Rev Saúde Pública

The effect of noise on the voice of preschool institution educators ABSTRACT OBJECTIVE: To analyze the association between noise levels present in preschool institutions and vocal disorders among educators. METHODS: Cross-sectional study conducted in 2009 with 28 teachers from three preschool institutions located in the city of São Paulo (Southeastern Brazil). Sound pressure levels were measured according to Brazilian Technical Standards Association, with the use of a sound level meter. The averages were classifi ed according to the levels of comfort, discomfort, and auditory damage proposed by the Pan American Health Organization. The educators underwent voice evaluation: self-assessment with visual analogue scale, auditory perceptual evaluation using the GRBAS scale, and acoustic analysis utilizing the Praat program. To analyze the association between noise and voice evaluation, descriptive statistics and the chi-square test were employed, with signifi cance of 10% due to sample size. RESULTS: The teachers' age ranged between 21 and 56 years. The noise average was 72.7 dB, considered as damage 2. The professionals' vocal self-assessment ranked an average of 5.1 on the scale, being considered as moderate alteration. In the auditory-perceptual assessment, 74% presented vocal alteration, especially hoarseness; of these, 52% were considered mild alterations. In the acoustic assessment the majority presented fundamental frequency below the expected level. Averages for jitter, shimmer and harmonic-noise ratio showed alterations. An association between the presence of noise between the harmonics and vocal disorders was observed. CONCLUSIONS: There is an association between presence of noise between the harmonics and vocal alteration, with high noise levels. Although most teachers presented mild voice alteration, the self-evaluation showed moderate alteration, probably due to the diffi culty in projection.


INTRODUCTION
The teacher's voice has been the focus of research that investigates, prevents or minimizes occupational risks, since this professional category presents higher frequency of vocal alterations than the population in general. 2,6,24 is necessary to implement actions that lead to the effective reduction in dysphonia within the teaching category, preferably during professional education. 2,4,16,24ese actions should consider behavioral and health aspects, work environment and organization, in addition to questions of relationship with students, families and the institutions' management.It is necessary to move from the preventive and individual focus towards health promotion targeted at the collective dimension. 10,16,23ograms that only provide information have a limited effect, since it is not the mere prescription of guidelines that will lead the teacher to change his/her vocal behavior. 3,9,15,19,20The union between informative or indirect approaches and more direct approaches seems to be the tendency of current studies in the area, as both complement one another. 7recent literature review 24 about high prevalence of vocal alteration among teachers argues that the main risk factors are: more advanced age and more years working as teachers, female sex, psychoemotional alterations, stress, high number of students per classroom and excessive background noise.The environment's high noise level, associated with the intense use of the voice, generates high vocal loading.It is considered an important public health problem also in leisure activities and on the streets.In schools, it interferes in the children's concentration and learning. 18The discomfort that is felt in its presence can be described as indisposition, irritability and stress and is not directly related to the level of exposure. 22Communication diffi culties have also been described, as well as headaches, sleep alterations, dizziness, hearing loss and tinnitus. 22Individual and collective preventive measures regarding noise exposure are indicated. 14Brazilian study with approximately 2,000 teachers found that half of them had complaints about high or unbearable noise levels at school, inside or outside the classroom and associated with worse voice-related quality of life. 10A literature review 13 analyzed studies in which the average noise level in educational institutions exceeded 70 dBA, which has also been observed in Brazilian day-care centers. 21Averages between 70 dBA and 85 dBA have been found in primary school classrooms in which the teachers presented many symptoms that were possibly associated with noise exposure, among them, dysphonia. 22a Constant exposure to sound pressure levels of 70 dBA or more is considered a permanent health risk situation, increasing wear of organism and risks of infarction, stroke and infections.b Teachers complain about the diffi culty in speaking loudly and being heard/understood in noisy environments. 17Speaking in the presence of high noise levels can lead to an increase in fundamental frequency and effort or vocal hyperfunction, mainly in non-trained speakers. 3In view of the fact that a large part of the teachers present restrictions in voice modulation and in vocal projection, 7 the presence of high noise levels at schools becomes even more relevant.
The improvement in the acoustic conditions of classrooms would be one of the main aspects to be tackled in an intervention program for teachers' adequate vocal use. 3 Furthermore, the program should approach vocal techniques and involve the use of individual amplifi cation devices.To achieve this, it is necessary to have better knowledge of the interrelation between voice and noise.
This study aimed to analyze the association between noise levels in preschool institutions and the presence of vocal alterations among educators.

METHODS
Cross-sectional observational study carried out at three preschool institutions in the city of São Paulo (Southeastern Brazil) in 2009.The institutions had established agreements with the Municipal Government, were managed by the same social institution and were located in the west zone of the city.They were selected by the criterion of convenience due to a partnership between the speech pathologists and the supporting institution for the development of health promotion actions.The 28 educators were invited to participate.Those who were absent or ill on the recording day would be excluded, but there were no losses.
The sound pressure levels were measured in the spaces of daily routine activities: classrooms, refectories, corridors, outdoor recreation areas, service areas and administrative rooms, in the presence and absence of the children.
A digital sound level meter of the brand Center, model 322, with Data Logger was used, and the norms of Associação Brasileira de Normas Técnicas (ABNT -Brazilian Technical Standards Association) c were followed.The device was calibrated and was operated in slow response curve in a sound level range between 30 dBA and 130 dB.The measurements were performed in the absence of atypical sound sources like rain or strong wind.Different points were considered in each measurement place, with a minimum time of fi ve minutes per point, and measurement device in the fast operation mode.The measurement was performed in a central point in small spaces (up to 10 m 2 ), two diagonal points were considered in medium spaces (larger than 10 m 2 and up to 20 m 2 ) and three points in large spaces (larger than 20 m 2 ).The device was positioned heading to the center of the spaces at a distance of, at least, one meter above the fl oor and away from the walls or outside walls.
The minimum, medium and maximum values of the obtained sound pressure levels were considered and the means were classifi ed according to what the World Health Organization recommends for comfort, discomfort and damage.b The results were dichotomized into adequate (comfort)/altered (discomfort or damage).
The educators underwent voice evaluation: self-assessment, auditory-perceptual evaluation and acoustic evaluation of the voice.
A visual analogue scale was used with 10 cm of length for the vocal self-assessment, in which 0 (zero) represented absence of vocal alteration and 10, maximum vocal alteration.The educators should mark a vertical line representing the status of their voice in the last days.This line was subsequently measured with a ruler and the value was classifi ed according to cut-off points per tertiles: no alteration (zero), mild alteration (values between 0.1 and 3.4), moderate alteration (values between 3.5 and 6.7) and extreme alteration (values between 6.8 and 10).
Voice samples of each educator were individually recorded: sustained emission of the vowel /a/ and counting numbers from 1 to 10.This recording took place at the institutions, in distant and silent rooms, with maximum noise level of 50 dB.The voices were recorded directly in a desktop computer and the acoustic analysis program Praat was used, together with a microphone of the make AKG.
The auditory-perceptual evaluation of the educators' vocal quality was performed using the GRBAS scale. 8lthough this scale analyzes vocal quality only at the glottal level and is subjective, it was selected because it is utilized in studies across the world and because its application is simple and quick, as the educators would not be able to be absent from the classroom during a long period of time.The voices were analyzed by a speech pathologist who has been working in the area for more than ten years.
The educators were divided into three groups: adequate voice (AVG): G (general degree) = 0; altered voice in a mild degree (MDG): G = 1; moderate vocal alteration (MODG): G = 2; and extreme vocal alteration (EDG): G = 3.The parameter G was chosen because it represents the global impression that we have of the voices.The other parameters -roughness, breathiness, asthenia and strain -were used in the correlation with the other data.The acoustic analysis was performed with the program Praat, using the medial portion of the sustained emission, because it is more stable.The automatic measures of interest were extracted: fundamental frequency, jitter (perturbation measure related to the frequency of the voice), shimmer (perturbation measure related to the intensity of the voice), and harmonics-tonoise ratio.A qualitative analysis of the spectrographic tracing of the same emission was carried out using the program Spectrogram, version 16.0, which was chosen due to the quality of the defi nition of the spectrogram.
Aspects related to orofacial motricity, resonance and speech were observed and registered during the contact with the educators to record their voices.
The AVG, MDG, MODG and EDG groups were compared in relation to the minimum noise levels that were observed (adequate, altered) and to the other fi ndings: vocal alteration degrees indicated in the educators' vocal self-assessment (adequate, mild, moderate or extreme); automatic acoustic measures and spectrographic analysis (adequate, altered); observed orofacial motricity, resonance and speech (adequate or altered).
The statistical program SPSS was utilized and the fi ndings were analyzed by descriptive statistics and application of chi-square test.A level of signifi cance of 10% was considered due to sample size.
The research was approved by the Ethics Committee of the institution (006/11).The institutions' principals and educators were informed about the project and signed a consent document.

RESULTS
The educators' age ranged between 21 and 56 years (mean of 30 years).
Approximately 25% of the educators presented adequate voice and 75%, altered voice in the auditoryperceptual assessment of the voice by means of the GRBAS scale, considering the global degree of alteration (G).Among the altered voices, 67% were classifi ed as mild alteration and 33% as moderate alteration.None were classifi ed as extreme vocal alteration; therefore, the EDG group was not constituted.
The mean noise level in the different spaces varied between 58.1 dB (discomfort 2, i.e., altered) and 83.7 their daily work and are considered inherent in the profession. 10,16Rough-breathy voice was found in half of the educators, which may indicate presence of vocal fold nodules. 1,19Laryngeal assessment was not performed due to the professionals' diffi culty in leaving the classroom.Roughness, breathiness and hyperfunction/strain have been prevalent in other studies, 13,19 but only one 19 found alterations related to phonotrauma by means of laryngeal assessment.
No association was observed between voice alteration in the speech pathologist assessment and the educators' dB (considered damage 2, also altered).The general mean was 70.4 dB (damage 2).The maximum levels ranged between 58.7 dB and 100.5 dB, with mean of 82.7 dB; and the minimum levels between 37.8 dB and 70.6 dB, mean of 56.1 dB.
The educators' vocal self-assessment ranked an average of 5.1 on the visual analogue scale, being classifi ed as moderate alteration.
Approximately half of the educators presented f0 values out of the expectation in the acoustic assessment, and the major part presented lower frequencies.The jitter values were altered for 32%, the shimmer ones for 43% and the harmonics-to-noise ratio, for 64%.
The majority of the aspects of the spectrographic tracing were altered with noise between the harmonics, irregularity and interrupted tracing, as well as alterations in the defi nition of the harmonics, reduced number of harmonics and low frequency concerning the defi nition limit.In addition, 57% of the educators presented alterations in orofacial motricity, resonance or speech, in isolation or in combination.Speech alterations were observed in 36%, resonance alterations in 21% and orofacial motricity alterations, in 18%.The following were observed with highest frequency: locked joint (24%), hyponasal resonance (19%), mouth breathing (19%), open bite or crossbite (14%) and frontal lisp (14%).
Educators with mild or moderate vocal alteration were exposed to levels that were considered inadequate with higher frequency than the educators from the group without alteration (Table 1).
As for the vocal self-assessment, orofacial motricity/ resonance/speech and acoustic measures data, there was no difference among the groups, except for the jitter values, which were considered to be altered with higher frequency in the groups with vocal alteration (Table 1).
Altered tracing and substitution of harmonics for noise were observed with higher frequency in the groups of educators with vocal alteration (Table 2).

DISCUSSION
The vocal alterations were more prevalent in institutions with higher noise level.High occurrence of alteration was observed in the vocal assessment performed with the GRBAS scale, in agreement with the literature review, 24 although the methodological procedures have been varied.
Mild alterations predominated and no case of extreme alteration was found.Only extreme vocal alterations lead to absences or sick leaves, as, in the teachers' perception, mild or moderate dysphonias do not limit self-assessment.Educators with adequate voice in the speech pathology assessment evaluated themselves as altered, the majority with mild alteration.Those with mild alteration in the speech and audiology assessment considered themselves mainly with extreme vocal alteration and two evaluated themselves as normal.Educators with moderate alteration were distributed among selfreported mild, moderate or extreme alteration.Teachers usually have diffi culty in perceiving their voices, 16,19 but self-perception can be an important tool for the early detection of vocal problems and adherence to treatment, when indicated.The educators do not refer only to vocal quality in the self-assessment, but also to aspects like diffi culty to speak loudly or in the presence of noise, 17,19 which was reported by participants in this study.
The spectrographic analysis showed the professionals' diffi culty to speak in the presence of noise: the analyzed aspects were altered in educators with adequate and altered voice, like harmonics defi nition, harmonics interruptions, noise between harmonics and graphic irregularity.These alterations are associated with little vocal brightness and projection, i.e., diffi culties to speak in strong intensity, to make oneself be heard in large spaces and in the presence of competitive noise.Little projection and vocal modulation was observed in another study with teachers. 17The richer the harmonics series, the better the vocal quality and the glottal closure, 1 conditions that facilitate vocal projection.
Altered spectrographic tracing and substitution of harmonics for noise were associated with vocal alteration.This indicates that global alterations of the spectrogram enabled to differentiate adequate voices from altered voices.Besides, the substitution of harmonics for noise is related to vocal fold nodules and glottal gap, 1 whose main auditory correlate would be roughness or roughness with breathiness.The altered vocal quality that was the most found in this study was the rough-breathy one.
In this study, f0 was altered for more than half of the participants, who presented mainly f0 lower in the analysis of the acoustic measures of automatic extraction.This decrease can indicate difficulties in vocal extension or vocal fatigue in educators without alterations in the GRBAS scale; altered voice can be related to the presence of mass lesions in the vocal folds, 1 which should be investigated in future studies.The vocal folds may become overloaded if the vocal extension is reduced.
The harmonics-to-noise ratio was altered for more than half of the educators and did not differentiate between altered and adequate voices, indicating that vocal extension limitations may generate fatigue, whose signs appear on the spectrum, but are not perceived in auditory terms.The shimmer values were adequate to the majority of the participants, like the jitter values, but the latter was able to differentiate the voices, reiterating that it correlates more directly with voice alterations than shimmer. 1 Thus, the educators with altered voice presented more diffi culty in controlling the vibration of the vocal folds than alterations in glottal resistance. 1 Jitter is a measure of aperiodicity that loses reliability as the aperiodicity of the voice increases. 25There were no voices with extreme alterations in this research.
Speech, resonance and orofacial motricity were altered, and the main alterations were locked joint or imprecise articulation, hyponasal resonance and mouth breathing, as well as bite alterations and presence of frontal lisp.These data are directly related to the vocal projection diffi culties that were reported and observed in the spectrographic analysis, as well as to an intense overload of the vocal tract.
Imprecise articulation indicates diffi culties to control the phono-articulatory dynamics and can favor hypertonic laryngeal patterns as an attempt to compensate for speech intelligibility problems. 1It is also associated with lack of coordination between breathing and speaking, an aspect that needs to be intact for an adequate professional use of the voice.Impaired intelligibility may interfere in the teaching-learning process, mainly if it occurs in sound competition situations.Articulatory imprecision is associated with the psychodynamics of lack of clear ideas and unwillingness to communicate. 1 Altered resonance negatively affects the amplifi cation of the harmonics and indicates a restriction for the modifi cation of adjustments in the vocal tract, 1 which will be made inadequately and with effort.The orofacial motricity alterations and breathing specially through the mouth can lead to phonation with compensatory effort, causing laryngeal hyperfunction. 1 Furthermore, respiratory alterations negatively affect sleep quality and attention capacity.The lack of adequate rest can reduce the educators' concentration levels and hinder the adequate performance of their work. 6ese speech, resonance and orofacial motricity alterations were observed in educators with adequate voice and in those with altered voice.To the latter, it may be one more overload, and there is the need of speech pathology, ENT and dental assessment and treatment.
After the recordings, the educators received a feedback on the most relevant aspects that were observed.The spectrogram obtained from their emission was utilized, as the use of images of the acoustic analysis enables a better understanding of the vocal problem. 25The educators agreed with the observations, mainly those concerning the diffi culties in vocal projection and resonance alterations.
High noise levels were observed in the three institutions, like in other studies; 3,11,13,21 therefore, they represent risk environments for the professionals' and children's health in many aspects, which need to be modifi ed.High noise levels interfere in communication, and they also hinder attention and concentration, interfere in memory and contribute to the occurrence of stress and excessive fatigue.In addition, they expose the worker to factors that may trigger occupational accidents. 5Many teachers are able to perceive the noise and associate it with vocal alterations, with impacts even on voice-related quality of life. 10e means of the minimum values found were used to analyze the noise in relation to the educators' voice.This is the only way in which there would be a group exposed to noise levels within what is considered adequate.An association between mildly or moderately altered voice and exposure to noise levels that are considered inadequate was observed, as is shown in the literature, which indicates that the changes in the activities of the muscles related to vocal production, in an attempt to overcome the noise, will possibly lead to vocal quality deterioration. 3,4,9,10,15,16,23,24Another study has found an association between high noise levels and vocal alteration in teachers of one out of three analyzed schools. 13e adhesion of all the invited educators shows interest and need of research in the voice area in this professional category.The lack of studies on the theme points to the need of research that aims at environmental characterization and the problems it may cause to professionals and students, and also at evaluating multidisciplinary actions of health promotion in the search of a better quality of life to those who are involved.
The way in which each individual adjusts his/her voice to speak in the presence of noise can be peculiar, 12 which indicates the need of more studies.
Our data strengthen the need of multisector measures to reduce noise levels and to promote the adequate use of the voice at the workplace, aiming to improve the quality of life of these professionals and of the children.It is fundamental that measures that have a collective focus are developed; such measures should emphasize the improvement in the acoustic comfort, the awareness-raising of the educators in relation to its alterations and to the importance of the treatment, when indicated.Likewise, the need of complementary laryngeal assessment will be discussed with the institutions' managers so that it is possible to think of a more specifi c instrumentalization of these educators to the professional use of the voice.

Table 1 .
Comparison among the groups AVG, MDG and MODG in relation to noise levels, vocal self-assessment and acoustic measures.São Paulo, SP, 2009.

Table 2 .
Comparison between the groups AVG, MDG and MODG in relation to the spectrographic analysis.São Paulo, SP. 2009.