Prevention of occupational dysphonia in scientific university personnel: A cross-sectional study

University teachers are professional speakers and belong to the risk group of developing a voice disorder as a result of their job (Tormin and Bock 2018). They are a particularly vulnerable group due to their above average high level of speaking in their everyday work and the demands placed on their voice quality. At the end of 2018, around 400,100 people were employed as scientific staff at German universities and university clinics. This included around 47,900 professors who taught and researched at German universities and universities of applied sciences (Statistisches Bundesamt 2018). Many studies found evidence of an increased prevalence of voice disorders due to excessive speech strain. While the frequency of voice disorders in the general population is around 6%, professional speakers suffer from it at two to three times the frequency (Fritzell 1996; Sliwinska-Kowalska et al. 2006; De Jong et al. 2006). In the study by Higgins (2006), 45% of the university teachers surveyed stated that they had a voice disorder. Of these dysphonic speakers, 6.7% were chronic, that is, showing a dysphonia that lasts for more than 4 weeks, and 93.3% were acute (Higgins 2006). A systematic review and meta-analysis by Azari et al. (2022) reported an overall prevalence of voice disorders in university professors of 41%.

Despite the overall high prevalence, only a few risk factors have been identified that have a significant influence (Korn et al. 2018). One influencing variable is gender: here, a gender difference to the disadvantage of female lecturers shows that females are significantly more frequently affected by voice complaints than their male colleagues (Korn et al. 2018; Gomes et al. 2018; Kyriakou et al. 2018). Perceptions of voice problems and stress are also cited as influencing factors (Korn et al. 2018). Female university faculty members are more likely to perceive their vocal problems when teaching and perceive their working conditions as more stressful (Korn et al. 2018). According to observations of Korn et al. (2015), female university teachers who reported being stressed or nervous were also more likely to experience vocal problems. Upper respiratory tract infections were found to be significantly more common among university teachers (Higgins 2006; Korn et al. 2015), and a relationship between hoarseness in teaching staff and the work environment in terms of loudness, noise levels and poor air quality has been found (Korn et al. 2015). The louder the area we are in, the louder we try to make our voices so that we can still be heard. This noise can be caused by human voices, such as by seminar participants, but also from the interaction of functional noises of projectors, air conditioning systems and other devices (Nollmeyer 2012). Lecturers who have not learned any techniques to generate the necessary vocal force often speak with too much force and develop hyperfunctional dysphonia and, as a result, a hyperfunctional speaking style (Rittich 2018). In this way, the vocal musculature can be damaged (Rittich 2018). There are rooms that have unfavourable hall properties for speakers (Rittich 2018). These are, for example, rooms with straight, smooth walls, large windows and non-textile floor coverings. The reverberation of what is said can mix with the next words, which impairs intelligibility. Speakers have to counteract this by increasing the volume of their speech, usually with additional effort (Rittich 2018). But also the opposite, an acoustic environment that is too anechoic, can be disadvantageous for speakers, because every word in the room is immediately “swallowed” (Rittich 2018). This makes permanent adjustments for the voice and articulation more difficult, which can lead to non-physiological vocal pressure. Similarly, excessively dry air conditions, such as those we experience in many heated seminar rooms, can negatively affect our pharyngolaryngeal mucous membranes, including vocal fold epithelia. They become dry—an extremely unfavourable situation for intensive use of the voice. Drinking water frequently can be a simple remedy to this. Furthermore, a correlation was found between vocal problems and one’s own vocal abilities (Richter and Sandel 2013). If these are low, more vocal problems result. Any healthy voice is capable of being subjected, at times, to unusually intense work for more than 6 to 8 hours a day, or for several days at a time. If performance decreases or even hoarseness sets in, then this is usually a warning sign but can usually be remedied with a few hours or even a few days of vocal rest (Richter and Sandel 2013). If vocal stress becomes permanent for a longer period of time, say, almost every day at work, the voice, which is still in need of rest, is challenged again and again. This leads to the failure of compensation mechanisms, the overload becomes chronic, and functional dysphonia develops (Seidner 2013). This effect is exacerbated if the speech performance is permanently monotonous. Another significant relationship was found between voice problems and high speech rate, high pitch and sustained loudness (Gomes et al. 2018). During intense vocal use, laryngeal and pharyngeal muscles are placed under strong, unaccustomed tension. This, combined with high mental presence and possibly stage fright, results in increased whole-body tone. In the speech process, in which already renewed phonation takes place during the exhalation phase, a breath break is hardly possible. This would be important, however, so that the respiratory muscles are released and the diaphragm in particular has a chance to regenerate briefly. If the rhythmic breathing process is not given, however, this can lead to a muscular imbalance of the respiratory musculature, which in turn triggers unfavourable compensation mechanisms in the vocal musculature. It is supposedly assumed that various job demands (e.g. inattentive students) require a higher or even lower position than one’s primary voice pitch because it is believed a pitch change will gain respect (Tormin and Bock 2018). However, if one does not speak in one’s natural physiological voice pitch in the long run, one risks unfavourable tension conditions in the laryngeal musculature. Prolonged loudness was also a significant influencing variable in the context of a large student population (Gomes et al. 2018). If there is a need for a dynamic increase in intensity to call voice in untrained personnel, especially those with constitutionally unfavourable preconditions, it may lead to compensatory overtonation with a gradual increase of the mean pitch of the speaking voice and a corresponding subjective symptomatology of overload (Rittich 2018).

When evaluating the influencing variables, a differentiation must be made between, on the one hand, the factors that affect action—such as vocalisation—and, on the other hand, the factors that would call for an adjustment of environmental factors or conditions. For example, the risk factor duration of vocal stress could be addressed both behaviourally by incorporating more individual work into the seminar concept of lecturers, and relatively by offering further training in didactics. Further behavioural preventive measures would be to help those affected to find their own main range of speech and to adapt it adequately to professional requirements (Rittich 2018), and also to avoid a non-physiological posture by first informing lecturers about the connections between posture and voice and then finding ways to improve posture in the relevant speaking situations.

Studies show that university teachers are at risk of developing work-related dysphonia over the course of their professional lives as a result of increased vocal demands (Vilkman 2004; Kooijman et al. 2006). This not only has unfavourable health consequences (voice problems, poorer well-being, more stress) for those affected (Smith et al. 1996) but also negative effects on the learning success of students (due to a loss of quality in teaching) (Rogerson and Dodd 2005) and, in the long run, also on the economy (more sick days, costs for treatment and early retirement) (Verdolini and Ramig 2001). In the worst case, chronic occupational dysphonia can develop, which has not yet been recognised as an occupational disease in Germany (Hammer and Teufel-Dietrich 2017; Kutej 2011). Although the voice is essential for working at a university, surprisingly little is done to preserve it. There is no training or appropriate preparation with regard to the increased vocal stresses in this profession (Boss-Ostendorf and Senft 2018). A congenital voice weakness is unchangeable, but the correct speaking and breathing technique can be learned (Tormin and Bock 2018). This is where an existing discrepancy between professional requirements and insufficient preparation for university teachers in the relevant training and study programmes comes into play (Tormin and Bock 2018). Furthermore, no vocal fitness examination takes place before entering the profession. Hammer and Teufel-Dietrich (2017) emphasise that a vocal fitness examination can reveal a need for prevention. One often searches in vain for offers in the field of prevention and early detection. Appropriate structural and personnel framework conditions should therefore be created in the university setting.

Prevention would be all the more successful the more it is oriented towards the importance of the professional and personal context. Since the ICF model focuses on strengthening individual resources in everyday life, it therefore applies to the conceptual work of the voice prevention programmes in question (Rittich 2018) (see Fig. 1).

By means of its interrelated components of structure/function, activity, participation and contextual factors, the model takes into account the multidimensional causal structure of occupational dysphonia. In particular, it makes sense to consider the participation component, since the distinguishing feature of the group of professional speakers from other potentially voice-affected people is part of everyday professional life, that is, participation (Ehlert 2011). The holistic perspective on the effects and conditions of voice complaints in university teachers seems particularly suitable for the implementation of preventive intervention measures.

The aim of the study was to identify the group of scientific staff at German universities as suitable recipients for health-promoting offers in the area of voice. It sought to show why this target group, in particular compared to the non-scientific administrative staff also working at universities, would be a preferred candidate for such offers. For this purpose, the effects of voice problems on physical, psychological and functional aspects of the life of university teachers were considered. Subsequently, influencing factors that promote or prevent the development of voice disorders in university teachers were identified.

Finally, possible implications for operational practice at universities have been derived from these findings and measures for prevention and early detection formulated in order to avoid voice disorders in university teachers. In order to examine whether university teachers are more frequently affected by voice problems than university administration employees, whether job-related and individual factors influence the occurrence of voice problems among university teachers and whether there is a connection between voice problems and physical (e.g. more sick days), psychological (more stress, worse well-being/quality of life) or functional (reduction of didactic activities) changes, the following hypotheses were motivated:

The study population included all university faculty and administrators.

The sample was a total of 101 people who took part in the questionnaire study in the period from 14 June until 14 July of 2021. Of these, eight participants had to be excluded because they did not meet the inclusion requirement of “belonging to the teaching or administrative staff of a university” in Germany.

The test data set included the data of 93 participants, which were evaluated with the statistical programme R (https://​www.​r-project.​org/​) with the data science package collection tidyverse (Wickham et al. 2018). Fifty-nine participants assigned themselves to female gender and 22 to male gender. Twelve participants did not indicate their gender. The average age of the participants was 49.5 years, ranging from 27 to 65 years. The average age of the male participants was 47 years, and 44.1 years for the female participants; see Fig. 2.

In order to answer the questions and to examine the hypotheses, an empirical-quantitative approach, in the form of a cross-sectional study, was chosen using a questionnaire design. For this purpose, a test group (teaching staff) and a control group (administrative staff) were recruited. The allocation to the groups was made by an initial corresponding question in the questionnaire; see Appendix A. All participants then received the same questions, with the group of teaching staff including additional questions about the teaching situation, previous investigations, training courses and measures or about wishes for the employer. The groups were compared for the presence of voice problems. The groups were then examined to determine whether there is a connection between workplace-related conditions and/or individual factors and the presence of voice problems. The groups were analysed according to whether there is a connection between the severity of symptoms of voice problems and loss of vocal production and the number of sick days.

In order to reach as many potential respondents as possible, a questionnaire in online format was chosen. The online questionnaire was implemented using SoSci Survey (Leiner 2019) and made available to the participants on www.​soscisurvey.​de after technical and content-related pre-testing.

The theoretical foundation of the questionnaire was based on the application of the International Classification of Functioning, Disability and Health (ICF) model (see introduction).

The questionnaire began with the welcome page; see Appendix A. The participants were informed of the use and the storage of their data. By clicking the continue button, the participants declared their consent. Page 2 asked for socio-demographic information (gender, age, professional group, function and professional experience at the university). The third and fourth pages contained the questions that were aimed exclusively at the teaching staff. These were job-related questions (page 3) such as the size of the student groups, the weekly teaching load, the type and duration of the teaching, the use of technical aids and questions about previous training and studies, whether a voice aptitude test took place before the start of the teaching activity (page 4), whether a voice seminar was attended as part of the university work and whether first-aid measures for the voice are known. The last question asked about wishes to the employer for preserving a good voice. Pages five to ten were aimed at all participants and asked about the voice profile, voice activities with singing or modulation and activity and participation using five-stage multi-category answer formats in the area of voice function. For this, items from existing self-assessment questionnaires were used and supplemented by the specialist literature. This applied to items 14 to 35 (excluding item 17) from the Vocal Profile for Professional Speakers (SPBS) (Ehlert 2011), the studies by Besser et al. (2020), Assuncau et al. (2012), Depolli et al. (2019) and de Medeiros et al. (2012). Item 17 was taken from the Voice Symptom Scale (Deary et al. 2003). The questions on environmental factors on pages 10–11 in multi-category response format (items 36-40) were taken from studies on external risk factors for dysphonia (Vilkman 2004; Kooijman et al. 2006). The items relating to personal factors (items 41 to 48, pages 12 and 13) were taken from the studies by de Almeida et al. (2010) and Alva et al. (2017). Except for question 46, which was asked in mixed response format, all items were created in multi-category response format.

To recruit participants, the author of the study wrote an e-mail to all members of universities in Germany who were in her professional and private network and who worked in the areas of teaching and administration. In addition, the freely accessible e-mail addresses of employees of German universities who met the inclusion criterion of belonging to university staff in the areas of administration or teaching were selected from the corresponding university websites. In addition to the link that led to the questionnaire, the invitation e-mail contained a brief description of the author and the university for which the survey is being carried out. In addition, the questionnaire was briefly presented, and it was explained how the data obtained will be used further. Colleagues proved to be helpful supporters by distributing the invitation to the questionnaire and study via internal university mailing lists and promoting participation in the study (snowball principle).

There was no reward for the participants. The data was entered by the participants according to the input fields in the data mask of the online questionnaire. Data protection was applied in that the data has only been stored in an anonymous form and then processed further. If the purpose of storage is finally no longer applicable, all data will be deleted. The questionnaire could be terminated at any time without giving a reason. By clicking the continue button on the first page of the questionnaire, the participants confirmed their consent to the storage and use of their data as detailed previously on the questionnaire.

Table 1 shows the results of the survey of the teaching staff on their conditions at work, previous examinations/training courses in the field of voice, and their wishes.

The university teachers surveyed teach predominantly (64.4%) small student groups (21 to 40 people) or very small (42.5%) student groups (1 to 20 people). The majority of the university teachers taught three times a week (38.6%) with generally (39.7%) more than four teaching units per day. At the time of the survey (under pandemic conditions caused by the infectious disease COVID-19), they mostly taught virtually (71.2%) and/or by being present (60.3%), with mostly no (41.1%) to rare (19.2%) use of technical aids such as microphones, audio systems or headsets. Only four (5.5%) of the participants stated that they had undergone a vocal aptitude test with an ENT (ear, nose and throat) doctor/otorhinolaryngologist)/phoniatrist before they began teaching, and 87.7% denied this to be the case. A clear majority (79.5%) stated that they had not completed any further education or training in the field of voice. In addition, 48 (65.8%) of the university teachers surveyed did not know any first-aid measures for the voice. It is therefore not surprising that many concrete wishes were formulated for the university as an employer, which primarily expressed training offers in the areas of voice, relaxation and coaching. In addition, there were requests for medical advice from the company doctor, for optimisation of the technical and acoustic equipment or technical support in all teaching rooms, and for a better distribution of teaching with longer breaks and the avoidance of daily teaching units that exceed four units.

All participants were asked to rate their voice themselves. Table 2 shows the results of the self-declared voice profile.

Overall, only minor vocal anomalies could be identified. However, the teaching staff, in contrast to the administrative staff, more frequently had problems with the parameters clearing their throat and coughing. A total of 31.5% of the teaching staff reported difficulties sometimes and 39.7% rarely. For administrative staff, 25% reported problems sometimes and 50% rarely. The parameter voice rough, hoarse, raspy was mentioned more frequently (31.5% sometimes and 45.2% rarely amongst teaching staff; and 10% sometimes and 55% rarely for administrative staff). The teaching staff also more frequently (26% sometimes and 39.7% rarely) stated a change in the sound of the voice over the course of the day than the administrative staff (35% never and 35% rarely). The parameters from the voice profile clearing of the throat/cough and change in voice tone over the course of the day were related to age. There was a connection of the effect that clearing the throat, coughing and a change in the tone of the voice over the course of the day among university teaching staff with the increasing of age.

Figures 3 and 4 illustrate that there is a connection to the effect that clearing the throat, coughing and a change in the tone of the voice over the course of the day among university teaching staff increases with age.

The questionnaire also recorded information on potentially influencing environmental factors at the university; see Table 5. The university teachers reported background noise more frequently than the participating administrative staff (32.9% sometimes and 34.2% rarely vs 10% sometimes and 45% rarely), likewise a dry indoor climate (37% sometimes and 28.8% rarely vs 20% sometimes and 50% rarely), similarly poor spatial acoustics (31.5% sometimes vs 15% sometimes), as well as missing breaks (32.9% sometimes vs 10% sometimes).

Finally, individual data of the participants were collected that may have an impact on the voice; see Table 6.

Here, 20 (27.4%) university teachers stated that they only sometimes or rarely got enough sleep compared to eight (40%) employees in administration. The remaining parameters were mostly classified as not conspicuous.

To check the hypotheses, ordinal regression analyses were carried out using the proportional odds (PO) model with the R data science package collection VGAM (version 1.1-5) (Yee 2015).

The first hypothesis was confirmed by the results of the items clearing/coughing and voice rough, hoarse, raspy. According to this, in the sample examined, university teachers are affected by voice problems more often than administrative employees in their professional environment. Based on proven differences in the two aforementioned items, the hypothesis of the equality of conditions can be refuted. The results thus confirmed the literature (e.g. Higgins 2006) which found an increased prevalence of voice disorders in university teachers due to excessive speaking load. However, the absence of a significant difference in the item change in voice over the course of the day must not be interpreted as evidence that the two groups have no differences. The non-existence of differences on this item could not be proven beyond doubt.

The second hypothesis could be confirmed by the parameters dry room climate, poor room acoustics, background noise and lack of breaks based on the available data with the available test power. Unfavourable conditions in the workplace, such as a dry room climate, poor spatial acoustics, background noise and the lack of breaks, are more prevalent in the group of university teachers that are vocally impaired compared to administrative employees. The results pointed in the same direction as the findings by Korn et al. (2015) who reported an association between hoarseness among teaching staff and their work environment in terms of volume, sound level and poor air quality. The results relating to the connection between voice problems and poor spatial acoustics were also confirmed in the literature by Rittich (2018). Previously, the findings by Gomes et al. (2018) have shown that voice overload can have adverse consequences. The results of the examined sample pointed in a similar direction, since university teachers more than randomly mentioned missing breaks as the cause of their voice problems.

The third hypothesis could be verified in the examined sample with the existing test power for the parameter age. Age is a sensitive factor that favoured the development of a voice disorder. Age-related, physiological changes in the voice apparatus can no longer be adequately compensated for by additional factors. This is confirmed by previous work on age-related changes and their consequences for members of speaking professions (Tormin and Bock, 2018). One of the additional factors could be the stress factor, which was not explicitly asked about in this survey. Findings by Korn et al. (2015) have confirmed a connection between stress, nervousness and vocal problems in university teachers.

The fourth hypothesis agreed with the study data for the items cough/clearing throat, hoarse, hoarse, raspy voice and long vocal onsets. This was also confirmed by the results of Rittich (2018), who found a compensatory over-toning with a gradual increase in the middle speaking voice range and associated subjective symptoms of overload in the direction of vocal fatigue, especially in untrained personnel. The majority of the sample examined was voice untrained.

This study had several limitations that constrain its generalisation. First of all, it was based only on an online questionnaire. Due to the pandemic conditions which prevailed at the time the study was undertaken, no clinical evaluation was performed with the subjects who participated in the study.

Second was the small random sample size of the group of administrative staff, which comprised only 20 participants. According to Ioannidis and Lau (1997), studies with a small number of subjects only provide unambiguous answers to research questions if the study is well defined and the population is homogeneous. A possible solution would be a collection of various small groups that are as homogeneous as possible. That offers the opportunity to test different groups and differently defined studies in order to create tailor-made offers, such as in training/consulting. For this, the recruitment channels and the wording of the invitation text would have to have been reconsidered. Presumably, members of the administrative staff per se participate less in scientific studies and therefore need different information or incentives to motivate them to participate.

Third, the group of university staff stated that they mostly taught with small or very small groups of students. This can lead to a selection bias due to a lower speaking load, as could be the case in larger groups of students. As a result, this could systematically have distorted the results with regard to the information on the vocal profile, vocal activity, especially with singing, modulation and participation. University teachers with larger groups of students should therefore be enabled to be included in a targeted manner in departments with more heavily frequented courses.

In addition, there was another form of distortion: the loss bias due to omissions of answers. This happened up to 26 times in several questions. This could easily be counteracted by designing the online questionnaire in such a way that participants are reminded of unanswered questions, or filling it out is mandatory in order to be able to continue processing the questionnaire.

There may also have been a measurement bias due to systematic differences in the way the result parameters were collected. So, for example, only two questions were asked related to stress (whole-body exertion and jittery, tight voice). Detailed questions about the current stress level, general well-being, psychological stress, self-regulation strategies and self-efficacy would have been beneficial here, as there is an evident connection between psychological condition, emotional state and voice. Numerous studies show a relationship between functional dysphonia and psychological characteristics or disorders such as anxiety, stress and depression. For example, almost 30% of patients with functional dysphonia showed psychiatric abnormalities (Stapleton et al. 2020; White et al. 2012; Dietrich et al. 2008). In addition, a connection has been shown between certain previous illnesses, their medication (reflux disease, depression, hearing loss, respiratory diseases, severe general diseases, hormonal disorders) and the occurrence of voice disorders. To this end, questions about previous illnesses and any medication intake could have been included in the list of questions for this study.

Despite the aforementioned limitations, the study had several strengths. In fact, the questionnaire mainly included short questions, and the scales were frequently used in the reviewed literature and were validated outside Germany. The study included 93 university professionals, which increased the precision of the outcomes and might have ensured a good representativeness of the German university teaching professionals, especially professors and lecturers, which may have improved the generalisability of the results. However, due to the limited time frame of the project, the ideal sample size could not be achieved. The ideal sample size was calculated using Qualtrics' sample calculator (https://​www.​qualtrics.​com/​de/​erlebnismanageme​nt/​marktforschung/​stichprobenrechn​er/​) and given with a confidence level of 95%, a population size of 400.100 and a margin error of 5%, an ideal sample size of 384 probands. Therefore, the results cannot be transferred to the population of the sample.

Moreover the results should be interpreted cautiously, because all data referred to Germany. Generalisation to other countries with comparable academic conditions is conceivable and would have to be examined accordingly.

Finally, additional research should be conducted using other methods, for example, longitudinal studies or qualitative studies, in order to support the findings.