Effects of noise and acoustics in schools on vocal health in teachers.

Previous studies on the influence of noise and acoustics in the classroom on voice symptoms among teachers have exclusively relied on self-reports. Since self-reported physical conditions may be biased, it is important to determine the role of objective measurements of noise and acoustics in the presence of voice symptoms. To assess the association between objectively measured and self-reported physical conditions at school with the presence of voice symptoms among teachers. In 12 public schools in Bogotα, we conducted a cross-sectional study among 682 Colombian school workers at 377 workplaces. After signed the informed consent, participants filled out a questionnaire on individual and work-related conditions and the nature and severity of voice symptoms in the past month. Short-term environmental measurements of sound levels, temperature, humidity, and reverberation time were conducted during visits at the workplaces, such as classrooms and offices. Logistic regression analysis was used to determine associations between work-related factors and voice symptoms. High noise levels outside schools (odds ratio [OR] = 1.83; 95% confidence interval [CI]: 1.12-2.99) and self-reported poor acoustics at the workplace (OR = 2.44; 95% CI: 1.88-3.53) were associated with voice symptoms. We found poor agreement between the objective measurements and self-reports of physical conditions at the workplace. This study indicates that noise and acoustics may play a role in the occurrence of voice symptoms among teachers. The poor agreement between objective measurements and self-reports of physical conditions indicate that these are different entities, which argue for inclusion of physical measurements of the working environment in studies on the influence of noise and acoustics on vocal health.

Introduction

The places where we live and work can present hazards to our health and wellbeing.[1] Several studies in school environments have investigated the effects of environmental factors, such as noise levels, acoustic conditions, and indoor air quality, on children's health and performance.[2345678910] However, these environmental factors may also influence the health and wellbeing of teachers.[1112] Teachers have been recognized as one of the largest groups of professional voice users.[1314] In general, voice disorders are more prevalent among teachers than in other occupational groups, but the reported prevalences range from 9%[14] to 94%,[15] strongly depending on definition of severity and duration of symptoms. This wide range hampers a clear evaluation of the contribution of work-related factors to the occurrence of voice disorders among teachers.

Previous studies on work-related factors of voice disorders among teachers have relied on self-reported physical conditions, such as high background noise and poor acoustics in the classrooms.[16] de Medeiros et al. (2008) found that teachers who reported high background noise levels in the classroom or outside school had twice as often voice symptoms than teachers who did not report these conditions. Other studies have shown that poor acoustics in the classrooms was associated with voice symptoms among teachers with odds ratios (OR) ranging from 1.80[17] to 2.69.[18] Poor ventilation (OR = 2.84)[19] and large changes in temperature (OR = 1.48)[17] were also associated with voice disorders among teachers. Other studies have focused on objective measurements of physical conditions (noise, reverberation time [RT], temperature) in the classrooms. These studies have concluded that noise and acoustic conditions are the primary uncomfortable factors in teachers΄ workplaces with background noise levels up to 87 dB[20212223] and RTs higher than 0.50 s.[202224] To the best of our knowledge, only one study has indicated that voice disorders were more prevalent among teachers who worked in schools with higher noise levels.[25] Hence, there is a lack of studies using objective measurements of work-related factors (such as background noise levels, RT, humidity, temperature), and the presented associations may suffer from information and recall bias.

Although objective measurements of physical conditions at the workplace are more expensive with high technical and personnel requirements compared with self-reports, it is recommended to complement self-reports with such measurements in order to describe comprehensively the working conditions. Self-reports are influenced by the level of satisfaction with physical conditions of the workplaces,[22] whereas objective measurements provide exposure levels as well as guidance to determine the technical requirements of the physical environment.[202426] Thus, there is a need for observational studies that investigate associations of objective measurements of physical conditions in the classroom with the presence of voice symptoms among teachers.

Therefore, we conducted a cross-sectional study within 682 Colombian school workers (621 teachers and 61 non-teachers) at 377 workplaces in 12 schools. The aims of the study were to assess the agreement between objective measurements and self-reports of physical conditions at the workplace and to evaluate their associations with the presence of voice symptoms among teachers.

Methods

Design and participants

This cross-sectional study was carried out in 12 public schools in Bogotá, Colombia (1449 teachers, and 143 non-teachers). The Department of Education of Bogota selected by convenience sampling the primary and secondary schools to participate in this study. The principal researcher had group and individual meetings with the head teachers of those schools in order to describe the purpose and requirements of the study and to invite them to participate. After, the principal researcher had group meetings with school workers to inform about the aims of the study and the voluntary and confidential nature of participation. Other characteristics of the sampling procedure have been described previously.[27] The study protocol was agreed by the Department of Education of Bogotá, the Universidad del Rosario in Bogota, and Erasmus University Medical Center in Rotterdam. The study was approved by the Medical Ethics Committee of the Universidad del Rosario in Colombia, and complied with the ethical principles embodied in the Declaration of Helsinki.

Data collection procedures

Data collection took place in February and March of 2012 (at the start of the school year). The questionnaire was designed to collected information on individual characteristics, voice functioning, lifestyle habits, work-related conditions, and health conditions possibly related to voice disorders. In addition, short-term objective measurements of physical conditions at the workplace were conducted.

Questionnaire

For this study, we developed a questionnaire that was based on existing English-language questionnaires described in the literature[141728] and consisted of 71 questions for teachers and 63 questions for non-teachers. Other characteristics of the design and characteristics of the questionnaire have been described previously.[27]

The first part of the questionnaire contained questions about sex, age, and education. The second part of the questionnaire contained questions about the presence of voice symptoms in the past month (tired voice, vocal fatigue, dry throat, itchy sensation and pain in throat, hoarseness, weak voice, voice spasms, voice loss, strained voice, breathiness); frequency (once, once every couple of weeks, weekly, daily), severity (mild, moderate, severe) and duration (open question) of these symptoms; whether the voice was affected while singing or speaking; aggravation of voice symptoms at work over time; and improvements during non-work periods in weekends and holidays. The presence or absence of voice symptoms was determined using the dichotomous question “Have you had voice symptoms in the past month?”[1428]

The next part of the questionnaire contained questions about working conditions,[29] including five questions about work-related factors, such as noise, acoustic conditions, temperature, humidity, and dust. For these physical factors, participants were asked to indicate whether they considered them uncomfortable: Always, often, sometimes or never. Since the frequency of answers for some categories was low, in further analyses physical factors were used as dichotomous variables, with subjects who answered “always” or “often” considered as being exposed. We also included questions about the presence of health conditions known to be associated with voice symptoms, such as respiratory diseases, gastrointestinal diseases; and hearing impairments.[173031]

Objective environmental measurements

We conducted objective measurements of sound levels (SLs) using the frequency weighting (A), temperature, humidity, and RT at the workplaces and SL outside schools. In the workplaces, SL, temperature and humidity were measured during actual teaching or working activities at three different locations to cover the complete workplace. The first position was defined as the most common location where the teacher (or worker) was located most of the time during the lecture (work time). The second position was close to the door, and the third position was close to the windows. The duration of each measurement was 1 min each.[32] The RT measurements were performed in non-occupied workplaces during weekends or non-lectures times. The measurements outside the schools were aimed at identifying the highest noise level at a distance of 2 meters from walls.[20] SLs (dB), temperature (°C) and humidity (% RH) were measured with the 4 in 1 digital multi-function Environment-Meter Mod WK040, which integrates the functions of SL meter, light meter, humidity meter, and temperature meter. Measurements of RT (RT60) were performed at 40 Hz, 50 Hz, 63 Hz, 80 Hz, 100 Hz, 125 Hz, 160 Hz, 200 Hz, 250 Hz, 315 Hz, 400 Hz, 500 Hz, 630 Hz, 800 Hz, 1000 Hz, 1250 Hz, 1600 Hz, 2000 Hz, 2500 Hz, 3150 Hz, 4000 Hz, 5000 Hz, 6300 Hz and 8000 Hz by the Room Acoustic Measurement System. In all statistical analyses, we used the average of the three measurements performed at the workplaces. In addition, for further analysis, we calculated the average value of RT of all 24 frequencies ranging from 40 Hz to 8000 Hz. All the objective measures of environmental factors were dichotomized, whereby subjects under the 75th quartile were used as a reference group.

Statistical analysis

Epi-info 3.5.3. (CDC/2011) software developed by Center for Disease Control and Prevention (CDC) in Atlanta (USA)[33] was used for data entry, and SPSS 20 software, one of the brands under IBM software Group΄s Business Analytics Portfolio, in New York (USA)[34] was used for statistical analysis. The statistical analysis was conducted on the study population with complete information on all variables. Since for some independent variables a few missing values occurred, multiple imputation was performed. Descriptive statistics was used for characteristics of the study population. The Shapiro-Wilk test was used to evaluate whether variables were normally distributed. Since menopause-related hormonal changes that may affect the voice of both men and women start around the age of 50 years,[35] we dichotomized the variable age using a cut-off value of 50 years of age. Because teachers could work in more than one classroom within a school, we calculated the average value of all environmental measures across all classrooms as the exposure measure per teacher. Since physical characteristics of workplaces may vary within and between schools, an analysis of variance was used to estimate the proportion of variance due to schools and workplaces within the schools. To assess the association between the objectively measured and self-reported work-related factors, we calculated the mean difference in objective measures between subjects with self-reported exposure to these factors and those subjects without. We used multiple logistic regression analysis to investigate associations between objectively measured and self-reported work-related factors with voice symptoms. Variables with a P value below 0.20 in the univariate analyses were included in the multivariate analysis in order to avoid residual confounding,[36] and were only retained when the P value reached the conventional level of significance of 0.05. In the final multivariate analysis, associations were adjusted for socio-demographic characteristics and health conditions. The magnitude of the association was expressed by the OR, and the statistical significance as the 95% confidence interval (95% CI).

Results

Participant characteristics

In total, 682 participants were enrolled in this study with the same response among teachers and non-teachers (43%). A non-response analysis showed no association between the proportion of participants and prevalence of voice symptoms among the participating schools. Compared with non-teachers, teachers were younger and more often women. No differences were observed in self-reported occurrence of health conditions or work-related factors between teachers and non-teachers. As shown in Table 1, teachers (71%) were more likely than non-teachers (54%) to report voice symptoms in the past month (OR = 2.03, 95% CI: 1.19-3.46). The four self-reported physical conditions were weakly correlated (Spearman's rho correlation coefficients between 0.17 and 0.20).

Table 1

Sociodemographic characteristics, voice symptoms, health conditions and work related factors of teachers and nonteachers in 12 public schools in Bogotá D.C., Colombia

Variable	Teachers (n = 621)	Nonteacher (n = 61)
	n (%)	n (%)
Prevalence of voice symptoms in past month	438 (71)	33 (54)
Sociodemographics
Female gender*	444 (71)	34 (56)
Age (years)*
<50	402 (65)	26 (43)
50+	219 (35)	35 (57)
Postgraduate studies	175 (28)	23 (38)
High school and Bachelor	225 (36)	12 (20)
Other levels of education	221 (36)	26 (43)
Self-reported health conditions
Respiratory diseases	228 (37)	27 (44)
Gastrointestinal diseases	305 (49)	28 (46)
Hearing impairment	186 (30)	14 (23)
Self-reported work-related factors
High noise in workplace	413 (67)	34 (56)
Poor acoustics in workplace	385 (62)	32 (52)
Dry air in workplace	300 (48)	33 (54)
Large changes in temperature in workplace	345 (56)	29 (48)

Chi-square test, P < 0.05

The study population worked at 377 workplaces (345 classrooms, 12 playgrounds, 19 offices, and 1 library) of which 338 workplaces (90%) could be visited for objective measurements. Since RT measurements were performed in non-occupied workplaces and availability was not always guaranteed, this factor was measured in 248 workplaces. On average, 31 workplaces were measured per school with little variation in a number of workplaces measured per school. Table 2 shows noise levels, relative humidity, temperature, and RTs in workplaces and noise levels outside schools. Differences in physical conditions were much larger between workplaces within schools than between schools. As shown in Table 3, for subjects who reported the presence of uncomfortable physical conditions the objective measurements at their workplaces showed similar mean values than for the workplaces of subjects without reporting exposure to these physical conditions.

Table 2

Physical characteristics of 338 workplaces in 12 public schools in Bogotá D.C., Colombia

Measure	Workplaces	Mean	SD	Interquartile range	Sources of variance
					Between schools	Between workplaces within schools
				25-75%	Percentage	Percentage
Background noise levels in workplace (dB(A))	338	72	7	68-76	12	82
RH in workplace (%)	338	52	7	47-57	48	52
Temperature in workplace (°C)	338	21	2	20-23	36	64
Reverberation time in workplace (s)	248	1.82	1.79	0.91-2.01	6	94
Noise outside school (dB(A))	12	74	7	69-80	100	Not applicable

SD = Standard deviation, RH = Relative humidity

Table 3

Relationships between self-report and objective measures of physical work-related factors in 12 public schools in Bogotá D.C., Colombia

Objectively measured work-related factors	Self-reported work-related factors		Difference
	Uncomfortable	Comfortable	Mean	SD
Noise outside school (dB(A))	73	72	1	12
Background noise in workplace (dB(A))	71	70	1	14
Reverberation time in workplace (seconds)	1.84	1.74	0.10	3
Humidity in workplace (RH)	51	50	1	15
Temperature in workplace (°C)	22	22	0	5

SD = Standard deviation, RH = Relative humidity

Work-related factors of voice symptoms

Table 4 describes that high noise outside the school (OR = 1.90) and poor acoustics in the workplace (OR = 2.44) were associated with the occurrence of voice symptoms, whereas no statistically significant associations were observed for other physical conditions. Participants who worked in schools with high noise levels in the surroundings reported voice symptoms more often than participants who worked in schools with noise levels below 80 dB(A) (OR = 1.88). The results of the multivariate analysis showed that the associations between high noise levels in the surroundings and self-reported poor acoustics in the classrooms remained statistically significant and changed little after adjustments for socio-demographic factors and health conditions.

Table 4

Associations between work-related factors and presence of voice symptoms in 12 public schools in Bogotá D.C., Colombia

Work-related factors	Multivariate analysis$
	Crude analysis		Objective measures		Self-reports		Full model
	OR	95% CI	OR	95% CI	OR	95% CI	OR	95% CI
Objectively measured work-related factors
High noise outside school (dB(A))	1.88*	1.19-2.95	1.91*	1.19-3.08			1.83*	1.12-3.00
High background noise in workplace (dB(A))	0.90	0.55-1.46
Long reverberation time in workplace (seconds)	1.03	0.69-1.54
High humidity in workplace (RH)	0.95	0.62-1.44
High temperature in workplace (°C)	0.91	0.63-1.32
Self-reported work-related factors
High noise in workplace	1.90*	1.35-2.67			1.33	0.91-1.95	1.34	0.92-1.95
Poor acoustics in workplace	2.44*	1.73-3.45			2.33*	1.60-3.39	2.34*	1.61-3.41
Dry air in workplace	1.35+	0.96-1.89			1.20	0.82-1.74
Large changes in temperature in workplace	1.37+	0.99-1.90			1.04	0.71-1.50

*P < 0.05, +P < 0.20, considered for inclusion in the multivariate logistic regression analysis, $Adjusted for sex, age, level of education, respiratory diseases, gastrointestinal diseases and hearing impairment. OR = Odds ratio, CI = Confidence interval, RH = Relative humidity

Discussion

In this study, we investigated the association between objectively measured and self-reported physical conditions at school with the presence of voice symptoms among teachers. Our findings showed that the noise in the surroundings of schools and self-reported poor acoustic conditions were important work-related factors of voice symptoms. We found poor agreement between objective measurements and self-reports on noise and acoustics. In conclusion, this study suggests that acoustics and noise may be important elements to take into account in the design of schools in order to reduce voice disorders among teachers.

The univariate analyses of potential risk factors for voice symptoms showed that objectively measured noise outside the school and self-reported high noise levels and poor acoustics in the workplaces were strongly associated with the presence of voice symptoms. The multivariate analysis showed that only objectively measured high noise levels in the surroundings of school and self-reported poor acoustics remained associated with voice symptoms. In the current study, it was not possible to disentangle completely the relative importance of acoustics and noise in schools since both factors were interrelated. However, the multivariate analysis suggests that voice symptoms were stronger associated with poor acoustics than with noise.

We suggest three reasons to explain the lack of association between objectively measured environmental aspects (noise, RT, temperature and humidity) with voice symptoms. First, the physical conditions were short-term measurements on a single day. The results may not necessarily reflect well the average conditions encountered by teachers during their school year. Second, average noise and RTs were high, which may indicate a lack of discriminatory power to compare teachers in good physical working conditions with those with high exposure. Third, individual sensitivity may vary whereby some teachers will experience already voice symptoms at much lower levels of exposure to environmental risk factors than others.

Although classroom acoustics guidelines have not been fully developed in most countries, national and international recommendations on acceptable exposure levels at the workplace have recommended that the noise level should not exceed 50 dB(A) and that the RT should be below 0.6 s for optimal student learning.[3738] In Colombia, the Department of Environment, Housing and Land Development recommends a Maximum LAeq in school zones during daytime of 65 dB(A),[39] and a Maximum LAeq of 55 dB(A) inside classrooms.[40] In this study we found, on average, background SLs around 72 dB(A), and 1.78 s of RT. The interquartile range of the distribution across workplaces shows that there were few workplaces with good acoustics, which may have limited our ability to demonstrate associations between these physical conditions and the presence of voice symptoms. The average values far exceed national and international recommendations, which will have important implications for the vocal health of teachers. Voice use in noisy or acoustically poor environments requires repetition and loud voice use without distinction of occupation. However, it seems likely that teachers may require loud voice use and repetition more often than non-teachers in order to maintain the attention of the students and to overcome poor acoustics and noise in classrooms. Permanent loud voice use under these conditions seems to contribute to increased loading of vocal organs and thereby, contributing to the higher proportion of voice symptoms among teachers.[18193141] Activity noises were measured during actual teaching or working activities at three different positions for 1 min each. However, these moments were scheduled by the head-teachers of each school. Each head-teacher informed us the day and hour that were available to perform the environmental measurements in each school. No specific conditions were determined to choose the measurement moments, except that the day of measurements (noise, temperature and humidity) was a regular day of academic activities.

In order to propose appropriate noise reduction and acoustic solutions strategies inside classrooms, it is important to identify the effect of particular noise sources on specific performance variables. We found that high noise levels in the surroundings of the school building were strongly associated with voice symptoms. Schools with higher outside noise levels were those located near to main streets, commercial areas, or to the airport. This finding is in concordance with previous studies that have shown that road traffic and aircraft noise interferes with speech and teaching inside classrooms since external high noise levels may influence internal noise levels.[1142] Therefore, lower RTs and lower noise levels are important elements to consider in the design of schools in order to reduce voice disorders among teachers. It is recommend to take into account the location of school buildings and sound insulation of the building when planning their construction, since it seems that external high noise levels have important effect on internal noise levels and thus on vocal health in teachers.

A major limitation of this study was the cross-sectional study design, which does not allow insight into the causality of the reported associations: We have no information on the relationship over time between the potential risk factors identified and the onset and perseverance of voice symptoms. Another limitation is the low response of the participants. However, the non-response has most likely not biased the prevalence of the voice symptoms because there was no association between high response and prevalence of voice symptoms among the participating schools. A third limitation was that random sampling of schools and teachers within schools was not feasible. Since selection of school and participants was not based on prior knowledge on the occurrence of voice symptoms or noise levels in the classroom, a systematic bias due to convenience sampling seems unlikely.

Conclusion

This study presented some indications that poor acoustics and high noise levels at the workplace may contribute to the occurrence of voice symptoms among teachers. However, these associations were based primarily on self-reports and could not be corroborated by objective measurements of physical conditions at the workplace.