Voice Disorders in Teacher Students—A Prospective Study and a Randomized Controlled Trial

ARTICLE IN PRESS Voice Disorders in Teacher Students—A Prospective Study and a Randomized Controlled Trial *Ann-Christine Ohlsson, *Eva M. Andersson, †Maria Södersten, ‡Susanna Simberg, §Silwa Claesson, and *Lars Barregård, *§Gothenburg, Sweden; †Stockholm, Sweden; and ‡Turku, Finland Summary: Objectives. Teachers are at risk of developing voice disorders, but longitudinal studies on voice problems among teachers are lacking. The aim of this randomized trial was to investigate long-term effects of voice education for teacher students with mild voice problems. In addition, vocal health was examined prospectively in a group of students without voice problems. Methods. First-semester students answered three questionnaires: one about background factors, one about voice symptoms (Screen6), and the Voice Handicap Index. Students with voice problems according to the questionnaire results were randomized to a voice training group or a control group. At follow-up in the sixth semester, all students answered Screen6 again together with four questions about factors that could have affected vocal health during their teacher education. The training group and the control group also answered the Voice Handicap Index a second time. Results. At follow-up, 400 students remained in the study: 27 in the training group, 54 in the control group, and 319 without voice problems at baseline. Voice problems had decreased somewhat more in the training group than in the control group, but the difference was not statistically significant (P = 0.1). However, subgroup analyses showed significantly larger improvement among the students in the group with complete participation in the training program compared with the group with incomplete participation. Of the 319 students without voice problems at baseline, 14% had developed voice problems. Conclusions. Voice problems often develop in teacher students. Despite extensive dropout, our results support the hypothesis that voice education for teacher students has a preventive effect. Key Words: Teacher students–Voice problems–Long-term voice training–Screening instrument–Voice Handicap Index. INTRODUCTION Teaching is a professional activity with high demands on communicative capability and voice function. Most teachers are dependent on a voice that functions well for many hours every day throughout their career. At the same time, the teacher’s voice must be clear and comfortable to listen to in order to make learning easy for the pupils. Teachers are exposed to many environmental factors in the workplace that might affect their voice and eventually lead to voice problems. The most obvious risk factor is speaking against high background noise, in poor room acoustics, and in rooms with overly dry air. Furthermore, feelings of stress or distress may influence the voice in a negative way. The importance of voice rest for teachers with heavy vocal load has also been stressed, but often teachers are short of time for vocal recovery during the workday.1–6 Voice problems also undermine the teachers’ interaction with their pupils and affect learning.7–9 Teachers are overrepresented among patients with voice disorders.10–12 Common symptoms are vocal fatigue, hoarseness, and pain or lump sensation in the throat without having a

Accepted for publication September 3, 2015. From the *Occupational and Environmental Medicine, Institute of Medicine, University of Gothenburg, Gothenburg SE-405 30, Sweden; †Department of Clinical Science, Intervention and Technology, Division of Speech and Language Pathology, Karolinska Institutet, Stockholm SE-141 86, Sweden; ‡Åbo Akademi University, Logopedics, FIN20500 Turku, Finland and Department of Special Needs Education, Faculty of Educational Sciences, University of Oslo, Oslo 0317, Norway; and §Department of Pedagogical, Curricular and Professional Studies, University of Gothenburg, Gothenburg SE-405 30, Sweden. Address correspondence and reprint requests to Ann-Christine Ohlsson, Institute of Medicine, Occupational and Environmental Medicine, University of Gothenburg, 405 30 Gothenburg, Sweden. E-mail address: [email protected] Journal of Voice, Vol. ■■, No. ■■, pp. ■■-■■ 0892-1997 © 2015 The Voice Foundation. Published by Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jvoice.2015.09.004

cold. The prevalence of voice problems among teachers varies from 5% to 80% depending on the type of teachers studied (eg elementary school teachers, high school teachers, music teachers, science teachers), the phrasing of the questions, and the response rate.13–15 Lyberg Åhlander et al16 reported a prevalence of 13% among teachers in 23 randomly selected schools in Sweden, and a study in Finland found that the prevalence of voice problems among teachers increased from 5% to 20% over a period of 12 years.17 In a study by Roy et al,18 1243 teachers and 1288 nonteachers from Iowa and Utah were randomly selected for a telephone interview regarding voice problems. The prevalence of current voice problems was higher among the teachers than among the nonteachers (11.0% vs. 6.2%); the lifetime prevalence of voice disorders was higher among the teachers (57.7% vs. 28.8%); and more teachers than nonteachers had consulted a health care professional because of a voice disorder (14.3% vs. 5.5%). Similar results were presented by Behlau et al19 in a study of 1651 teachers and 1614 nonteachers in Brazil. A higher prevalence of current voice disorders was found among teachers compared with nonteachers (11.6% vs. 7.5%), and 12.1% of the teachers reported that they had been absent from work because of voice problems on five or more days in the past year, in comparison with 2.4% among the nonteachers. One potential risk factor for developing voice problems among teachers is insufficient education in voice ergonomics and lack of voice training, as highlighted by Fritzell,11 Vilkman,20 and others. One definition of voice ergonomics is “awareness of workrelated risk factors for voice disorders and knowledge about how to improve voice production and speech intelligibility in different working environments with the goal to prevent occupational voice disorders.”21 Research during recent decades has shown the importance of preventative voice education for teachers.22–26

ARTICLE IN PRESS 2 In a study by Bovo et al,27 a group of 21 female teachers received preventative voice treatment and guidance over 3 months, and the results from voice assessments were compared with those in a control group of teachers without treatment. The participants answered questionnaires on vocal self-evaluation, and their voices were recorded for acoustic and auditory perceptual voice analyses. All participants were also examined by videolaryngoscopy. There was an evident improvement in the treated group directly after the intervention, with results from questionnaires such as the Voice Handicap Index (VHI) showing a significant decrease in voice symptoms at 3 months post baseline. However, it is crucial to also study the effects of preventative voice care on a long-term basis; so far, no long-term study has been conducted. Preventative voice education for teachers and teacher students in university programs is still rare in Sweden.28,29 Of 25 university programs for teacher students in Sweden, only nine provide compulsory voice education, though a few also teach voice ergonomics. International studies show that teacher students with weaknesses in voice function are at risk of developing a voice disorder during their teaching career.30,31 A study by Simberg et al32 showed that 20% of 226 teacher students in Finland had voice disorders. Similar results were found by Ohlsson et al,33 showing voice problems in 17% of 1250 Swedish teacher students in their first semester (mean age 23 years). Individual risk factors for the students with voice problems included previous vocal and speech problems in childhood or adulthood, frequent throat infections, airborne allergy, smoking, hearing problems, voice-demanding work, and voice-demanding hobbies. Voice problems were more common among women. In another study comparing different student groups, Simberg et al34 found that voice disorders were more common among teacher students than among other student groups. The results indicate that special attention should be paid to the voice care of teacher students. Prevention can be planned and implemented on three levels, according to the time of intervention: primary, secondary, or tertiary.35 Primary prevention is implemented before the problem has even occurred, whereas secondary prevention signifies identification of the problem and early intervention. Tertiary prevention aims at remediating the problem, that is, to minimize the impairment, disability, and handicap of a condition. This article reports on a prospective and randomized controlled trial for secondary prevention in students with voice problems, and a prospective study of students without voice problems. The aims of this work were (1) to investigate the long-term effects of a program for voice education, and (2) to study possible changes in vocal health over time in students without voice problems at baseline. METHODS Design and participants The randomized controlled study was carried out at the teacher programs of two universities in southern Sweden, University of Gothenburg and Linnaeus University, between 2009 and 2013. The design of the 3-year longitudinal study is shown in Figure 1.

Journal of Voice, Vol. ■■, No. ■■, 2015

FIGURE 1. Study design. A letter providing information about the study was sent to the new teacher students before the start of their first semester, together with the general welcome letter from the departments of education. Later, at the general introduction for the new students, further information about the study was presented by the project leader and the students were invited to participate. Of the 1250 students who agreed to participate in the autumn 2009/ spring 2010 semester (76% out of the 1636 that were approached), 208 had voice problems according to their answers to the questionnaires, as described below.33 Five of these left the teacher program before randomization, leaving 203 students for randomization (Figure 1); 107 were randomized to voice training and 96 to the control group. Because of early dropouts, voice screening was performed again with new first-semester teacher students in the autumn 2010/spring 2011 semester; this provided 88 additional students, of whom 53 were randomized to voice training and 35 to the control group. Thus, in total, 160 students were randomized to voice training and 131 to the control group. The reason for oversampling students for the intervention

ARTICLE IN PRESS Ann-Christine Ohlsson et al

Voice Disorders in Teacher Students

group was an anticipated higher dropout rate in this group. There were 133 dropouts in the voice training group: 64 rejected participation before start of the course, 23 did not turn up to the first session, 36 left the teacher program, and 10 left for other reasons. There were 77 dropouts in the control group: 28 left the teacher program and the rest dropped out for no known reason. Thus, at follow-up, 27 students remained in the voice training group and 54 in the control group. Of the 1042 students in the group without voice problems at baseline, 607 remained in the teacher program in the end of this study and 319 of those participated at the follow-up. Thus, the results of the present study are based on 400 participants: 27 in the voice training group, 54 in the control group, and 319 in the group without voice problems at baseline. At follow-up, questionnaires were administered in two ways. Students ending the teacher program in autumn 2012 and autumn 2013 completed the questionnaire on the web, and students ending the program in spring 2013 were mailed the questionnaire in paper form. The results showed no difference in response rate between web and paper; responses were received to 316 of 538 web forms (59%) and 84 of 150 paper forms (56%). During their teacher education, some of the students in both groups (22% in the training group and 30% in the control group) had voice training other than the intervention offered in this study. This took place within compulsory courses for future music teachers, in drama courses for future preschool teachers, and during leisure time (controls only). Procedure and questionnaires Baseline—the first semester All of the teacher students who agreed to participate signed a letter of consent and answered three questionnaires: one about background factors, one about voice symptoms, and the Swedish version of the VHI. Questionnaire about background factors33. The questionnaire included 11 questions about individual health- and workrelated risk factors: vocal-fold problems in childhood or as an adult, frequent throat infections, airborne allergy, smoking, hearing problems, voice-demanding work or hobbies, previous speech therapy or voice training, and mother tongue other than Swedish. The answer alternatives were “yes” and “no,” with space provided for the students’ comments. Questionnaire for screening vocal symptoms (Screen6)33. The Screen6 questionnaire was developed by Ohlsson et al33 on the basis of a larger questionnaire used for voice screening by Simberg et al.36 Screen6 includes six questions about vocal symptoms in the absence of a cold: (1) Does your voice become strained or tired? (2) Does your voice become low or hoarse? (3) Does your voice break? (4) Do you have difficulty being heard? (5) Do you need to clear your throat or cough? (6) Do you have a sensation of pain or lump in the throat? Each response alternative was scored on a four-point scale (0 = never; 1 = less often; 2 = every week; 3 = every day). The results from Screen6 were used to categorize the students into

3

those with voice problems and those with no voice problems. Students who stated that they experienced at least two of the six voice symptoms daily or weekly were classified as having voice problems and were randomized to a voice training group or a control group. Students who had fewer than two symptoms daily or weekly were assigned to the group without voice problems. Swedish version of the VHI37. The VHI contains 30 statements representing three different aspects of voice experience in everyday life: physical (P), functional (F), and emotional (E). Each aspect is represented by 10 statements, for example: “I run out of air when I talk” (P), “I use the phone less often than I would like” (F), and “I am tense when talking with others because of my voice” (E). Each statement is scored on a five-point scale (0 = never; 1 = almost never; 2 = sometimes; 3 = almost always; 4 = always). The summed score for each subdimension ranges from 0 to 40 points. The highest total score of 120 points indicates the highest possible negative impact of voice problems on everyday life. Follow-up—the last (sixth) semester All students answered Screen6 a second time at follow-up, including those with voice problems at baseline (voice training group and control group) and those without. They also answered four additional questions about circumstances during the teacher program that could have affected their vocal behavior or vocal health: voice training other than that offered in the project, voice symptoms during the period of practice, sick leave or other interruptions of the teacher education program, and current smoking. These four additional questions were administered on the web or on paper by mail together with Screen6. The VHI questionnaire was, however, not included. The reason was the large proportion of dropouts (see the section “Dropouts and study power” in the discussion section). To minimize dropout at followup, the questionnaire had to be short, and therefore VHI was excluded. However, VHI was administered a second time to the voice training group (all 27 students responded) and to the controls (25 of which responded) either at the occasion of the interview described below or by e-mail. At the end of the last semester, all students in the voice training group and 10 students chosen at random from the control group were interviewed to elicit their thoughts about whether and how future compulsory voice education for teacher students should be designed. The voice education The aim of the voice education program was to start the process of increased awareness of voice ergonomics among the participating students during their time in the teacher education program, and to give them basic tools for long-term professional vocal improvement. After voice screening and randomization for intervention during their first semester, the students started their voice education with an individual voice consultation with the voice teacher. They received information about their questionnaire responses at baseline and about the voice education program. The voice teacher was a speech language pathologist with long experience in the field (the project leader and the first author). During the second semester, a basic voice course of 12 hours

ARTICLE IN PRESS 4 was held in groups of five to seven students. The voice training technique taught by the voice teacher followed the socalled verbal instruction model.38 In this model, instead of demonstrating the voice task with her own voice and body, the speech therapist verbalizes the task to the patient. While the patient investigates different options for performing the task, the preponderate importance of the speech language pathologist’s support is on the patient’s search and discovery of performance options, more than on the patient’s vocal behavior. The purpose is to start the patient’s process of transfer-to-speech and carry over of the new voice behavior into daily life right at the beginning of the treatment. The course taught vocal physiology and voice ergonomics along with basic training of voice coordinated with breathing, posture, relaxation, and body movements. Voice function was trained in playful exercises for resonance and articulation in words and texts, and in simulated situations. Tempo, pauses, and eye contact were trained together with voice and body gestures in stories and conversations. Exercises for projecting the voice were used to help extend the pitch range and intensity range. The students received a course compendium describing voice physiology and ergonomics, and presenting the content of each training session together with the texts used. The voice training tasks performed in the different training sessions were described in detail. The compendium also contained a checklist of all parts of the voice education program, such as voice physiology, ergonomics, mindfulness, basic voice training, resonance training, feedback during their own teaching, and homework. The students also received a pamphlet on voice function and voice ergonomics, produced by the Swedish Teacher Union, and a logbook to record their own notes and questions about voice function during everyday life. Between the voice lessons, the students had individual homework such as basic voice training at a certain time every day, transfer-to-speech once a day in a certain situation, and consciously using a pause between two sentences once a day in communication. At the end of the course, the students were recommended to continue selftraining with help of the course compendium, to continue making notes in the logbook, and to contact the voice teacher for individual support if needed. A few students took up the offer to contact the voice teacher for support. During the third and fourth semesters, the students were invited to a group training session in which 20 of 27 participated; those who did not participate in these sessions had individual consultations instead. All students were offered feedback from the voice teacher during one lesson held during a practice period, and 15 of the 27 participants used this opportunity. Before the lesson started, the student and the voice teacher agreed on what the student wanted to train during the lesson and what the voice teacher should observe and give feedback on. After the lesson, the student and the voice teacher evaluated the lesson in terms of voice and speech aspects. From the fifth semester, the students were given the initiative to contact the voice teacher for training and consultation. Half of the group used this opportunity. During the students’ last semester in the teacher education program, they all had an individual consultation with the voice teacher, covering the students’ own experiences until then regarding their own voice use and the voice

Journal of Voice, Vol. ■■, No. ■■, 2015

training, as well as thoughts about how to continue voice training in their future profession. In all, the schedule of the voice education program contained 20 hours of training (group and individual) for each student, complemented with consultations on the students’ own initiatives. Moreover, extra training hours were given outside the schedule for students who were unable to participate at the scheduled times. Interview regarding the students’ views on future voice education at the teacher program At the last meeting with the voice teacher, the students were also interviewed about their thoughts and recommendations for how future voice education for teacher students should be shaped. The open questions were based on the following issues: Should there be compulsory voice education? If yes, where should it be placed in the teacher program? What form should it have? What content should it have? How can students continue their voice education in their professional life after teacher education? The questions were put orally and individually to all 27 participants in the training group and to a random selection of 10 of the 54 participants in the control group at the end of their last semester. The aim of asking these questions was to capture the opinions of students ending their teacher education program who had fresh experience of the complete program. Statistical analysis Screen6 allows participants to report between 0 and 6 frequently occurring symptoms (FOS), that is, symptoms with a score of 2 or 3. In addition to analyzing the number of FOS, responses were dichotomized into few FOS (0 or 1) or many FOS (2 or more), and a transition from many FOS at baseline to few FOS at follow-up was considered to indicate an improvement. Each Screen6 symptom was dichotomized into “infrequent” (score 0 or 1) or “frequent” (score 2 or 3), and each symptom could be infrequent at baseline but frequent at follow-up (deterioration), frequent at baseline but infrequent at follow-up (improvement), or have the same classification both at baseline and at follow-up (no change). For VHI, the total score as well as the score for each subdimension (F, P, and E) was calculated for each person at baseline and at follow-up. The total VHI score was also dichotomized into “low” (<20) or “high” (≥20 or more). A change from low to high was considered to constitute deterioration, a change from high to low was considered to constitute an improvement, and having the same classification both at baseline and at follow-up was considered to indicate no change. For number of FOS and the VHI score, changes within groups (between baseline and follow-up) were examined with a paired t test. Group comparisons were made using either the Wilcoxon two-sample test or the t test for independent groups for the semicontinuous variables (number of FOS and VHI score), whereas group comparisons for the dichotomized variables (few FOS vs. many FOS, and low VHI vs. high VHI) were made using either the chi-square test or Fisher’s exact test. The choice of test depended on the group sizes. The compared groups consisted of (1) training versus control; (2) the following four questions: “Voice training other than that offered in the project” (yes/no), “Voice

ARTICLE IN PRESS Ann-Christine Ohlsson et al

5

Voice Disorders in Teacher Students

symptoms during the period of practice” (yes/no), “Sick leave or other interruptions of the teacher education program” (yes/ no), and “Current smoking” (yes/no); (3) voice coaching or not; and (4) participating in all training sessions versus absence from at least two of the eight sessions. Regression analysis was performed with baseline factors as potential explanatory variables (vocal-fold problems in childhood or adulthood, frequent throat infections, airborne allergy, smoking, hearing problems, voice-demanding work, voicedemanding hobbies, previous speech therapy or voice training, mother tongue other than Swedish), both for the change in the number of FOS (linear regression) and for the dichotomized FOS at follow-up (logistic regression). Regarding the proportion of participants with an improvement, deterioration, or no change (in FOS and VHI respectively), chi-square tests were used to test the group difference between the voice training group and the control group. The change within each group between baseline and follow-up was tested using a McNemar test. The correlation between the change in VHI total (between baseline and follow-up) and the change in FOS in Screen6 was assessed with Spearman correlation. Ethical approval was given for the study by the regional ethical review board in Gothenburg, May 4, 2009 (Approval No. 155-09).

RESULTS The numbers of students in the three groups (voice training, control, and no voice problems at baseline) are given in Table 1. There were no significant differences between the groups regarding the individual risk factors33 (vocal-fold problems, frequent throat infections, airborne allergy, smoking, hearing problems, mother tongue other than Swedish, voice-demanding work or hobbies, previous speech therapy or voice training). When the baseline results were compared with the followup results, a significant decrease was found in the number of FOS in Screen6 for the group with voice problems. At the same time, there was a significant increase in the number of FOS in the group without voice problems at baseline (Table 2). The number of FOS decreased more in the training group than in the control group (−1.85 vs. −1.28), but the difference was not statistically significant (P = 0.10, t test). Number of FOS was also analyzed as a dichotomized variable (few vs. many FOS). The results showed that a larger proportion of the students in the voice training group than in the control group no longer had many FOS (≥2) at follow-up (Table 3), but the difference between groups was nonsignificant. At the same time, 14% of the students in the group without voice problems at baseline had two or more FOS at follow-up.

TABLE 1. Age (at Baseline) and Gender in the Three Groups of Students Voice Problems at Baseline (Screen6) Age, mean (range) No, of women/men

No Voice Problems at Baseline

Voice Training, N = 27

Control Group, N = 54

N = 319

26 (19–41) 23/4 (85%/15%)

27 (18–40) 36/18 (67%/33%)

22 (18–45) 248/71 (78%/22%)

TABLE 2. Results From Screen6 Showing Number of FOS (Two or More Symptoms Weekly or More Often); Change Within Groups From Baseline to Follow-up Voice Training Group, n = 27 Mean Baseline Follow-up Difference P value (t test)

3.07 1.22 −1.85

SD 1.33 1.53 1.35 <0.0001

Range 2–6 0–5 −5–1

Control Group, n = 54 Mean 2.83 1.56 −1.28

SD 1.09 1.38 1.65 <0.0001

Students Without Voice Problems, n = 319

Range

Mean

2–6 0–5 −6–2

0.16 0.58 0.42

SD 0.37 1.17 1.16 <0.0001

Range 0–1 0–6 −1–6

TABLE 3. Number of Students With at least Two FOS at Follow-up Follow-up Baseline

Voice training group (at least 2 FOS) Control group (at least 2 FOS) Group with no voice problems (0 or 1 FOS)

(0 or 1 FOS)

(at least 2 FOS)

18 (67%) 28 (52%) 275 (86%)

9 (33%)* 26 (48%)* 44 (14%)

Notes: Voice training group (n = 27), control group (n = 54), and group without voice problems at baseline (n = 319). Percentages are within parentheses. * P = 0.20 for chi-square test comparing voice training group and control group.

ARTICLE IN PRESS

59% (16) 11% (3) 52% (14) 4% (1) 56% (30) 33% (18) 37% (20) 15% (8) Control group n = 54

Pain or Lump in Throat Throat Clearing

67% (18) 30% (8) 37% (10) 0% (0) 59% (32) 41% (22) 24% (13) 6% (3) 48% (13) 19% (5) 30% (8) 0% (0) 39% (21) 26% (14) 20% (11) 7% (4)

Difficulty Being Heard Voice Breaking

44% (12) 15% (4) 33% (9) 4% (1) 31% (17) 15% (8) 22% (12) 6% (3) 41% (11) 22% (6) 22% (6) 4% (1) 48% (26) 19% (10) 33% (18) 4% (2)

Low, Hoarse Voice Strained, Tired Voice

48% (13) 26% (7) 30% (8) 7% (2) 50% (27) 22% (12) 35% (19) 7% (4) Baseline Follow-up Improved Worse Baseline Follow-up Improved Worse Voice training group n = 27

In the training group, five of the 27 students were absent from two of the eight training sessions. Comparison between groups with complete and incomplete participation revealed that the proportion of students with at least two FOS at follow-up was considerably lower in the complete participation group than in the group with incomplete participation in the voice training (18% vs. 100%; P = 0.0016, Fisher’s exact test). At follow-up, the students answered four questions about background variables (see Statistical analysis), and the responses to these questions were tested against change in the number of FOS. One question was regarding voice symptoms during periods of practice; for students without voice problems at baseline, the proportion experiencing voice symptoms during periods of practice was 37% (118 of 319). These students showed a larger increase in the number of FOS at follow-up than those without voice symptoms during practice (0.81 vs. 0.18; P = 0.0001). Among students with voice problems at baseline (training group and control group), the proportion of students experiencing vocal symptoms during practice was around 65% (33 of 54 controls, and 17 of 27 in the training group). Decrease in the number of FOS was significantly larger among those without voice symptoms during practice (−2.13, n = 31) compared with those with voice symptoms (−1.06, n = 50; P = 0.0024, t test). In the training group, there was no difference in the change in the number of FOS at follow-up between those with and without voice symptoms during practice (−1.59 vs. −2.30). For the control group, on the other hand, the decrease in the number of FOS at followup was smaller for students with voice symptoms during periods of practice than for those without (−0.79 vs. −2.05; P = 0.0052). Students who received feedback from the voice teacher during periods of practice were compared with students without such feedback. The results showed that the proportion of students with at least two FOS at follow-up was significantly larger in the group without feedback than in the group who received feedback (54% vs. 14%; P = 0.0461, Fisher’s exact test). Regression analyses showed that some of the individual risk factors at baseline were associated with the change in the number of FOS between baseline and follow-up. In the training group, students with airborne allergy decreased their voice symptoms significantly more than did students without airborne allergy (−2.67 vs. −1.59). In the group without voice problems at baseline, students with airborne allergy showed an increase in voice symptoms compared with students without airborne allergy (0.65 vs. 0.32), and those who had been smokers increased their voice symptoms in comparison with nonsmokers (0.79 vs. 0.34). Regression analyses showed no association between risk factors at baseline and the proportion of students with at least two FOS at follow-up in the voice training group. In the larger group without voice problems at baseline, the proportion with at least two FOS was higher among students with airborne allergy compared with the students without this allergy (22% vs. 10%), and also higher among previous smokers compared with nonsmokers (24% vs. 12%). The most common symptoms among the students with voice problems at baseline were pain or lump in throat, and throat clearing (Table 4). Although in four of the six symptoms the improvements were larger in the training group, there was no

Journal of Voice, Vol. ■■, No. ■■, 2015

TABLE 4. Proportion of Students Experiencing Each of the Six Symptoms at least Every Week, at Baseline and at Follow-up (Number of Students Within Parentheses)

6

ARTICLE IN PRESS Ann-Christine Ohlsson et al

60

Voice training Control

50 Difference (%)

7

Voice Disorders in Teacher Students

40 30 20 10 0 Strained, Low, hoarse Voice breaks Difficulty tired voice voice being heard

Throat clearing

Pain or lump in throat

FIGURE 2. Difference between percent improved and percent deteriorated from baseline to follow-up in the voice training group and the control group for each of the six symptoms in Screen6.

significant difference between groups (Figure 2); the smallest P value was 0.22 for pain or lump in throat (chi-square tests). Figure 2 shows the difference in change of symptom prevalence from baseline to follow-up for the voice training group and the control group. For each symptom, the proportion with improvement was larger than the proportion with deterioration in both groups. In the voice training group, 48% (13 students) frequently TABLE 5. VHI Total Scores and Differences Between Baseline and Follow-up for Voice Training Group and Control Group

Voice training group Baseline Follow-up Difference P value (t test) Control group Baseline Follow-up Difference P value (t test)

N

Mean

SD

Min; Max

27 27 27

20.67 17.52 −3.15 0.13

14.01 11.06 10.50

3;52 3;56 −29;17

25 25 25

19.80 19.48 −0.32 0.86

11.17 12.57 9.04

5;42 3;57 −11;20

experienced a strained, tired voice at baseline, and 26% (seven students) reported the same at follow-up (Table 4). Also in this group, 30% (eight students) had improved and 7% (two students) had deteriorated, giving a difference of 23% (dark bar). In the control group, 35% (19 students) had improved and 7% (four students) had deteriorated, giving a difference of 28% (light bar). Total VHI scores did not differ significantly between baseline and follow-up for either group (Table 5). For the three aspects of VHI (F, P, and E), there was one significant change; the score for emotional aspect in the voice training group decreased from 5.04 at baseline to 2.81 at follow-up (t test, P = 0.03). The mean total VHI score decreased more in the training group than in the control group (−3.15 vs. −0.32), but the difference between groups was not significant (P = 0.30, t test). None of the three individual aspects (F, P, or E) differed between the groups. For the four background questions (other voice training, voice problems during practice, long absence from teacher program, smoking), there was no significant difference between groups regarding change of VHI score between baseline and follow-up. There was no significant difference between baseline and follow-up in the proportion of students that scored at least 20, in any of the groups (Table 6).

TABLE 6. Number of Students Scoring at least 20 for VHI Total, at Baseline and Follow-up Baseline Score <20 Voice training group (n = 27)

Follow-up

Score <20 Score ≥20

14 1

P value (McNemar) Control group (n = 25)

Follow-up

Score <20 Score ≥20

12 3

P value (McNemar)

Score ≥20 5 7 0.10 2 8 0.65

ARTICLE IN PRESS 8

Journal of Voice, Vol. ■■, No. ■■, 2015

FIGURE 3. Correlation between difference in VHI score and difference in the number of frequently occurring Screen6 symptoms (difference between baseline and follow-up). The improvement seemed to be larger in the training group (18% vs. 8%) and the deterioration smaller (4% vs. 12%), but the difference was not significant (P = 0.33, chi-square test). Regression analyses showed no association between risk factors at baseline and change in total VHI score at follow-up in the training group or in the control group. The correlation between the difference in total VHI score and the difference in the number of FOS in Screen6, between baseline and follow-up, was 0.41 (P = 0.03; Figure 3). Results from the interview with the students at their last individual consultation with the voice teacher are presented in Table 7. The students were asked about their opinions regarding future voice education of teachers and student teachers. DISCUSSION The aims of this prospective study were (1) to investigate longterm effects from preventive voice education in teacher students with mild voice problems, and (2) to study possible changes in vocal health during teacher education in students without voice problems. The students participated for more than 3 years, from their first to their last semester, which makes this study unique. The intervention part was planned to consist of 100 students in voice training and 100 students in the control group. However, a large dropout decreased the number of students in both groups, to 27 students in the voice training group and 54 students in the control group, which caused a decrease in the statistical power. The results of the intervention study showed a somewhat larger improvement regarding FOS in the trained group, but the difference was not statistically significant (P = 0.1). The most frequent symptoms in both groups were “a need to clear the throat

or cough” and “a sensation of pain or lump in the throat.” In patients with functional voice disorder, these two symptoms are typical and among the most frequent. In this study, both symptoms showed the largest improvement in the training group, compared with the other four symptoms. There was no background question specifically about reflux or of taking medications for that. However, it is likely that some students in our study had reflux, probably in both groups, and we do not think it affects the group comparison. Because of limited resources, laryngological examinations could not be performed on the students. Persistent voice problems were reported at follow-up by 33% of the trained group and 48% of the controls, meaning that improvements were seen not only in the training group but also in the control group. In comparison, results from a similar study in Finland with a more homogenous group of teacher students with mild voice problems39 showed a significantly larger proportion of students with improvement in the voice training group than in the control group. However, in that study, the posttraining assessments were performed at a shorter term: 2 weeks after training and follow-up 9 months later. After 2 weeks, 5% of the students in the treatment group and 45% of the controls still had FOS, whereas 9 months later the corresponding proportions were 15% in the treatment group and 50% in the control group. In the present study, voice problems persisted after 3 years in 33% of the trained students and 48% of the controls. The intervention study included only students with voice problems at baseline, which were randomized to either the training group or the control group. The fact that improvement was seen also in the control group may be due to regression toward the mean; because voice symptoms usually are fluctuating, it is likely

ARTICLE IN PRESS Ann-Christine Ohlsson et al

9

Voice Disorders in Teacher Students

TABLE 7. Interview Responses From Students About Future Voice Education in the Teacher Program Issues Compulsory voice education Placement of voice education in the teacher program

Form of the voice education

Content of the voice education

Continuation in the profession

Training Group n = 27

Controls n = 10

All students answered yes.

All students answered yes.

Lecture (physiology, voice ergonomics): year 1. Basic course (group training): year 2. Intensified voice training and guidance: during practice periods; feedback during students’ teaching; students carrying out their own specific voice work. Voice education of tutors. Like the project’s group training but much more time and together with classmates.

Lecture: at start of teacher education. Voice course and training in groups: in the middle of teacher education. Main training, feedback, and voice tasks: during practice periods.

Theory (voice physiology and ergonomics). Practical training of voice and clearness of speech, and own verbal and nonverbal communication. Feedback and voice training above all at practice periods. Examination tasks. Recurring training, each semester. Education of head teachers at schools and personnel at occupational health consultations.

that, even without treatment, some students with voice problems at baseline will be symptom-free at follow-up. However, this was a long-term study, and the follow-up was performed 3 years after the start of the study and only once. Thus, uncontrollable extraneous factors can also bias the results. We do not know the extent to which individual students were exposed to vocal load during their time in the teacher program, and variability in vocal load will make it more difficult to detect an effect of voice training. Also, the students constituted a heterogeneous group in terms of their education, with future professions including preschool teacher, high school teacher, or teacher of specific subjects such as music or sports. For some of the students, other voice training was offered within compulsory courses for future music teachers and in drama for future preschool teachers, and some of the controls had participated in voice courses during their leisure time. In all, 22% of the students in the training group and 30% of the controls had participated in other voice training during their teacher education, and it is possible that some students in the control group improved because of this. In our opinion, there are several results indicating a protective long-term effect from the voice training. First, even if not statistically significant, there was a strong tendency (P = 0.1) toward an overall effect of training. Second, the comparison between subgroups showed a significantly lower proportion of students with at least two FOS at follow-up in the complete participation group (18%) than in the group with incomplete participation (100%). Third, the proportion of students with at least two FOS at follow-up was significantly lower among students in the training group who had received feedback from the

Basic course (physiology, work environment, group training). Main training and feedback during students’ teaching during practice periods. Practical training for voice and speech in small groups with classmates. Feelings, gestures, and voice. Voice tutoring and training during practice periods. Feedback of speech and voice during own teaching. Examination tasks for voice. Recurring training (at study days or other). Education of leaders at school and of staff at occupational health consultations.

voice teacher during periods of practice compared with the students without feedback on voice during practice (14% vs. 54%). Another indication of training effect is that in the training group the students with airborne allergy at baseline had decreased their voice symptoms at follow-up significantly more than the students without allergy. It is possible that some voice symptoms among the students with allergy could partly be explained by strained vocal behavior because of the allergy. Similar results were reported by Millqvist et al,40 Simberg et al,41 and Ohlsson et al.42 After participating in the voice training, these students might have learned new and better strategies for voice function that relieved some of their voice symptoms. In our study, the group without voice problems at baseline had no voice training, and at follow-up 14% of them had developed problems. Thus, these students’ voices deteriorated during their time in teacher education. Part of this change could be explained by regression toward the mean, but vocal strain during their education and practice as teachers could have also been responsible. A larger proportion of the students with airborne allergy at baseline in this group had voice problems at followup than the students without allergy, indicating an injurious effect on voice from airborne allergy.40–42 It had been interesting to include also, in the randomized trial, students without voice problems at baseline (with the aim of primary prevention), but the training resources of the present study did not allow it. Another background factor affecting voice in the group without voice problems at baseline was smoking. When comparing students who were smokers at baseline to the nonsmoking students, a larger proportion of smokers had voice problems at follow-up.

ARTICLE IN PRESS 10 A study by Simberg et al43 found no difference in reported voice symptoms between male smokers and male nonsmokers. However, female smokers had more voice symptoms than female nonsmokers. The results are in accordance with those in our study, where most students were women. Vocal fatigue or other vocal symptoms during periods of practice were also experienced by students in the larger group without voice problems at baseline. In this group, the students with vocal fatigue or other symptoms during periods of practice showed larger deterioration (increase in FOS) at follow-up compared with those without vocal symptoms during practice. The result indicates a long-standing negative effect from teacher practice on the voices of the students in the group without voice problems at baseline. At follow-up, the VHI was assessed only for the voice training group and the control group. To minimize dropout, we decided not to include VHI in the follow-up questions to the group without voice problems at baseline. At that point of time, we anticipated large dropout due to high proportion of teacher students interrupting their studies. Results from VHI assessments showed no significant change in total VHI score in either of the groups, whereas one of the subscales (emotional aspect) showed significant improvement only in the training group. It is possible that worries about voice function decreased in the training group, owing to their increased knowledge from participating in the training, which was not the case in the control group. There was only a moderate correlation between the difference in VHI score and the difference in the number of FOS in Screen6 (rs = 0.41), which is in accordance with findings that VHI and subjective voice ratings seem to assess different aspects of voice function.37,44 The VHI items refer to everyday experiences of problems appearing due to voice dysfunction, whereas the subjective ratings are more related to symptoms. The comparisons between students with complete participation in voice training and students with incomplete participation, as well as the comparisons between students with and without voice coaching at practice periods, indicate that it would be beneficial to provide more time for voice training than that offered in this study. Because of the design of the study, the number of training sessions had to be restricted, as the voice education had to be performed outside schedule. In addition, the randomization implied that the students in the training group had to participate without their classmates. The need for training was confirmed in the interview responses from the students at the end of their last semester. All the students in the training group found the design of the voice education in the study to be a good model for future voice education in the teacher program, but they felt it should be given within the teacher education schedule, be given much more time, and conducted together with their classmates. Dropouts and study power The risk of dropouts in longitudinal studies is well known.45–48 Moreover, participants were also lost to follow-up in the present study because a large proportion of students at Swedish teacher education programs interrupt their studies. A Swedish report published in 200849 stated that 35% of teacher students interrupt their

Journal of Voice, Vol. ■■, No. ■■, 2015

studies, and the proportion has increased rapidly during later years. In comparison, only 9% of students in other university programs interrupt their studies. In addition, the number of students applying for teacher education in Sweden has decreased successively since 2001. The abovementioned report49 proposes that the negative media attention regarding teacher education in Sweden in recent decades may have contributed to this decline in enrollment in teacher education. In the present study there were 1042 students without voice problems at baseline and 203 students with voice problems at baseline. Because of early dropouts in the intervention part of the project, voice screening was later performed with new teacher students in their first semesters. Students without voice problems at this later screening were not included; that is, the group of students without baseline voice problems still consisted of the original 1042. This study consists of two arms, a secondary prevention part (students with voice problems) and a nonintervention followup (students without voice problems). It is possible that it would have been valuable to include the students without voice problems in the intervention; that is, a primary prevention study should have been performed instead. However, in the present study, primary prevention would have decreased the statistical power because fewer students would have had symptoms at followup. We would then have needed larger groups (than 100 + 100) for randomization, and we did not have the resources for training of so many students. Thus, a total of 291 students with voice problems were included in the intervention study, and 81 of them responded, giving a response rate of 28%. The proportion of students that reported on frequent voice symptoms (a symptom that occurs at least every week) at baseline did not differ significantly between the training group and controls, which suggests that the randomization result was correct despite the dropouts. In the group without voice problems at baseline, 723 of 1042 were lost to follow-up, giving a response rate of 31%. With respect to voice symptoms, if those who drop out are different from those who respond, it would bias our results. The large number of dropouts decreased the statistical power to show a significant effect of voice training. The original plan was to follow about 100 students with voice training and 100 controls, which would have provided an acceptable power. When the high dropout rate was noted, additional students were recruited, but the number of students completing the training was still limited. The reasons for the dropouts in the group without voice problems at baseline were not studied. The primary reasons for the decreased number of students in the group with voice problems at baseline were that they rejected participation before the start of the voice education or that they had interrupted their teacher education. The students often reported stress during their first semester of teacher education. Gustavsson50 points out that when students start their teacher education they often misunderstand or have difficulty interpreting what is critical in teaching practice. The teacher’s performance in the classroom is one such critical factor, and an important aspect of teacher performance is nonverbal communication, including the voice. International studies have shown that teacher students have low awareness of themselves as future professional voice users.30,31 However, vocal

ARTICLE IN PRESS Ann-Christine Ohlsson et al

Voice Disorders in Teacher Students

behavior is considered an integrated part of bodily expression and is regarded as a professional tool closely connected to the development of the pedagogic self-image.51 Deficient knowledge about the high vocal demands of the teacher profession could be one reason for the students’ rejecting participation in the intervention part of this study. Another related aspect could be that some students rejected participation because they might have found it inconvenient to participate in a situation where they had to display themselves and their voices in interaction with unknown persons, which was the case for the group training in this project. A study by Jensen et al52 about dropout from psychotherapy found that motivation to work with interpersonal issues was important for compliance, and that social functioning was associated with dropout. In addition, a study by Hagger and McIntyre53 showed that some teacher students lack the necessary social competences. A voice education program performed within the schedule of teacher education would imply that the students could participate together with their classmates, which might offer the student a higher degree of social safety than participating in a random group of unknown participants, as was the case for this study because of the randomization. The most common reason for rejecting participation was that the voice education was not included in the teacher program schedule, so the students would have had to find extra time for this. Another common reason for students rejecting participation in the study, and which was also commented on among the students in the training group, was that it was not possible for them to participate in the voice education together with their classmates. It is likely that participation together with classmates would have motivated some of the dropouts to participate in the voice training. Thus, on the basis of our experience from this study, we recommend future researchers to have the following in mind before performing the baseline assessments: (1) find out the expected proportion of students interrupting their studies and the typical time for interrupting; (2) examine which student groups will already get some form of voice training within the program; (3) choose homogenous student groups, for example, students who plan to become preschool teachers, elementary school teachers, music teachers, sports teachers, etc; (4) choose students from the same classes, making it possible to have voice training together with their classmates; (5) perform baseline assessments and start the intervention study not too early in the teacher program, preferably after the first practice period; (6) if possible, perform the voice training within the schedule of the teacher program. CONCLUSIONS Voice problems often develop in teacher students. Despite extensive dropouts, the results from this randomized longitudinal trial support the hypothesis that voice education for teacher students has a long-term preventive effect. The training part of preventive voice education for teacher students should not start until after their first practice period. It is crucial that the design of the voice training course allows it to be included in the schedule for the teacher program and that the students can participate together with their classmates.

11

Acknowledgments This study was supported by the Swedish Council for Working Life and Social Research (2011-0470). REFERENCES 1. Kooijman PG, Thomas G, Graamans K, et al. Psychosocial impact of the teacher’s voice throughout the career. J Voice. 2007;21:316–324. 2. Smolander S, Huttunen K. Voice problems experienced by Finnish comprehensive school teachers and realization of occupational health care. Logoped Phoniatr Vocol. 2006;31:166–171. 3. Sliwinska-Kowalska M, Niebudek-Bogusz E, Fiszer M. The prevalence and risk factors for occupational voice disorders in teachers. Folia Phoniatr Logop. 2006;58:85–101. 4. Thibeault SL, Merrill RM, Roy N, et al. Occupational risk factors associated with voice disorders among teachers. Ann Epidemiol. 2004;14:786–792. 5. Vilkman E. Occupational safety and health aspects of voice and speech professions. Folia Phoniatr Logop. 2004;56:220–253. 6. Martins R, Pereira E, Hidalgo C, et al. Voice disorders in teachers. A review. J Voice. 2014;28:716–724. doi:10.1016/j.jvoice.2014.02.008. 0892-1997/ $36.00_2014, The Voice Foundation. 7. Lyberg Åhlander V, Haake M, Brännström J, et al. Does the speaker’s voice quality influence children’s performance on a language comprehension test? Int J Speech Lang Pathol. 2015;17:63–73. doi:10.3109/17549507 .2014.898098. 8. Morton V, Watson D. The impact of impaired vocal quality on children’s ability to process spoken language. Logoped Phoniatr Vocol. 2001;26:17–25. 9. Rogerson J, Dodd B. Is there an effect of dysphonic teachers’ voices on children’s processing of spoken language? J Voice. 2005;19:47–60. 10. Titze I, Lemke J, Montequin D. Populations in the US workforce who rely on voice as primary tool of trade—a preliminary report. J Voice. 1997;11:254–259. 11. Fritzell B. Voice disorders and occupations. Logop Phoniatr Vocol. 1996;21:7–11. 12. Lyberg Åhlander V, Pelegrín García D, Whitling S, et al. Teacher’s voice use and speakers’ comfort in teaching environment: a field study using ambulatory phonation monitor. J Voice. 2015;doi:10.1016/j.jvoice .2014.03.006, 841.e5–841.e15 (in press). 13. de Jong FI, Kooijman PG, Thomas G, et al. Epidemiology of voice problems in Dutch teachers. Folia Phoniatr Logop. 2006;58:186–198. 14. Preciado J, Pérez C, Calzada M, et al. Epidemiological study of voice disorders among teaching professionals of La Rioja, Spain. J Voice. 2008;22:489–508. 15. Sala E, Laine A, Simberg S, et al. The prevalence of voice disorders among day care center teachers compared with nurses. A questionnaire and clinical study. J Voice. 2001;15:413–423. 16. Lyberg Åhlander V, Rydell R, Löfqvist A. Speaker’s comfort in teaching environments. Voice problems in Swedish teaching staff. J Voice. 2011;25:430–440. 17. Simberg S, Sala E, Vehmas K, et al. Changes in the prevalence of vocal symptoms among teachers during a twelve-year period. J Voice. 2005;19:95– 102. 18. Roy N, Merrill RM, Thibeault S, et al. Prevalence of voice disorders in teachers and the general population. J Speech Lang Hear Res. 2004;47:281– 293. 19. Behlau M, Zambon F, Guerrieri AC, et al. Epidemiology of voice disorders in teachers and non-teachers in Brazil: prevalence and adverse effects. J Voice. 2012;26:665.e9–665.e18. doi:10.1016/j.jvoice.2011.09.010. 20. Vilkman E. Voice problems at work. A challenge for occupational safety and health arrangement. Folia Phoniatrica et Logopedica. 2000;52:120–125. 21. Sala E, Sihvo M, Laine A. Röstergonomi—Rösten ett Fungerande Arbetsredskap [Voice Ergonomics—Voice as a Functional Work-tool]. Helsingfors, Finland: Institutet för Arbetshygien, Arbetarskyddscentralen; 2005. 22. Ohlsson AC. Utbildningsprogram för förebyggande röstvård [A program for preventing voice disorders]. Arbetsmiljöfondens Sammanfattningar 1371. Stockholm. 1990.

ARTICLE IN PRESS 12 23. Ilomäki I, Leppänen K, Kleemola L, et al. Relationships between selfevaluations of voice and working conditions, background factors, and phoniatric findings in female teachers. Logoped Phoniatr Vocol. 2009;34:20–31. 24. Laukkanen AM, Leppänen K, Ilomäki I. Self-evaluation of voice as a treatment outcome measure. Folia Phoniatr Logoped. 2009;61:57–65. 25. Pasa G, Oates J, Dacakis G. The relative effectiveness of vocal hygiene training and vocal function exercises in preventing voice disorders in primary school teachers. Logoped Phoniatr Vocol. 2007;32:128–140. 26. Hazlett DE, Duffy OM, Moorhead SA. Review of the impact of voice training on the vocal quality of professional voice users: implications for vocal health and recommendations for further research. J Voice. 2011;25:181–191. 27. Bovo R, Galceran M, Petruccelli J, et al. Vocal problems among teachers: evaluation of a preventive voice program. J Voice. 2007;21:705–722. 28. Söderlundh K. Röstvården på lärarutbildningarna “[Voice care in the teacher programs]. Röstläget. 2006: 5–6. 29. Ek H, Elfvengren J. Röstvård i lärarutbildningen vid Göteborgs universitet [Voice care in the teacher program at Gothenburg University]. Teacher education program, Gothenburg University. 2006. 30. Schneider B, Enne R, Cecon M, et al. Effects of vocal constitution and autonomic stress-related reactivity on vocal endurance in female student teachers. J Voice. 2006;20:242–250. 31. Thomas G, Kooijman PG, Cremers CW, et al. A comparative study of voice complaints and risk factors for voice complaints in female student teachers and practicing teachers early in their career. Eur Arch Otorhinolaryngol. 2006;263:370–380. 32. Simberg S, Laine A, Sala E, et al. Prevalence of voice disorders among future teachers. J Voice. 2000;14:231–235. 33. Ohlsson AC, Andersson EM, Södersten M, et al. Prevalence of voice symptoms and risk factors in teacher students. J Voice. 2012;26:629–634. 34. Simberg S, Sala E, Rönnemaa AM. A comparison of the prevalence of vocal symptoms among teacher students and other university students. J Voice. 2004;18:363–368. 35. Duffy OM, Hazlett DE. The impact of preventive voice care programs for training teachers: a longitudinal study. J Voice. 2004;18:63–70. 36. Simberg S, Sala E, Laine A, et al. A fast and easy screening method for voice disorders among teacher students. Logoped Phoniatr Vocol. 2001;26:10–16. 37. Ohlsson AC, Dotevall H. Voice Handicap Index in Swedish. Logoped Phoniatr Vocol. 2009;34:60–66. 38. Ohlsson AC. Verbal Instruction Model (VIM) in voice therapy. Logoped Phoniatr Vocol. 2014;Early Online 1–6. doi:10.3109/14015439.2014.949303.

Journal of Voice, Vol. ■■, No. ■■, 2015 39. Simberg S, Sala E, Tuomainen J, et al. The effectiveness of group therapy for students with mild voice disorders: a controlled clinical trial. J Voice. 2006;20:97–109. 40. Millqvist E, Bende M, Brynnel M, et al. Voice change in seasonal allergic rhinitis. J Voice. 2008;22:512–515. 41. Simberg S, Sala E, Tuomainen J, et al. Vocal symptoms and allergy—a pilot study. J Voice. 2010;doi:10.1016/j.jvoice.2007.03.010. 42. Ohlsson AC, Drevsäter A, Brynnel M, et al. Allergic rhinitis and voice change. Logop Phoniatr Vocol. 2015;doi:10.3109/14015439.2015.1049288 (in press). 43. Simberg S, Udd H, Santtila P. Gender differences in the prevalence of vocal symptoms in smokers. J Voice. 2015;29:588–591. 44. Speyer R, Wieneke GH, Dejonckere PH. Self-assessment of voice therapy for chronic dysphonia. Clin Otolaryngol Allied Sci. 2004;29:66–74. 45. Moreno-Betancur M, Chavance M. Sensitivity analysis of incomplete longitudinal data departing from the missing at random assumption: Methodology and application in a clinical trial with drop-outs. Stat Methods Med Res. 2013;doi:10.1177/09622280213490014 [Epub ahead of print]. 46. Damen NL, Versteeg H, Serruys PW, et al. Cardiac patients who completed a longitudinal psychosocial study had a different clinical and psychosocial baseline profile than patients who dropped out prematurely. Eur J Prev Cardiol. 2015;22:196–199. 47. Groeneveld IF, Proper KI, van der Beek AJ, et al. Factors associated with non-participation and drop-out in a lifestyle intervention for workers with an elevated risk of cardiovascular disease. Int J Behav Nutr Phys Act. 2009;6:80. doi:10.1186/1479-5868-6-80. 48. Ulmer M, Robinaugh D, Friedberg JP, et al. Usefulness of a run-in period to reduce drop-outs in a randomized controlled trial of a behavioral intervention. Contemp Clin Trials. 2008;29:705–710. 49. SOU 2008:109. En hållbar lärarutbildning [A sustainable teacher education]. Stockholm: Fritzes Offentliga Publikationer. 50. Gustavsson S. Resistance and meaning in student teachers’ seminar discussion. Dissertation. Göteborg: Studies in Educational Sciences 266, Gothenburg University; 2008. 51. Furu AC. A journey in the landscape of voice. A narrative study of how teachers become professional voice users. Doctoral Thesis. 2011. Faculty of Pedagogics. Åbo Academy, Finland: Åbo Academy University Press. 52. Jensen HH, Mortensen EL, Lotz M. Drop-out from a psychodynamic group psychotherapy outpatient unit. Nord J Psychiatry. 2014;68:594–604. 53. Hagger H, McIntyre D. Learning Teaching from Teachers. Realising the Potential of School-based Teacher Education. Maidenhead: Open University Press; 2006.