noise & Health, September-October 2014, Volume 16
activities during weekend days contributed only a small fraction of exposure. The relationship between measured mean 1-min LEQ levels and categories of perceived noise from time-activity logs was monotonic and strong for the four lowest categories for office workers and flight technicians (data not shown). The trend was somewhat weaker among day care workers, and failed for all groups at the fifth category. Trends were statistically significant for all three groups and overall. Office workers had a mean cumulative LEQ, Cumi exposure of 77.5 ± 12.5 dBA, lower than day care workers (79.6 ± 4.7 dBA) and flight technicians (82.5 ± 7.5 dBA) [Figure 1d]. As with LEQ(168) exposures, variability in LEQ, Cumi was smallest among day care workers and greatest among flight technicians. When continuous exposures were collapsed into four categories, 13 of 15 (87%) office workers were in categories 1 (70 P (% dose) n % >70 P (% dose) dBA Work Nonwork dBA Work Nonwork dBA Work Nonwork Work Nonwork 86.7 42.5 57.5 15 73.3 24.3 75.7 16 100 55.4 44.6 14 85.7 48.1 51.9
44 44 45 45 45
68.9 80.0 76.3 82.2 71.1
57.2 59.9 62.4 60.8 46.0
42.8 40.1 37.6 39.2 54
14 14 15 15 15
46.7 66.7 66.7 46.7 40
35.1 27.2 25.0 37.5 37.2
64.9 72.8 75 62.5 62.8
16 16 16 16 16
100 100 93.8 100 100
73.9 83.0 83.3 78.7 68.7
26.1 17 16.7 21.3 31.3
14 14 14 14 14
57.1 71.4 64.3 100 71.4
60.5 66.9 80.0 65.6 34.9
39.5 33.1 20 34.4 65.1
44 44
60.0 36.0
5.0 6.4
95 93.6
14 14
53.3 26.7
0 8.1
100 91.9
16 16
68.8 50.0
0 0.2
100 99.8
14 14
57.1 28.6
16.4 12.3
83.6 87.7
n Weekly LEQ(168) Weekday LEQ(24) Monday Tuesday Wednesday Thursday Friday Weekend LEQ(24) Saturday Sunday
% >70 dBA
dBA = A-weighted decibels
a
c
b
d
Figure 1: Measured and estimated noise exposures. Red reference lines indicated recommended 24-h exposure limit. (a) Measured Activity-specific LEQ (n = 45 subjects). (b) Measured LEQ(24) (n = 311 days). (c) Measured LEQ(168) (n = 45 subjects). (d) Estimated LEQ, (n = 45 subjects) Cum
Risk factors for elevated hearing thresholds Risk factors were not varied across frequencies or between ears within frequency in our logistic regression analyses [Table 4]. Only two factors, extreme noise exposure and categorized LEQ,Cumi exposure, appeared in models at all frequencies. Extreme noise exposure was protective in the left ear, but was a significant risk factor in the right ear. Cumulative LEQ,Cumi was a significant risk factor for the 4 kHz and average 3, 4, and 6 kHz models. Categorized age was a significant risk factor at 6 kHz and average 3, 4, and 6 kHz. Job tenure was also a significant factor at 275
4 and 6 kHz. Overall, the models indicated that extreme noise exposure (an indicator of possible acoustic trauma) and chronic noise exposures were consistent risk factors for elevated HTLs.
Discussion This study provides useful insights into noise exposures, the contributions of occupational and nonoccupational activities towards total exposure, and risk of NIHL for three diverse groups of Swedish workers (office workers, day care workers, Noise & Health, September-October 2014, Volume 16
Neitzel, et al.: Occupational and nonoccupational noise exposures in Sweden Table 4: Logistic regression models for risk of HTL >10 dB HL Frequency/variable 4 kHz Extreme noise exposure Cumulative LEQ category Painkillers Job tenure Tinnitus 6 kHz Extreme noise exposure Age category Cumulative LEQ category Ear infections Family hearing loss HPD user Painkillers Job tenure Tinnitus Average 3, 4, and 6 kHz Extreme noise exposure Age category Cumulative LEQ category
Model pseudo R2 0.23
Left ear OR
SE
P value
0.17
0.17
0.07
6.31 1.14 0.05
6.92 0.08 0.08
0.09 0.04 0.06
0.26 6.04 3.01
0.20 3.80 1.78
0.08 0.004 0.06
0.38
Model pseudo R2 0.26
Right ear OR
SE
P value
1.16
0.09
0.02
1.17 0.04
0.08 0.07
0.01 0.04
4.31
2.89
0.03
4.36 0.09 0.03
3.73 0.13 0.05
0.09 0.09 0.03
1.26 0.07
0.11 0.11
0.007 0.09
4.41 1.25
2.81 0.07
0.02 0.04
0.45
5.05
4.92
0.10
0.29
0.24 0.37 3.27
0.22 1.61
0.09 0.02
HTL = Hearing threshold level, dB HL = Decibels hearing level, OR = Odds ratio, HPD = Hearing protection device, SE = Standard error
and flight technicians). Our analyses suggest that most workers in these groups are exposed above limits recommended to prevent NIHL in the public, that nonoccupational exposures can contribute a substantial fraction of total exposure, that workers’ perceived exposures are highly correlated with measured exposures, and that workers with even moderate exposures may be at risk of elevated hearing thresholds from noise. The first aim of our study was to characterize noise exposures among the three groups of workers assessed. The occupational exposures of the groups did not meet our expectations. While we anticipated work exposures of approximately 60, 70 and 85 dBA for the office workers, day care workers, and flight technicians, respectively, observed exposures were around 70, 82, and 84 dBA. Group exposure levels were ordered as expected but showed smaller-than-expected separation. The unexpectedly high occupational exposures among day care workers compared to the expected levels is likely a reflection of noisy activities associated with children (e.g., noisy play, singing or playing of musical instruments, shouting, speech, etc.,), whereas the unexpectedly high occupational exposures among office workers may be due to the social and meal gatherings which are common in Swedish office environments. The vast majority (about 88%) of weekly average exposuresand a slightly lower fraction of LEQ(24) exposures - exceeded the WHO/EPA 24-h exposure limit of 70 dBA. This finding is remarkably consistent with other dosimetry studies on urban dwellers in the United States (70% exceedance),[10] Spain (84% exceedance),[9] and China (85% exceedance).[11] Noise & Health, September-October 2014, Volume 16
LEQ(24) exposures among these various studies have ranged from 74.9[9] to 79 dBA,[10] again quite consistent with our mean LEQ(24) of 73.6 dBA. This consistency increases our confidence in the generalizability of our findings, and also suggests similarities in total noise exposures in diverse urban environments around the globe. The noise levels and contribution of exposures associated with various occupational and nonoccupational activities generally agree with previous dosimetry-based studies. The levels measured by Diaz and Pedrero[9] on workers in Madrid are similar to those measured here for home, shopping, sleep, and commuting, though weekend days were noisier than weekdays, in contrast to our findings. Nonoccupational activities assessed by Diaz and Pedrero contributed about 65% of total noise dose, which compares well to the nearly 58% nonoccupational contribution measured here. Zheng et al.[11] found that, on average, work contributed about 70% of total noise dose for office workers and teachers in Beijing, compared to 24 and 55%, respectively, in the current study. Levels for sleep are lower here than previously reported in the US[22] and China.[11] This is due at least in part to our low measurement range (40-110 dB) and absence of a 70 or 80 dBA threshold level, which more accurately captured low-level exposures, but which may have resulted in censoring of very high exposures. Self-assessments of “typical” occupational noise exposure levels generally showed good agreement with 1-min LEQ and LEQ(24) exposure levels. These findings are generally consistent with the performance of this self-report noise scale in US workers,[20,21] as well as musicians[23] and others,[24] and 276
Neitzel, et al.: Occupational and nonoccupational noise exposures in Sweden
suggest that self-report can be a useful measure of noise levels over short (i.e., days to weeks) and possibly longer periods. Our second aim was to evaluate the contributions of occupational and nonoccupational activities to total exposure. The average individual received 57.5% of their weekly exposure from nonoccupational activities. The distribution of exposure differed between groups: Office workers received about 76% of their exposure from nonoccupational activities, flight technicians about 52%, and day care workers about 45%. On weekend days the vast majority of exposure was from nonoccupational activities, and weekend days had lower LEQ(24) noise exposures than weekdays. These results suggest that workers with low occupational exposure, such as office workers, may receive the majority of their exposure from nonoccupational activities, but also suggest that workers with somewhat higher occupational exposures can still receive a majority of their exposure from nonoccupational activities. This finding is consistent with our recent study of noise in New York City.[25] We noted statistically significant (though generally small) differences in mean nonoccupational LEQ levels between groups, suggesting that workers may have systematic differences in the way in which they are exposed during nonoccupational activities. Nevertheless, certain aspects of nonoccupational exposure were shared among all groups: For example, sleep and time at home had the lowest mean LEQ levels, while social activities had the highest, consistent with other dosimetry-based studies.[11,22] Our third and final aim was to evaluate the association between noise exposure and other risk factors for NIHL and elevated hearing thresholds. We noted that HTLs were generally worse at 6 kHz than at 4 kHz, a finding that highlights the need to include 8 kHz HTLs in any audiometric analysis in order to confirm that the audiometric configuration is consistent with NIHL, e.g., that a “noise notch” is present. Risk factors that were consistently associated with elevations in HTLs included extreme noise exposures, categorized cumulative LEQC exposure, age category, and job tenure. Our model results were likely influenced by two factors. First, most subjects had normal HTLs — that is, thresholds
