Indirect Vs Direct Voice Therapy for Children with Vocal Nodules: a Randomized Clinical Trial

Supplementary Online Content

Hartnick C, Ballif C, De Guzman V, et al. Indirect vs direct voice therapy for children with vocal nodules: a randomized clinical trial. JAMA Otolaryngol Head Neck Surg. Published online December 28, 2017. doi:10.1001/ jamaoto.2017.2618

eAppendix 1. Trial Protocol eAppendix 2. Statistical Analysis Plan eAppendix 3. CONSORT Checklist eAppendix 4. CONSORT Flowchart eTable 1. Eligibility Criteria eTable 2. Description of Treatment Protocols eFigure 1. Pediatric Voice Related Quality of Life (PVRQOL) Form

This supplementary material has been provided by the authors to give readers additional information about their work.

Downloaded From: https://jamanetwork.com/ on 10/02/2021 Online-Only Supplement

1. Trial Protocol 2. Statistical Analysis Plan 3. CONSORT Checklist 4. CONSORT Flowchart 5. eTable 1 and 2 6. eFigure 1

Downloaded From: https://jamanetwork.com/ on 10/02/2021

“A Randomized Controlled Trial on Voice Therapy on Children with Vocal Nodules”

PROTOCOL REVISION 2/2015

1.0 SPECIFIC AIMS AND OBJECTIVES

Voice disorders place over five million school-aged children [2, 3] at increased risk for inferior school performance, dysfunctional social development, and higher participation in criminal activities [4-6]. The lesions most commonly associated with pediatric vocal dysfunction are reported to be vocal fold nodules, which are present in up to 21% of the general pediatric population [7, 8], and are associated with behavioral problems and inferior quality of life [9, 10]. Voice therapy with a speech-language pathologist is recommended by 95% of otolaryngologists for management of vocal fold nodules [11]. When implemented, this voice therapy requires significant time investment and typically results in regular absence from the classroom over a period of 1-3 months or longer, depending on the protocol and the setting [12-19]. Regular absence from school has been shown to place children at risk for poorer scholastic performance [20-22], but the benefits of this voice therapy have yet to be demonstrated in any rigorous prospective, controlled trial. Thus, the most widely used treatment for the most common lesion associated with pediatric voice disorders has not been rigorously tested in a randomized controlled trial to determine whether it may have benefits.

A randomized clinical trial addressing whether voice therapy improves voice-related quality of life for children with apparent vocal fold nodules is a logical step toward developing an evidence-based treatment plan to optimize outcomes for this sizable population of at-risk children. There are currently no well-powered, prospective, controlled studies which compare voice therapy versus control in children who present with apparent nodules. Our goal is to remedy this deficiency through testing of the following primary null hypothesis: There is no difference in the change in PVRQOL scores after 3 months of treatment with voice therapy versus office instructions.

The primary objective of this study is to determine the impact of voice therapy on voice-related quality of life in children age 6-10 years old with apparent vocal fold nodules, as measured by the validated Pediatric Voice-Related Quality of Life Instrument (PVRQOL) [23], ], administered 4 weeks after completion of voice therapy.

Specific Aim #1: To compare the impact of formal voice therapy versus office instructions alone on changes in voice-related quality of life in children with apparent vocal fold nodules.

The secondary objective of this study is to examine the relationship between the primary outcome measure and other measures of voice and vocal fold function. As stated above, PVRQOL scores will be the primary outcome measure. All currently validated pediatric voice-related quality instruments, including the PVRQOL, are designed for parent, rather than child, administration. The parents or caregivers are blinded to which form of vocal instruction their children will receives but are informed that their child will receive one of two forms of voice therapy to compare the efficacy of one in comparison to the other. The rationale for blinding is to prevent any parental bias towards either arm. Since there is currently no standard for voice therapy nor is there information available via the internet to define these standards, parents will have no ability to know pre-hoc whether the therapy their child receives is the control or study arm. Scores from the PVRQOL are also correlated to perceptual measures of voice quality (namely measures derived from Consensus Auditory Perceptual Evaluation of Voice Cape-V stimuli) as well as to change scores in phonation threshold pressure and harmonics-to-noise ratio. Finally, the laryngeal exams of each child are examined pre- and post therapy by each of the two senior laryngologists and assessed for change in appearance of the nodules both by gross characterization as well as by assessment according to a validated grading scale [24].

Downloaded From: https://jamanetwork.com/ on 10/02/2021

Specific Aim #2: To assess the agreement between the primary outcome measure, the parent-indicated PVRQOL score voice quality rating with phonation threshold pressure, harmonics-to-noise ratio, and nodule ratings.

Tertiary objectives are to examine corollary issues around voice therapy for children that may provide insights which can be more fully pursued in future studies. Specific Aim #3: To examine mechanisms for possible benefits shown by voice therapy or the control intervention, including patient demographics, disease characteristics, voicing modality, and patient compliance. An additional aim is to provide data on cost utilization in voice therapy versus the control intervention, in terms of fiscal and temporal costs associated with the interventions.

2.0 BACKGROUND AND SIGNIFICANCE

“Western cultures like ours, it is said, do not suffer the speech handicapped gladly and thus the persons with whom we work come to us with a special kind of human misery.” [25, 26]

2.1 Pediatric dysfunctional voice, which is a common problem associated with inferior quality of life and social development, is most frequently caused by apparent vocal fold nodules.

The prevalence of voice disorders within the pediatric population has been estimated at 11% to 36% [7, 27-30], resulting in over five million affected school-aged children [2, 3] who are at risk for developmental and behavioral problems. These pediatric voice disorders may result in retardation of communication skills, with resultant inability to participate in classroom discussions, as well as disadvantages in developing appropriate social interactions and psychosocial functioning [4]. Furthermore, a persistent voice disorder has been shown to interfere with productivity, self-esteem, and well-being [5]. It has also been shown to increase feelings of anger, sadness, embarrassment, and nervousness, and also raise self-consciousness in children as young as six years of age [31]. In addition, communication disorders are associated with higher crime rates in juveniles [6]. Pediatric voice disorders also have been linked to adult voice disorders which ultimately require further ongoing treatment [32]. Because of this propensity for major developmental ramifications, federal law mandates that services must be provided for children with voice disorders affecting their educational performance (Individuals with Disabilities Act) [33]. By most conservative estimates, at least three million children are referred for such formal intervention because of persistent dysphonia (dysfunctional voice) [34], and nearly one third of speech-language pathologists report that they treat children specifically with voice/resonance disorders [32].

The lesions most commonly associated with pediatric dysphonia are reported to be vocal fold nodules [7, 35, 36], which are implicated in up to 35%-78% of cases [30, 37-40]. These lesions typically result in inferior quality of life and social function [9, 10]. The prevalence of vocal fold nodules has been reported to be as high as 21.6% in males and 11.7% in females in the general population of children 7-16 years of age [7, 8]. Vocal fold nodules are protuberances which occur on the free margin of the true vocal folds, usually at the junction of the anterior and middle thirds (Figure 1). These lesions are best distinguished on stroboscopic exam or microsuspension laryngoscopy, given the inherent difficulties in accurately diagnosing lesions on single-view fiberoptic examination of hesitant children (Also Figure 1. Bilateral vocal fold nodules in see section 4.5.1 Diagnosis of Vocal Fold Nodules). a child. These are protuberant lesions Children with such nodules have significantly worse vocal on the medial surface of the true vocal function than normal children (as measured by voice range profiles and computer-assisted voice analyses) [41, 42]. folds, located at the junction of the Also, the severity of hoarseness, breathiness, straining, and anterior and middle one third of each lack of voice has been associated with the physical extent of

Downloaded From: https://jamanetwork.com/ on 10/02/2021 nodules [43]. In addition, children with vocal fold nodules have significantly inferior quality of life, as compared to normal children, according to prospective evaluation using validated instruments; a substantial impact is seen in total disease-specific quality of life scores, as well as in social/emotional and physical functioning domain scores [9]. Children with nodules also are significantly more likely to act out, have disturbed peer relations, and immature behaviors (according to the Walker Problem Behavior Identification Checklist), as compared to age-matched children without nodules [10].

2.2 Voice therapy, which is a commonly used management option for vocal fold nodules, accepts the opportunity cost of missed school and the potential associated problems of absenteeism to achieve resolution of dysphonia. However, the possible benefits of voice therapy have yet to be shown in any rigorous, randomized, controlled trial.

The most frequently utilized primary treatment for pediatric vocal fold nodules is voice therapy, with 95% of surveyed otolaryngologists choosing this method of treatment first for children [11]. The use of voice therapy is based on the theory that nodules are primarily the result of phonation patterns. The traditional view is that nodules’ origins are principally quantitative: an affected individual has phonated too much, too loudly, with insufficient recovery (i.e. engaged in so-called “vocal abuse”) [4, 29, 44]. However, recent thought emphasizes the notion that phonation quality is at least as important as phonation quantity in pathogenesis. Specifically, “pressed voice” – produced with strong medial compression of the folds -- is thought to increase the risk of nodules due to the link between vocal fold adduction and perpendicular impact stress to the folds[1, 45] and further between impact stress and tissue injury.[46] In both quantitative and qualitative schools of thought, the goal of voice therapy is to address putative pathogenic factors through behavior modification. Consistent with the quantitative view, traditional or “indirect” voice therapy tends to employ “vocal hygiene” education to inform children about voice production mechanisms and injury, and guide them towards the identification and reduction of “vocal abuse” including loud phonation, the imitation of animal and machine noises, hard glottal stops, coughing, yelling across a distance, and similar behaviors[12-14, 16, 18]. This approach is conceptually reasonable. However, recently concerns have been raised about its practicality. Specifically, questions have been raised about how feasible it is to tame a child’s phonatory output, especially where exuberant children most at risk for nodules are involved. The likelihood that therapy can successfully suppress vocal exuberance in the long term seems marginal at best, especially in light of a connection of exuberance to personality traits that tend to endure across the lifespan.[47, 48] Within this latter qualitative camp, the proposal has been that voice therapy should involve direct voice exercises that provide children with tools to produce voice loudly – if they are so predisposed – while limiting phonotraumatic impact stresses that are harmful to the tissue [1, 49]. Specifically, this “direct” approach targets the training of resonant voice, perceptually defined as easy voice involving anterior oral vibrations, which is shown to optimize the relation between voice output intensity (large) and vocal fold impact stress (small).[1, 50]. Stated differently, children who use resonant voice should be able to phonate loudly if they wish, while still remaining relatively protected from injury.

In spite of how commonly it is utilized, the possible impact of voice therapy has not been established in prospective controlled studies, and its actual benefit is controversial. Some voice therapists argue that “screaming is an entirely normal behavior of children; a history of shouting may in fact be less common among children with nodules than among those without,” and strenuously argue against the utility of voice therapy in the pediatric population [25]. In accordance with this opinion, retrospective literature shows that 77% of children with nodules still exhibited abnormal larynges, even after 2 months of voice therapy [51]. Against this opinion, however, prospective uncontrolled studies of voice therapy in children have shown that 68% of children have improved vocal function after 15 sessions of voice therapy [17]. Surveys confirm that opinion is mixed among those who actually give voice therapy, with only 48% of speech- language pathologists reporting that voice therapy was the preferable treatment for children with nodules [11].

Voice therapy is time-consuming and has a clear opportunity cost. Many clinicians are reluctant to provide voice therapy in the school setting due to lack of confidence in their abilities in voice therapy, uncertainty over how therapy can be justified vis-à-vis the Individuals with Disabilities Education Act (IDEA), and other factors [2]. Thus, it is likely that the predominance of voice therapy for children is provided in the outpatient setting. As a result, therapy is often associated with missed school and missed work for

Downloaded From: https://jamanetwork.com/ on 10/02/2021 parents. Formal data are not available. However, anecdotal reports indicate students may miss as many as 75 classes over a 12-mo academic period for therapy in the school setting, and up to 15 half-days of school (and thus an estimated 60 classes) over a 15-week period for outpatient therapy (American Speech-Language-Hearing Association Special Interest Division 3 Voice and Voice Disorders listserve survey, conducted by K. Verdolini, January 28 – February 1, 2009). Thus, voice therapy has potential deleterious effects in the context of having unclear benefits. Furthermore, the time investment itself is a significant burden, as it requires missed school and the associated problems of absenteeism. Absence from school for more than 10% of the time is associated with lower academic performance in children as young as those in kindergarten through third grade, even if absences are excused.[22] School absenteeism is also a key risk factor for high risk behaviors, psychiatric disorders, and economic deprivation [21, 22] and has even been touted as a proxy measure of child health status [52]. Thus, although voice therapy is the most frequently utilized treatment, there are no prospective controlled data that have established that its benefits outweigh the associated problems of regular school absenteeism.

2.3 A next step toward developing an evidence-based treatment plan for children with vocal fold nodules is to perform a randomized clinical trial evaluating the impact of voice therapy on voice-related quality of life.

Pediatric otolaryngologists use a variety of combinations of first line management options for vocal fold nodules, with voice therapy being used most commonly [53]. Our goal is to perform a randomized clinical trial to determine whether there are true benefits that outweigh the risks for this commonly used treatment (voice therapy), as there are no well-powered, prospective, controlled studies which do so.

The outcome of most interest for children affected by apparent vocal fold nodules is voice-related quality of life, as this measure encompasses the impact of the associated dysphonia on daily activities. Such voice-related quality of life can be quantified using a disease-specific, validated instrument such as the Pediatric Voice-Related Quality of Life Instrument (PVRQOL) [23, 54]. This instrument measures the impact of voice on daily activities, emotions, and social behavior. Post-treatment changes in voice have been demonstrated for other pediatric voice disorders using the PVRQOL, which has proven convergent and discriminant validity, reliability, and responsiveness [23, 54-56]. Given these properties, scores for this instrument are the primary outcome for this study.

The PVRQOL was chosen over other disease-specific options for measuring voice-related quality of life because it has previously established child-specific, normative, and nodule-specific characteristics. In addition to the instrument’s initial validation, normative data have been established according to age (2 to 17 years) and gender [57]. Also, results in common laryngeal disorders such as nodules have been published [9]. Other options for validated instruments that evaluate pediatric voice include the Pediatric Voice Handicap Index and the Pediatric Voice Outcomes Survey. The first potential alternative, the Pediatric Voice Handicap Index, is a 23-item parent proxy survey (all current pediatric voice related quality of life instruments have been designed for parent proxy administration) that has been validated for patients age 4-21 years old [58]. This index was validated through study of normal children and those with subglottic stenosis from prolonged intubation. This Index was not chosen because nodule-specific data have not been published, and because the higher item count represents a greater burden to the parent in the study. The second potential alternative, the Pediatric Voice Outcome Survey is a 5-item survey that was validated for study in children age 2-18 years of age [59], using a population of patients who had undergone tracheotomy in order to establish discriminant validity. This survey was not chosen for use in this study because: 1) nodule-specific data have not been published, and 2) social, emotional, and physical impact of dysphonia are not as specifically addressed (as they are in the PVRQOL). There are also a number of other instruments for voice evaluation that have been validated in adults. For example, the Glottal Function Index is a 4-item, self-administered symptom evaluation with good reliability, reproducibility, criterion-based and construct validity, but it is only validated in adults [60]. Likewise, the Voice Handicap Index, the Voice-Related Quality of Life scale, the Voice Activity and Participation Profile, the Voice Symptom Scale, the Patient Questionnaire of Vocal Performance, and the Voice Outcome Survey have been validated in adults, and have 5-30 items that focus on the range from functional to

Downloaded From: https://jamanetwork.com/ on 10/02/2021 physical to social to emotion symptomatology [61]. Since our study focuses on pediatric patients, none of these adult-validated instruments were preferred.

The PVRQOL was also chosen in order to measure disease-specific quality of life, rather than global quality of life. Although there are rigorously validated instruments to measure pediatric global health, such as the Child Health Questionnaire (a 50 item instrument) and its abbreviated counterpart (a 28-item instrument) [62], the focus of our inquiry is dysphonia and its consequences. Given the specific voice- related focus of our inquiry, a disease-specific measure was felt to be more appropriate than a global measure of quality of life.

2.4 Other Measures to Assess Function to Correlate with Voice Related Quality of Life

2.4.1 Perceptual Measures of Voice Quality The sound of the voice is one of the chief concerns that brings patients to voice disorder clinics. In the case of children, the concerns may be raised primarily by the parent or a physician, who notice something wrong with the child’s voice and therefore seek attention for it. As such, voice quality is one of the central parameters relevant to evaluating the success of a voice intervention. Unfortunately, perceptual measures of voice quality are subject to a series of concerns themselves, in particular those about measurement stability within and across raters. Numerous attempts have been made to offset such concerns by identifying instrumented correlates of voice quality, such as perturbation measures, signal-to- noise ratios, etc. The idea is that such measures should improve on the reliability of perceptual measures, and in some cases, they do. However, the problem is that no single instrumented measure and no set of instrumented measures has yet been identified that captures voice quality globally. Thus, measurement reliability may be improved with an instrumented approach, but improvements may come at the expense of validity. As a result, considerable efforts over the past years have been directed not so much towards the substitution of perceptual measures of voice quality in the evaluation of treatment success, but rather towards the development of methods to improve their reliability.

Efforts in this direction were undertaken by a special task force appointed by the American Speech- Language-Hearing Association (ASHA). The result was the generation of the “Consensus Auditory- Perceptual Evaluation of Voice” (CAPE-V) tool[63], which is currently undergoing evaluation by ASHA for its potential inclusion as preferred practice in ASHA guidelines for the perceptual evaluation of voice. The CAPE-V specifies standardized procedures for the perceptual evaluation of voice, including standardized recording conditions, word stimuli and rating procedures. Thus, some of the sources of variability inherent in voice quality ratings can be contained. However, the authors of the CAPE-V are also clear that problems with measurement stability will persist despite these attempts.[63] The reason is that in addition to differences in stimulus content and recording quality, several other sources of variability exist in perceptual ratings of voice quality. Among them are listeners’ varying internal standards for ratings, and listeners’ difficulty isolating different perceptual attributes of a voice.[64-66]

One interesting approach to address the first problem has been the use of “analysis-to-synthesis” techniques.[64] In that approach, a rater listens to a recording of a patient’s voice, and adjusts selected acoustic parameters of a synthesized voice until the real and synthesized voices are perceptually matched. The acoustic parameters selected in the matching process are then noted and used as indicators of voice quality. The basis for this quite ingenious approach is that it moves listeners’ rating standards from a personal idiosyncratic space to a “public” space that can be shared by all listeners. Results show that listener agreement is sharply increased by this approach in comparison to results using typical visual-perceptual rating procedures.[64] . However, the approach is not yet ready for clinical use or use in clinical studies. Specifically, the approach has been investigated for its utility in measurements of restricted aspects of voice quality (e.g. breathiness) during sustained vowels rather than voice quality as a whole for more complex utterances that are of clinical interest.[64]

Awaiting further advances in analysis-to-synthesis, a parallel approach has been developed by the same investigator group.[67] This approach involves paired comparisons. Specifically, raters listen to two recorded samples of identical utterances produced by a given individual at different times, for example before and after therapy. The listener then uses a forced choice paradigm to declare which of the

Downloaded From: https://jamanetwork.com/ on 10/02/2021 samples is “better” – e.g., in terms of “global voice quality.” The listener then indicates on a scale from 1-5 how confident he/she is of the decision (1=”positive;” 5=”wild guess”). Using this approach, the referent models for determination of voice quality are again moved “outside the head” to a common space available to all listeners (2 external samples to be compared rather than comparing 1 external sample to a personal internal standard). Moreover, the speaker also serves as his/her own control in terms of speech patterns and other speaker-specific idiosyncracies not relevant to the perceptual task at hand. As such, the perceptual attributes of interest can be moved to the perceptual forefront. Simple computational procedures described shortly are then applied to the results to obtain quantitative estimates of voice quality for each sample, for example samples before and after therapy.[67] Work by Kreiman and associates has indicated that this general approach more than doubles the likelihood that listeners will agree exactly in their voice quality assessments compared to typical visual-analogue scale procedures involving ratings of single voice samples at a time, because the approach controls for both listener referents and perceptual complexity of the rated samples [67]. Details regarding our experience with this approach are described in Preliminary Studies, which provide the basis for the approach we will use to rate voice quality in the present study.

2.4.2 Assessment of Nodules Many clinicians hold that laryngeal status is also a relevant parameter in the evaluation of voice therapy success. Consideration was given to making an assessment of nodule severity the primary outcome variable for this study. However, measurement questions are again posed. How is “nodule severity” to be measured? Potentially, severity could be perceptually evaluated, using procedures similar to those just described for voice quality. Unfortunately, our preliminary work reveals that rater agreements are considerably poorer for such ratings than for auditory ratings of voice quality, using this approach (see Preliminary Studies). Another approach is shown to be more promising. A grading system for vocal fold nodules has been validated with established inter-rater reliability for nodule presence and size [23]. Specifically, a 3-point grading scale has been established for nodules size, with grade 1 being small, grade 2 being moderate, and grade 3 being large. The contour of the nodule is also rated as either discrete or sessile (Appendix 3). The objective outcome of nodule grade (size) is an attractive metric for treatment outcomes. As such, nodule severity might indeed be considered a reasonable primary outcome for study, using this approach. However, there are data that have established that nodule size or resolution do not strictly correlate with improvements in voice [96,97]. In fact, patients may have resolution to normal voice, even in the presence of residual nodules. Furthermore, nodule size does not appear to correspond to the level of vocal dysfunction. [102] Thus, although it is a concrete physical measurement, it is still more appropriately chosen as a secondary outcome, rather than a primary outcome. Additional secondary outcomes will involve measures of voice quality already described, as well as instrumented measures of vocal function, described next.

2.4.3 Instrumented Measures of Vocal Function Selected instrumented measures are also relevant as potential indicators of therapy effects. As discussed, instrumented measures generally fail to capture fully integrated, global impressions about voice. However, they can sometimes point to discrete aspects of laryngeal performance that are illuminating. Moreover, many instrumented measures have greater reliability than perceptual ones. As such, they are appropriate to consider. Numerous measures are possible candidates, including acoustic, aerodynamic, and physiological measures. One approach would be to collect data using a large number of measures, in the spirit of exploring which ones might be most sensitive to change with treatment. The problem with that approach is that it lacks conceptual clarity and vastly increases the potential for alpha inflation in the overall data set (the likelihood of statistically detecting effects that are in fact not present). For the present protocol, we seek measures that are conceptually targeted to the types of changes we would expect to see with effective treatment for nodules, ideally with the greatest potential to reflect global aspects of laryngeal functioning. Phonation Threshold Pressure (PTP) and Harmonics-to-Noise Ratio (HNR) are among the most appealing measures in that regard. PTP reflects the minimum subglottal pressure required to initiate and sustain vocal fold oscillation. Both theoretically and empirically, PTP is sensitive to numerous aspects of laryngeal mucosal function, including (a) vocal fold thickness; (b) vocal fold viscosity; (c) the speed of mucosal wave propagation (related to tissue stiffness); and (d) pre- phonatory (and phonatory) intercordal distance[68, 69]. Of these parameters, tissue viscosity, tissue stiffness, and degree of inter-vocal fold gap are predicted to change with improvements in vocal fold

Downloaded From: https://jamanetwork.com/ on 10/02/2021 nodules. Thus, PTP is an attractive aerodynamic measure to target in our study. Among acoustic measures, arguably, the Harmonics-to-Noise Ratio (HNR) has appeal because it, too, reflects the vocal folds’ general ability to convert pulmonary air into systematic sound waves, and has been shown to be a valid indicator of laryngeal pathology (e.g. Shama, Krishna, & Cholayya, 2007[70]). Thus, in the interest of containing and focusing the corollary questions in the present study, we will use PTP and the HNR as secondary indicators of laryngeal state and function.

The ultimate goal is to perform a rigorous, prospective, well-powered clinical trial to determine whether outcome is better with either voice therapy versus office instructions alone. The results of this trial will provide initial data required to guide evidence-based practice for children with apparent vocal fold nodules. Results will also be useful as a platform for future trials that will expand the subject base and explicitly examine mechanisms that may mediate treatment benefits as well as factors predicting therapy success.

3.0 PRELIMINARY STUDIES

3.1 Validation of the Pediatric Voice Related Quality of Life Instrument (PVRQOL) The P.I. (CJH) at the coordinating center (MEEI) developed and validated the instrument used for the primary outcome measure for this study [23]. The PVRQOL instrument consists of 10 items completed by parent proxy administration (see Appendix 1). On average, it takes 5 minutes to complete [57]. The instrument proved reliable (Cronbach’s alpha value 0.96, p<0.001; weighted kappa 0.8), and had convergent and divergent validity[71]. One of the co-investigators on this grant(JK) and his research team evaluated a subset of 95 children with “healthy and normal” voices and found their PVRQOL scores to be 96.8 +/- 5.85 and then evaluated 13 children with vocal fold nodules and found their mean PVRQOL score to be 84.8 +/- 9.4[9]. The Co-investigator (JEK) additionally employed this instrument to assess PVRQOL in children with paradoxical vocal fold dysfunction, and to assess the impact of gender and age on PVRQOL scores.[9, 57, 72] Both the research team of the principal investigator (CJH) and that of the co-investigator then looked at a preliminary group of children with significant hoarseness enough to warrant voice therapy in order to obtain pilot data regarding what mean change score would be demonstrated. Collectively, 17 children age five to ten years underwent voice therapy. The mean pre and post therapy PVRQOL scores were 69.4 (SD 12.4) and 92.0 (SD 4.) with a mean change of 22.8 (SD 11.7). Close evaluation of these scores in conjunction with the assessments of the [parents and self- assessments by the children of their voices and when they felt a change had occurred suggested mean a 12 point change was determined to be the minimum change noticeable to the parents.

Furthermore, the coordinating center has additional experience with pediatric instrument use, including validation and subsequent studies using the Pediatric Voice Outcome Survey [59, 71], the Pediatric Cough Questionnaire [73], and the Pediatric Tracheotomy Health Status Instrument.[74] Thus, the coordinating center has significant experience with the utilization of disease-related quality of life instruments.

3.2 Data Management Center Experience The Cooperative Studies Program Coordinating Center (CSPCC) at the VA Hines provides biostatistical, data processing, project management and site management support for multi-center randomized clinical trials. It has been in existence since 1972 and has supported studies sponsored by the Department of Veterans Affairs, the National Institutes of Health and the private sector. The center director is Dr. Domenic Reda, who has more than 30 years of experience in the design, conduct and analysis of clinical trials and is a Fellow of the American Statistical Association. Staff size is 40 and includes biostatisticians, computer programmers, data coordinators, project managers, forms designers, data entry staff and an administrative core. The center is able to support studies outside the Department of Veterans Affairs through its affiliation with the CARES non-profit research foundation.

CSPCC has broad collaborative experience working with investigators in a variety of medical specialties. Specific to this study, CSPCC was the coordinating center for the Cochlear Implant Study,[81] the Hearing Aid study,[82] and the Hearing Aid follow-up study.[83] The center also supported the recently

Downloaded From: https://jamanetwork.com/ on 10/02/2021 completed Electronic Digital Noise Reduction Hearing Aid Trial (not yet published) and the ongoing NIDCD-funded Sudden Sensorineural Hearing Loss Trial.

4.0 STUDY DESIGN

4.1 Overview The current study is a randomized controlled multi-center trial of children with vocal fold nodules in two treatment arms: 1) voice therapy, 2) office instructions alone. Subjects in the voice therapy sessions arm receive 8 sessions for up to 12 weeks with a certified speech therapist, with the goal of implementing techniques of vocal hygiene and vocal resonance training [84]. Subjects in the office instructions alone (no voice therapy sessions) arm receive standardized verbal and written office instructions to discourage so-called “vocal abuse” during 6 sessions for up to 12 weeks. Boys and girls age 6-10 years of age with vocal fold nodules who satisfy eligibility criteria (please see section 4.3) will participate. The rationale for the 6-10-yr age range is threefold. First, children younger than age 6 are excluded in order to study only children who are in school full-time, and thus have similar educational and social exposures. Second, children older than age 10 are excluded to avoid patent changes in the larynx and laryngeal function that come with puberty, especially in boys. Third, the 6-10-yr age range roughly maps onto a proposed Piagetian developmental stage in which general cognitive capabilities are proposed to be fairly homogeneous.[85] Moreover, our clinical experience is that children in this age range can cooperate in laryngeal examinations and voice therapy. Of course, children across the identified range will demonstrate some degree of physical and cognitive heterogeneity. However, they represent the midrange of childhood, which is the target in this study. Further narrowing of the age range for inclusion would restrict the subject pool and as important, decrease the ability to generalize the results. Randomization is stratified according to participating study site. Three other factors were considered for stratification: baseline PVRQOL score, age, and gender. We chose two of these measures, age and gender, to additionally stratify by because our experience indicates these two are more closely related to change scores seen after therapy.

In order to minimize regression to the mean and the possibility we might randomize children with a transient voice quality problem, we will perform two baseline measurements of the PVRQOL before randomization, one at the time of initial screen and the next just prior to randomization and actual enrollment into the study itself. The parent or guardian who completes the initial PVRQOL must complete the survey at each data collection point for intra rater reliability. The final baseline PVRQOL will be used for comparison against later scores. Other baseline data (that will be recorded once before the beginning of the study) include demographic features and characteristics of voice use, as well as recordings to be used for voice quality ratings, phonation threshold pressure measures, harmonics-to- noise ratio, and nodule ratings. Follow up measurements of PVRQOL and other variables discussed shortly are targeted for immediately after termination of the interventions (ie. post therapy or control), and again 4 weeks after termination. A word of clarification is that the goal is indeed to complete the interventions over an 8 week period. However, prior experience is that there is great likelihood that some appointments during that window may be canceled due to child, parent, or clinician illness, vacation, and other factors. Thus, we plan for a “grace period” of up to 4 weeks, which brings the treatment period to up to 12 weeks, for children to complete the protocol, by which time we anticipate most children not involved in attrition will have received the full treatment. If interventions are completed within the 8-week target, follow-up measures will be made 8 and 12 weeks after baseline, or 12 and 16 weeks after baseline allowing for the grace period. For subjects who succumb to definitive attrition during the intervention period, follow-up data will be sought nonetheless and analyzed as “intent to treat” (section 4.10) Primary data analysis focuses on PVRQOL change scores at 16 weeks. Data for secondary outcome measures including voice quality and laryngeal appearance, Phonation Threshold Pressure and Harmonics to Noise Ratio measures will also be obtained from recordings made at baseline, 8 (or 12) weeks, and 12 (or 16) weeks. Finally, data on patient compliance with attendance and home instructions will be collected at every therapy session for both study arms. .

PLEASE NOTE: Circumstances arise in which subjects may not be able to complete treatment sessions and the follow-up visits within 16 weeks. Therefore, subjects may complete the study within 20 weeks from the day of their first treatment session, regardless of the treatment arm. However, this is only for

Downloaded From: https://jamanetwork.com/ on 10/02/2021 extenuating circumstances and on an as needed basis. All efforts will be made to complete treatment and follow-up within 16 weeks. (3/2013)

Data analysis is performed as described in section 4.10 Study subjects are enrolled at three centers: Massachusetts Eye and Ear Infirmary, Drexel University College of Medicine, and Medical College of Wisconsin. A fourth site (The Hospital for Sick Children) was added to aid in recruitment efforts (02/2012).

4.2 Power analysis and sample size calculations

4.2.1 Hypothesized treatment effect Our estimate of the expected treatment effect is based upon data from a pilot study of children both at MEEI and at Milwaukee who were evaluated for hoarseness, found to have vocal nodules, and underwent a course of voice therapy. Pre- and post-treatment PVRQOL scores were correlated with parental perception of clinical impression of vocal improvement. For a group of 17 with voice change noted, the mean change in PVRQOL was 22.8 with standard deviation 11.7. Because the proposed study is multicenter and randomized with a control group that will likely show some improvement over time we expect the difference in mean change in PVRQOL between the two treatment groups will be smaller. The sample size will be based on expecting an 8-point difference in PVRQOL change scores, which is approximately 1/3 the change associated with voice therapy found in the observational pilot study described above. The same pilot data also suggested that a 12-point change for an individual was clinically meaningful, which represents a difference in which one group on average continues to have a problem that is “as bad as it can be” while another group improves to have only a “small problem” in at least one domain. Generally, average group changes are smaller than individual changes. Therefore, the study is powered to detect an 8-point difference. In addition, data will follow a normal distribution [23, 54] based on initial data from validation and prior outcomes measurement with the PVRQOL instrument.

4.2.2 Sample size calculation For a mean difference in change scores of 8, standard deviation of the change 11.7, t-test for two independent groups as the statistical test, two-sided alpha of 0.05 and power of 0.90 the required number to be randomized is 46 per group, bringing total sample size to 92. The rate of withdrawal from the study during follow-up is anticipated to be less than 20%, given the relatively short time frame and number of follow up visits, and minimal burden of completing the 10 question PVRQOL instrument. Adjusting the sample size accordingly, 116 randomized participants, or 58 per group, will be needed (See Table 2).

Table 2. Total Sample Size Calculations with Two-Sided Alpha = 0.05, Power = 0.90, STD = 11.7 Under Various Assumptions.

Drop-out Rate (%) Difference between treatment and control group Total Sample Size 0.0 8 92 10 60 12 42 0.10 8 102 10 68 12 48 0.20 8 116 10 76 12 52 0.30 8 132 10 86 12 60

Downloaded From: https://jamanetwork.com/ on 10/02/2021 Due to lower than expected drop out rate (12%) as of March 31, 2014, the sample size has been adjusted to 106 subjects to establish the same mean difference of 8, standard deviation of the change 11.7, t-test for two independent groups as the statistical test, two-sided alpha of 0.05 and power of 0.90 (04/2014). 2/2105: As of February 2015, the withdrawal rate has increased. Therefore, the sample size has been readjusted to 114 subjects to establish the same mean difference of 8, standard deviation of the change 11.7, t-test for two independent groups as the statistical test, two-sided alpha of 0.05 and power of 0.90

4.3 Recruitment Plan Between the three clinical sites, based on review of patient databases over the past 24 months, 200 patients between the ages of 6 to years of age are seen on an annual basis and roughly two-thirds of those would be eligible for study inclusion. These children with vocal nodules and dysphonia are seen in the three recruiting centers “voice” practices as part of the normal clinical volume of patients presenting for treatment. Utilizing a rather conservative enrollment rate of 25% of eligible participants, which has been achieved in other clinical trials at each site, this will allow for 30-35 patients to be enrolled on a yearly basis.

Due to lower than expected recruitment in the first 6 months of the recruitment period, a new site has been added: The Hospital For Sick Children. The addition of the site will not affect the sample size nor the number of subjects expected to be recruited on a yearly basis. (02/2012)

4.4 Study Flow Study flow progresses according to the four stages of the trial – enrollment, baseline/allocation to intervention, follow up, and analysis – as delineated in the Consolidated Standards of Reporting Trials Statement [86], which has been shown to ultimately improve reports of randomized controlled trials in systematic reviews of the literature [87]. The number of participants and non-participants is determined at each stage. See Flow Charts in Appendix Ab for details.

4.5 Eligibility Criteria

4.5.1 Diagnosis of Vocal Fold Nodules The study of vocal fold nodules in children raises a diagnostic challenge, as it is often difficult to distinguish vocal fold nodules from other lesions on a single-view exam of a small moving target. Vocal fold nodules may be diagnosed, when in fact cysts, polyps, granulomas, or other distinct mass lesions are actually present. Fortunately, current technology such as videography, stroboscopy, and microscopic techniques allow increasingly accurate diagnosis of these lesions. Videography with digital recording allows for repeat review of endoscopic exams, which allows confirmation of findings, as well as independent reviews of the same exam. Stroboscopy allows for very precise evaluation of vocal fold mobility, mucosal wave, and mass lesions, and has been used successfully in children as young as 3 years of age [88]. Video-stroboscopy is the method utilized for diagnostic purposes during evaluation for enrollment into this study. Video-stroboscopy in children is performed via transnasal flexible fiberoptic laryngeal examination or via trans-oral technique depending upon the experience and preference of the examining laryngologist as well as that of the patient in terms of his or her being able to comply with the procedure [88]. Given that the differential diagnosis for vocal fold nodules is broad, and that diagnosis is based on the gross appearance of the lesion (histopathologic analysis of a biopsy specimen is typically not performed because of its benign nature and small size), each subject considered for enrollment will have the diagnosis of vocal fold nodules confirmed by two senior otolaryngologists’ (CJH &JK) video- stroboscopic examinations.[24] Diagnostic inclusion and exclusion criteria are as described in sections 4.3.2 and 4.3.3.

Operative microsuspension laryngoscopy is another means to obtain high quality diagnostic evaluations, but requires acceptance of general anesthesia and the associated risks. Nonetheless, some children undergo microscopic exam due to evaluation of another lesion, prior to presenting to a clinical setting where pediatric videostroboscopy is available. In these rare circumstances, these evaluations are

Downloaded From: https://jamanetwork.com/ on 10/02/2021 typically diagnostic; thus if a child presents with photodocumentation of operative microsuspension examination within 4 weeks of enrollment into the study, this diagnostic method is also considered acceptable. In other circumstances, children may be unable to tolerate office-based awake videostroboscopy, and the parents may feel that the diagnostic information gained by a planned examination under anesthesia is valuable given the severity of symptoms. If operative microsuspension laryngoscopy photodocumentation has been obtained for either reason, the diagnosis is also corroborated by thetwo evaluations by senior otolaryngologists ( CJH & JK), in order to confirm eligibility for enrollment. Subject selection criteria and mode of ascertainment is listed in Table 3.

4.5.2 Inclusion criteria 1. Eligibility is contingent on the presence of apparent vocal fold nodules, which are defined as bilateral, localized, benign, superficial growths with protrusion on the medial surface of the true vocal folds at the junction of their anterior and middle thirds. This examination is based on video-stroboscopic examination, or, in rare cases, on operative microsuspension laryngoscopy. In addition, strict exclusion criteria are applied toward excluding masses that are not nodules. Defining characteristics for these inclusion criteria are based upon the evaluations of twosenior otolaryngologists of each participant’s exam. Children evaluated by the senior otolaryngologist at each institution suspected to have vocal fold nodules are candidates for inclusion to this study. Patient de-indentified images of the stroboscopy are reviewed by each of the other two laryngologists; only those children with confirmed vocal fold nodules by the all two senior otolaryngologists will be considered for entry into this study. 2. Children 6 to10 years of age will be enrolled. The rationale for this age range was previously noted. In brief, children in this age range (a) have similar educational and social exposures; (b) will likely not yet encounter pubertal changes affecting the larynx; and (c) are relatively cognitively homogeneous. Moreover, children in this age range have the ability to cooperate with voice therapy and have also been shown to be tolerant of stroboscopic examination. Finally, the voice therapy regimen in this protocol has been shown to be clinically effective for many children in this age range. 3. Voice-related quality of life must be affected to the extent that baseline PVRQOL scores are <87.5 (on a scale of the worst, 0, to the best, 100) at the time of entry into the trial. Based on previously published data, this subset of scores will be clearly distinct from scores in children with normal voices. In addition, scores <87.5 represent worse than average scores among children diagnosed with vocal fold nodules. 4. Dysphonia duration prior to randomization must be at least 12 weeks, in order to ensure that vocal dysfunction is chronic in nature. 5. Hearing in better ear of 35 dB or better. 6. Agreement by informed consent from the parents and informed assent from the child participant with anticipated commitment to compliance throughout the follow up period of 3 months is necessary for enrollment which includes time commitment of up to 3 hours per week to therapy sessions and homework. 7. Vocal fold nodules are a pathology that predominantly affects males. A number of studies and the databases utilized to track this pathology at the clinical sites indicate that the male: female ratio is approximately 7:3. Given this background, the enrollment plan for the current investigation will seek to have a 30% female representation of the patients enrolled in this study to ensure appropriate gender representation. Additionally, patients with a diagnosis of pediatric vocal fold nodules predominantly are found in Caucasian populations. However, in accordance with the National Institute of Health Revitalization Act of 1993 the inclusion of minorities in this proposal will be targeted to the representation of minorities in the greater Boston, Philadelphia and Milwaukee metropolitan regions.

4.5.3 Exclusion criteria 1. Children with non-standard nodules or masses that are not unequivocally nodules are excluded, including those with irregular epithelium, concern for malignancy requiring biopsy, rapid growth pattern, airway obstruction requiring urgent or operative intervention, and prior laryngeal surgery. Furthermore, since there are several lesions that can present as bilateral true vocal fold masses (such as unilateral cysts or polyps with contralateral fibrovascular

Downloaded From: https://jamanetwork.com/ on 10/02/2021 reaction, contact granulomas, and Reinke’s edema), masses with the following characteristics will be excluded: a) yellow or blue color suggestive of a mucous- or fluid-filled component, b) pink friable lesions at the posterior aspect of the true vocal folds, c) watery edematous expansion without a distinct protrusive mass. Defining characteristics for these exclusion criteria are determined and corroborated by the evaluations ofttwo senior otolaryngologists of each participant’s videostroboscopic exam. This examination is based on video-stroboscopy, or, in rare cases, on operative microsuspension laryngoscopy. The video images will be placed on a password-protected server to which all two senior investigators will have access; these video images will not contain any patient identifying information but rather a study number that will be assigned to each participant. In addition, strict inclusion criteria are applied toward excluding masses that are not nodules. 2. Children who have previously received any form of speech therapy for longer than 2 weeks targeting voice and/or resonance with the exception of articulation or speech therapy. 3. Children for whom voice therapy protocols will be problematic are excluded: developmental delay, cognitive disorder, behavioral disorder, neurologic disorder, articulation disorder, phonologoical disorder, language disorder, expressive and/or receptive language delay, specific language impairment, central auditory processing disorder, fluency disorder, prior knowledge of voice therapy principles through previous interventions (with the exception of articulation or speech therapy), and inability to commit to at least 12 weeks of therapy as determined and evaluated by the treating speech language pathologist and confirmed by consensus of all two PIs (CJH and JK). 4. Children whose caregivers cannot complete the primary endpoint are excluded: non-English speaking as the PVRQOL instrument is administered in English, and those unable to commit to the 3-month treatment and follow-up schedule. 5. Children whose caregivers are not willing to commit up to 3 hours per week to vocal therapy including therapy sessions and homework for up to 12 weeks. 6. Children who cannot tolerate a conscious video-stroboscopic examination and whose parents feel that the diagnostic information gained by a planned examination under anesthesia is not valuable for the severity of presenting symptoms are excluded. 7. Children with significant confounders of voice-related quality of life are excluded: vocal fold paralysis, neurologic disorder of the larynx (dystonias, tics, tremors, etc), ongoing acute upper respiratory tract infection (defined by at least 2 of the following: sneezing, coughing, nasal congestion, runny nose, or temperature greater than 100.4° F (38.0° C), and untreated or unsuccessfully treated allergic rhinitis/post nasal drip.

Table 3. Participant Selection and Mode of Assessment Criteria Obtained By 6-10 years of age Parent report

Presence of vocal fold nodules Otolaryngologist video-fiberoptic exam

Voice related quality of life <87.5 Parent/Guardian assessment Dysphonia duration >12 weeks Parent report

Voice therapy for longer than 2 weeks prior to entry Speech-language pathologist report

Anticipated compliance Initial encounter, parent report Informed consent Parent compliance Concern for malignancy requiring biopsy Otolaryngologist assessment Airway obstruction Otolaryngologist assessment Irregular epithelium Otolaryngologist assessment

Downloaded From: https://jamanetwork.com/ on 10/02/2021 Prior laryngeal surgery Parent report, otolaryngologist assessment Laryngeal anatomy unfavorable to visualization via Parent report, pediatrician report, otolaryngologist laryngoscopy assessment Vocal fold paralysis Otolaryngologist assessment Neurologic disorder of the larynx (dystonias, tics, Otolaryngologist, speech-language pathologist tremors, etc) assessment Acute upper respiratory tract infection: defined by at Parent report, otolaryngologist assessment least 2 of the following: sneezing, coughing, nasal congestion, runny nose, or temperature greater than 100.4° F (38.0° C)

Untreated or unsuccessfully treated allergic Otolaryngologist, allergist assessment rhinitis/post-nasal drip.

Non-English speaking Parent report, investigator assessment Hearing in good ear better than 35 dB Investigator assessment

Inability to commit to the 3 month follow up schedule Parent report

Unwillingness to commit up to 3 hours per work for Parent report vocal therapy for up to 12 weeks

4.6 Baseline and Follow-Up Measurements We will perform two baseline PVRQOL measurements before randomization, one at the time of initial screen and the next just prior to randomization and actual enrollment into the study itself. The last measure taken will be used for comparison against post-intervention data. If either of the initial PVRQOL scores is greater than 87.5, the child will not be entered into the study as there is too little room to show improvement. Data collected just prior to enrollment will be used for data analysis. Baseline measurements include: age, sex, study center, initial PVQROL scores, duration of dysphonia prior to enrollment, presence/absence of putatively vocally harmful habits, and extent of school absenteeism.. Also data collection for associated measures of vocal function will be completed, including measures of nodule size (grade and contour), nodule description, voice quality, degree of resonant voice, phonation threshold pressure, and harmonics-to-noise ratio. All measures that could possibly change over the course of the protocol will be repeated at 8- and 12-week or 12 and 16 follow-up, as will data regarding patient compliance and school absenteeism. Obviously, baseline data that will not change (e.g. sex, study center, etc.) will not be re-collected. Finally, time and money cost analyses for the interventions will be completed at the conclusion of the study. Operational definitions of measures and procedures for data analysis are provided in the following section (4.6.1).

4.6.1 Operational Definitions of Measures and Procedures 1. Age is measured as the number of years and months since birth at the time of enrollment in the study. 2. Sex is recorded as male or female. 3. Study center is recorded as the site of enrollment and follow up (Massachusetts Eye and Ear Infirmary, Drexel University College of Medicine, Medical College of Wisconsin, and the Hospital for Sick Children) 4. Initial PVRQOL scores are measured via administration of this validated instrument to the caregiver (See section 4.6) 5. Nodule size (grade and contour) can be measured through a validated grading system [24] with proven inter-rater reliability. This grading system quantitatively addresses the objective presence, absence, and extent of this disease process (section 4.9.2). 6. Harmful vocal habits are assessed as parental report of the presence or absence of: a) excessive speech production, b) routine loud speech (i.e., as if to people with hearing loss or over background noise), c) frequent shouting/ yelling/ screaming, d) chronic cough and/or

Downloaded From: https://jamanetwork.com/ on 10/02/2021 frequent throat clearing, as reported by the parent [4, 29, 44]. The presence of any one of these habits will define the overall presence of harmful vocal habits. If none of these habits are present, however, then the subject will be noted to have an absence of harmful vocal habits. 7. Associated voice measures will be obtained at all enrolling centers. These measures will be made as secondary measures to assess the effects of voice therapy and control interventions for children with vocal fold nodules. Measures will include perceptual ratings of voice quality and of degree of vocal “resonance,” nodule grade and contour, phonation threshold pressure, and harmonics-to-noise ratio (section 4.9). Also data regarding patient compliance with the interventions will be collected at each therapy session for both study arms (section 4.8.2 and 4.8.3).

4.7 Randomization If the caregivers have provided informed consent, the participant has given assent, meets all inclusion criteria, meets none of the exclusion criteria, and has obtained consensus among all two senior otolaryngologists for the presence of vocal fold nodules, then the participant is eligible for randomization into either the “Adventures in Voice” - Therapy Group (Program Y) or “My Voice Adventure” - Hygiene Group (Program X). Assignment to either group is equally likely and possible. It is anticipated that at the end of the trial, there will be approximately the same number of participants in each group. The randomization process uses a web-based Electronic Data Capture (EDC) system with a backup telephone system. Permuted block randomization with random block sizes is used. Randomization is stratified by participating site, age (6-7 year old vs. 8-10 years old) and gender. The system requires input of certain baseline measurements and verification of certain inclusion and exclusion criteria as a quality control step.

4.8 Interventions

4.8.1 General Instructions and Specifications for Both Groups 1. An English-speaking parent or guardian should accompany the child to the baseline, 8 (or 12)-week, and 12 (or 16)-week visits. 2. No diet instructions will be specified for either group. 3. No subject will be allowed to switch from one group to the other during the protocol. Every attempt will be made to guard against drop out for whatever reason. However, drop out may occur for reasons beyond our control and is accounted for in sample size calculations.

4.8.2 Voice Therapy The “Adventures in Voice” program, a pediatric resonant voice therapy protocol developed by Katherine Verdolini-Abbott, PhD, an independent consultant, and her colleagues for use with generally cognitively appropriate children approximately ages 5 to 10 years with phonogenic conditions affecting voice is used for the voice therapy arm of this study. This program targets three issues in voice training: 1) biomechanical target of voice production, 2) establishing new voice patterns and overriding existing ones, and 3) compliance to therapeutic directives outside of the clinical setting. The biomechanical component of the program trains the use of barely adducted or barely abducted vocal folds for all voice use. The learning component highlights the importance of attention and perception in motor learning, particularly attention directed to movement outcomes rather than biomechanics. It utilizes practice of the target voicing pattern in multiple contexts specific to the particular learner to establish new voicing patterns. The last component, namely compliance with therapeutic directives, focuses on how a participant will reaches program objectives during the course of therapy through adherence to the program outside the clinic, such as completing homework, etc.

Information, training, and therapy directives are presented in 8 weekly therapy sessions conducted over a course of 8 to 12 weeks, with 8 weeks being the target time frame. These sessions last for 1 hour. A certified speech-language pathologist who has received

Downloaded From: https://jamanetwork.com/ on 10/02/2021 training in this particular resonant voice therapy program by Dr. Verdolini-Abbott and/or her designated associates conduct these sessions

Individual sessions focus on the following with principles and concepts presented in an age-appropriate manner: Session 1: • Description of basic fundamentals of vocal anatomy and function related to vocal fold nodule pathology which will include vocal hygiene education and establishment of a personalized vocal hygiene program. • Behavior modification of loud voice use with softer voice use as well as providing acceptable alternatives known as “Voice Healing Powers”. • Training of Voice Healing Power 1(water and steam), Voice Healing Power 2 (no voice, quiet, inside voice) and Voice Healing Power 3 (loud body not loud voice). Session 2: • Training in recognition and production of resonant voice using “easy vibrations”. • Training of Voice Healing Power 4 (easy vibrations) and Voice Healing Power 5 (“secret signal”). Session 3: • Review and practice of Voice Healing Powers previously trained. • Identification of self-correction strategies. • Training of Voice Healing Power 6 (recovery pull-outs). Session 4: • Review and practice of Voice Healing Powers previously trained. • Training in and practice of more appropriate voice projection/loud voice production. • Training of Voice Healing Power 7(strong clear voice). Session 5: • Review and practice of Voice Healing Powers previously trained. • Review and practice of proper loud voice use. Session 6: • Review and practice of Voice Healing Powers previously trained. • Training in loud voice use with background noise. • Training of Healing Power 8 (healing in noise). Session 7: • Review and practice of Voice Healing Powers previously trained. • Training of loud voice use in context of various emotions. • Training of Voice Healing Power 9 (letting go of excess laryngeal muscle tension) Session 8: • Application of voice use and Voice Healing Powers in real-life activities specific to and chosen from a closed set by the individual child. • Discussion of troubleshooting and application of Voice Healing Powers outside of clinical setting. • Training of Voice Healing Power 10 (If I get sick or hoarse…)

4.8.3 Office Instructions Alone (No Voice Therapy) The office instructions alone arm or vocal hygiene program entitled “My Voice Adventure” was created at Massachusetts Eye and Ear Infirmary. The “My Voice Adventure” voice hygiene program (MVA) is a representative compilation of typical programs currently used for children ages 6 to10 years by pediatric speech pathologists today (September 2010) to reduce and eliminate vocal behaviors that are known to result in phonotraumatic changes to the vocal folds. Its development was largely influenced by the vocal hygiene program developed for classroom use by Holly Nilson and Carl R. Schneiderman as published in

Downloaded From: https://jamanetwork.com/ on 10/02/2021 Language, Speech, and Hearing Services in Schools in 1983 entitled: “Classroom Program for the Prevention of Vocal Abuse and Hoarseness in Elementary School Children” combined with generally accepted practices and service delivery concepts for addressing phonotraumatic behaviors in school-aged children. Other published influences include the Vocal Abuse Reduction Program (VARP) [102] and Using Your Voice Wisely and Well: Vocal Awareness Activities for Children, 2nd Ed. [101]. The final product of “My Voice Adventure” includes principles and activities developed through combined years of clinical experience in general pediatric and targeted voice therapy by its authors.

“My Voice Adventure” is more than a checklist of “do’s and don’ts”. Its main focus is on the education of principles and concepts rather than on classical conditioning or direct behavioral modification of phonotraumatic behaviors. Rather than using the stimulus, response, antecedent format, “My Voice Adventure” uses education and discussion of principles with reinforcing games and activities to increase awareness sufficiently and motivate the child to change potentially phonotraumatic behavior.

This program takes the education and training elements typically addressed in school settings and divides them into six clearly outlined therapy sessions. Four main areas of focus comprise the first four sessions of the program and are followed by two sessions of review. Areas of focus include:

1. Basic education regarding the vocal mechanism including normal function and general care, 2. Identification and discussion of vocally phonotraumatic behaviors and ways to avoid or modify, 3. Identification and discussion of potentially phonotraumatic situations and strategies to avoid or modify these situations, 4. Discussion and identification of adequate verses inadequate voice quality and production.

Most therapy sessions for the “My Voice Adventure” program are designed to be 30 minutes sessions with the exception of the first session, which is longer to obtain necessary preliminary information from both the parent(s) and child, initiate development of clinician-patient rapport, and establish the therapeutic conditions including therapy rules and expectations.

Therapy sessions should occur every nine to ten days whenever possible. All six therapy sessions should be completed within eight weeks, and must be completed within 12 weeks for the purposes of this study.

All treating speech language pathologists (SLPs) involved in this protocol deliver both Adventures in Voice and My Voice Adventure. Therefore, training for the “My Voice Adventure” office instruction program should occur at the same time as training for Adventures in Voice. All treating SLPs are to attend a live training session provided by Dr. Verdolini or her designated trainers. Those unable to attend a live training session are required to view training video.

4.8.4 Control for Equipoise and Cross-Intervention Contamination There may be concern regarding clinician bias in favor of voice therapy or office-only instructions, and that this source of bias may have the potential to affect study results. This possibility is assessed in pre- and post-hoc testing. Based on past experience, enthusiastic training in both intervention modalities helps to offset the likelihood of bias. In two prior clinical studies, there was no evidence of any clear biases based on formal and informal observations (Verdolini et al., in preparation[a]; [b]). Moreover, based on common clinical practice, most clinicians have genuine concern for their patients, and it is anticipated that clinical instinct attenuates any weak tendency there may be for a disruption in equipoise. Should post-hoc testing of clinicians reveal any bias; this factor will be taken into consideration in the interpretation of the results.

A second concern pertains to the potential for “bleed”—or cross contamination—between the intervention programs. This concern arises primarily from the fact that all treating clinicians deliver both therapy interventions. The most straightforward solution is for interventions to be delivered by different clinicians. Instead, the importance of separating the two interventions through therapy materials both in content and the tools being utilized in each intervention has been repeatedly emphasized to clinicians.

Downloaded From: https://jamanetwork.com/ on 10/02/2021 Those parameters are indicated in the formal Fidelity Check protocol. Clinicians who are found to produce contamination will be duly informed, and steps will be taken to eliminate contamination in future sessions.

4.9 Outcome Measures

4.9.1 Primary Outcome: Voice-Related Quality of Life The primary endpoint of the study is voice-related quality of life in pediatric patients with vocal fold nodules, as measured by scores for the Pediatric Voice-Related Quality of Life Instrument (PVRQOL) [23] (See Appendix 1). The PVRQOL instrument has been validated in the pediatric population with voice disorders, and has demonstrated test-retest reliability and discriminatory validity. It consists of 10 items completed by parent proxy administration (see Appendix). On average, it takes 5 minutes to complete [57]. A change score of 12 on its scale (range 0 (worst) to 100 (best)) is deemed clinically significant and corresponds to a change in the status from “as bad as it can be” to causes “a lot” of problem in one of ten domains. Parents complete the instrument at screening, immediately prior to randomization, and during scheduled follow up visits at 8 (or 12) and 12 (or 16) weeks. The primary outcome measure is the mean score at 12 (or 16) weeks, although data from all follow up times will be examined in secondary analyses.

One of the inherent difficulties in assessing quality of life in children is the need to often utilize proxy reporting. Very large studies assessing the validity of parent-proxy reporting have been completed and have demonstrated the efficacy and utility of parent-proxy instruments when appropriately validated.[96]

4.9.2 Secondary Outcomes: Perceptual Voice Quality, Aerodynamic and Acoustic Measures, and Nodule Size The secondary objective of this study is to examine the extent to which variations in the primary outcome measure are reflected by similar variations in other measures of vocal function such as (1) perceptual measures of voice quality and degree of perceived resonant voice, derived from utterances using general recording procedures recommended by the Consensus Auditory Perceptual Evaluation of Voice (CAPE-V); (2) phonation threshold pressure obtained during aerodynamic testing, (3) harmonics-to- noise ratios obtained during acoustic recordings, (4) nodule size and contour using a validated nodule grading scale. The same data collection and data analysis procedures will be used for all baseline and follow-up time points. An important feature of data analysis is that data collection for all secondary measures, and all data analyses, will be completed by individuals who are blinded to subjects’ group (resonant voice therapy or office education) and where possible, also time-point (baseline, 8 (or 12)- week, or 12 (or 16)-week follow-up). Clinicians performing measures may recall having seen a child before, or may recognize the child has not been seen before, complicating blinding to time point. However, full blinding will be used with respect to all data extraction and statistical analyses.

Rationale for the measures was provided earlier in the document. Details regarding data capture and data extraction follow next. (For information about statistical analyses, see section 4.10)

4.9.2.2 Perceptual Measures of Overall Voice Quality and Resonant Voice Equipment and Procedures. The child will be seated comfortably in a quiet environment. The clinician will ask the child to produce the sentence, “We were away a year ago,” an all-voiced sentence from the CAPE-V -- after hearing a standardized, high-fidelity recording of the sentence made by a vocally healthy adult, male speech-language pathologist or otolaryngologist employed in the Voice and Speech Lab at the Massachusetts Eye and Ear Infirmary (MEEI). Selection of the speaker will be constrained by the absence of dialectical influences in speech, as judged independently by three speech-language pathologists in the MEEI Voice and Speech Lab. . Because we wish to capture the child’s typical voice and speech patterns as much as possible, the only instruction will be, “Please say this sentence,” prior to playing the recording in free field using high-fidelity speakers, at an intensity level of 75 dB SPL (+ 1 dB) at the child’s ears bilaterally. The rationale for having the child repeat the sentence after a recording as opposed to reading it is that some children in the study may not have adequate reading skills for the task and thus repetition will be required for them. We will thus present the recording to all children, so that procedures will remain consistent for all subjects. The same recording will be used at all study sites. The rationale for using a recording of an adult is that listening to another child’s voice could bias the child to

Downloaded From: https://jamanetwork.com/ on 10/02/2021 attempt to replicate the recorded voice because of its relative similarity to the child’s own voice, compromising our attempt to capture the child’s typical production as closely as possible. The rationale for using an adult male as opposed to adult female voice is that the male voice will differ from the child’s voice more than the female voice. In that way, production bias should be further reduced. The child’s replication of the recorded sentence will be captured using a condenser microphone placed at an azimuth of 45° from the front of the mouth, with a 4-cm microphone-to-mouth distance recorded in a quiet environment. Audio recordings will be digitized in a computer with 16 bits of resolution and a signal- sampling rate of no less than 20 KHz.

Data Extraction. After data collection has been completed for all subjects, the recordings will be analyzed using the paired-comparison approach described in Preliminary Studies. Procedures will replicate those established by Kreiman and colleagues [67]. In brief, three speech language pathologists at the Massachusetts Eye and Ear Infirmary Voice and Speech Lab under the supervision of the Senior Speech Language Pathologist for this study will receive a set of paired voice recordings derived from all possible paired comparisons of all subjects’ voices at each time point. Listeners will be blinded to subjects’ condition (resonant voice therapy or office intervention, and time point). Each pair of utterances in the set will involve some combination of recordings of the sentence, “We were away a year ago” made at baseline, 8 (or 12)_-week, and 12 (or 16)-week time points, from the same subject. Audio samples will be played to listeners in randomized order, re-randomized for each listener. Twenty-five percent of all pairs will be repeated in each set. The listener’s task will be to indicate, forced choice, which voice is globally “better.” Then, the listener will rate his or her confidence in the decision (1=”positive;” 5=”wild guess”). The final rating for each voice sample will be calculated on a scale derived from the combination of overall voice quality and confidence ratings, such that 1 = “positive voice is better,” 5 = “wild guess voice is better,” 6 = “wild guess voice is worse,” and 10 = “positive voice is worse.” After all listeners have completed ratings for all subjects at all time points, statistical analysis of the data will be completed as indicated in section 4.10 to assess intra- and inter-listener agreement as well as possible differences in pre-post therapy voice quality as a function of intervention type.

At a different time, and in counterbalanced order, the same listeners will rate the same samples for the degree of perceived “resonant voice” (“which voice is more resonant?” and “how confident are you in your decision?”). Again, listener agreement will be calculated from the forced-choice results. Results will be examined in relation to changes in the PVRQOL score across time points, to gain preliminary insight into whether the acquisition of resonant voice may have mediated changes in PVRQOL.

4.9.2.3 AerodynamicTesting Equipment and Procedures. All aerodynamic measures will be made using the KayPENTAX Phonatory Aerodynamic System (PAS) Model 6600 which was designed and developed to obtain phonatory acoustic/aerodynamic data (frequency, sound pressure, airflow, and air pressure) of speech and voice signals. The PAS is a sophisticated pneumotach consisting of a tube with a series of mesh screens which detect the drop in pressure as air passes through the tube. Specifically, the components of the PAS external module are the PAS handheld device, integrated microphone, airflow head (pneumotach), mask and airflow tube, intra-oral pressure kit and coupler.

The mask and intra-oral tube are the parts that have direct contact with the patient. The mask is held tightly against the patient’s face to avoid air leakage. The intra-oral tube is used to collect air pressure measures and rests inside the patient’s mouth without contact against the cheek or tongue. The assumption is that the oral pressure during occlusion of the vocal tract during production of the /p/ equals the lung pressure during the vowel production. [102] [103][97, 98] Minimum lung pressure required to initiate vocal fold oscillation, has been found to be sensitive to changes in vocal fold performance.[99] The measures will be taken following approximately 5-10 minutes of training in order to insure that the correct rate of syllable production is achieved while obtaining acceptably smooth pressure traces [104]. The aerodynamic measures will be recorded at the baseline assessment, and as follow-up measures at 8 (or 12) and 12 (or 16) weeks, with the primary data analysis focused on results at 12 (or16) weeks in comparison to baseline.

Downloaded From: https://jamanetwork.com/ on 10/02/2021 The aerodynamic measures obtained will focus on estimated Phonation Threshold Pressure (PTP), defined as the minimum amount of subglottal pressure required to initiate and sustain vocal fold oscillation. These measures will be obtained in a non-invasive manner from oral pressures generated during repetitive production of five consonant-vowel-consonant (CVC) strings (/pae pae pae pae pae/. The measures will be obtained at the quietest possible phonation levels—as encouraged, modeled, and monitored by the clinician-- at a rate of 1.5 syllables/sec and at the 80th percentile pitch established for the /a/ vowel at baseline individually for each child (from the lowest “growl” to the highest “screech” and measured using KayPENTAX’s Computerized Speech Lab (CSL). A total of 5 strings of /pae pae pae pae pae/ will be captured for each child at each time point. Strings the clinician considers are invalid due to the intrusion of patently suprathreshold phonation or subthreshold whisper will be recollected. Only valid trials will be included in data analysis.

Data Extraction. Information obtained by the KayPENTAX Phonatory Aerodynamic System (PAS) Model 6600 will be analyzed by software and stored on each center’s computer hard drive. The Microsoft Windows XP/Vista operating system is used to run the PAS program (version 3.2). Phonation threshold pressure measures will be obtained using the PAS Voicing Efficiency Protocol (VEP) which allows for the extraction of estimated subglottal pressure from intraoral pressure readings (air pressure with the patients mouth closed around the intraoral pressure tube during production of the /p/). Specifically, phonation threshold pressure will be obtained by selecting the middle three pressure peaks in each /pae pae pae pae pae/ sequence, and obtaining pressure values for each peak. Data from all pressure peaks will be included in data analysis (section 4.10.).

4.9.2.4 Acoustic Recordings and Analysis Equipment and Procedures. The Computerized Speech Lab (CSL) is KayPENTAX’s most advanced speech analysis system with uncompromised specifications and performance. CSL is a complete hardware and software system and is an extremely powerful yet very easy to use system for obtaining and analyzing speech signals. Vocal input can be read and stored at rates of up to 51,200 samples per second, providing detailed information to the clinician or researcher on a variety of speech parameters, including pitch, timing and energy.

Digital recordings of the patient's speech signal will be obtained in a double-walled sound attenuated booth using a Tascam DA-30 recorder, a Zygo head pointer mic holder, and a Sennheiser Model MKE 2microphone with a standard microphone to mouth distance of 15 cm. The patient will be instructed to sustain the /i/ vowel in his or her “typical voice,” on three repetitions.

Data Extraction. The primary acoustic measure obtained will be Harmonics-to-Noise [HNR) obtained from three steady-state vowel productions of the /i/ vowel. Recordings will be analyzed using KayPENTAX’s Computerized Speech Lab (CSL). To analyze the steady-state vowel productions, the steadiest one-second segment from the mid-portion of each vowel production will be selected from the digitized waveform for analysis, and HNR results for that segment will be obtained using CSL software. Results from all three samples will be retained in data analysis. Since all data analysis will be hand- sampled, 25 % of the samples will be selected randomly and re-measured at a later date to assess reliability.

4.9.2.5 Nodule Grade All children who can tolerate it undergo a follow up video-stroboscopic examination at the 12 week follow up visit. A grading system for vocal fold nodules has been validated with established inter-rater reliability for nodule presence and size [23] The 3-point grading scale for nodule size will be used with grade 1 being small, grade 2 a moderate nodule, and grade 3 a large nodule. The contour of the nodule will also be rated as either discrete or sessile. (also see Appendix 3). One secondary objective of the study is to compare nodule grade across baseline and 12 (or 16) weeks. Changes in grade are compared for subjects treated with voice therapy versus office instructions alone.

Consideration was given to making nodule grade (size) the primary outcome, and the objective outcome of nodule grade (size) is an attractive metric for treatment outcomes. However, there are data which have established that nodule size or resolution do not strictly correlate with improvements in voice [100, 101].

Downloaded From: https://jamanetwork.com/ on 10/02/2021 In fact, patients may have resolution to normal voice, even in the presence of residual nodules. Furthermore, the size of the nodule does not appear to correspond to the level of vocal dysfunction. [42] Thus, although it is a concrete physical measurement, it is still more appropriately chosen as a secondary outcome, rather than a primary outcome. Additional secondary measures such as changes in perceptual rating of voice obtained with the CAPE-V, phonation threshold pressure (PTP) changes, and/or harmonics to noise will be used to support any change noted in nodule grade with either voice therapy or office hygiene instruction only.

4.9.3 Other Measures and Indices Data about patient compliance will consist of (1) the number of sessions attended; and (2) number of homework assignments completed as prescribed, for each subject group. Data will be analyzed as indicated below. Also, at study completion, data will be tabulated regarding fiscal and time costs associated with each treatment, and will be reported, as will reasons for subject attrition.

4.10 Statistical Analyses This project identifies a primary outcome measure. Also a series of secondary and tertiary measures of possible interest are identified. The alpha level is set at .05 for the primary outcome. We will not correct for investigation-wide alpha error across remaining secondary and tertiary measures, as they are considered exploratory.

4.10.1 Primary Outcome Measure (S.A.1) The primary analysis will use intention to treat principles, i.e. all randomized patients will be included in the analysis even if their treatment was discontinued early. Patients will continue to be followed for the 12 (or 16)-week follow-up period if treatment is discontinued. The t-test for two independent samples will be used to compare mean changes from baseline in PVRQOL scores in the two treatment groups. The rationale for the use of the t-test is that the study was powered based on clinical data using this statistic. If a patient does not complete follow-up, then multiple imputation procedures will be used, i.e. a regression model will be developed based on patients with complete data to identify patient characteristics which are associated with the change in PVQOL. Then an estimated change score distribution will be generated for each patient with less than complete data and a value from that distribution will be randomly selected. As an additional (secondary) analysis of the primary outcome measure, we will use mixed-model Analysis of Variance (ANOVA), with time and group as independent variables, to model the set of responses over time and treatment differences will be assessed by the time by treatment interaction. Pattern mixture models will be used to account for missing data. If initial data analysis shows that PVRQOL scores do not follow a normal approximation, then the data is analyzed with non-parametric methods instead. The differences between baseline measurements and 12 (or 16) week measurements are calculated, and the medians compared. The null hypotheses are then tested using the Wilcoxon rank sum test. Two analyses will be done. The first will assign the highest (worst) rank for each patient who does not complete follow-up. The second analysis will be based on patients who complete follow-up.

4.10.2 Secondary Outcome Measures (S.A. 2) To parallel analyses for the primary outcome measure, t-tests for two independent samples and mixed- model ANOVAs as described for the PVRQOL will be done for the following: PVRQOL administered to speech-language pathologists Voice quality Nodule grade including nodule size and contour Phonation threshold pressure Harmonics to noise ratio

4.10.3 Additional Analyses about Possible Mechanisms Mediating Changes in Primary Outcome (S.A. 3: Exploratory) We will conduct several additional analyses of the primary outcome measure. 1. We will compare mean changes among those participants who completed the full treatment protocol.

Downloaded From: https://jamanetwork.com/ on 10/02/2021 2. We will conduct a multiple linear regression analysis to evaluate the magnitude of the treatment effects after adjustment for participant characteristics such as age, gender, and baseline vocal nodule grade, harmful vocal habits, and degree of resonant voice. 3. We will conduct a longitudinal analysis using mixed-model techniques comparing the trend of PVRQOL across the different treatment arms (interaction between treatment group and time) after adjusting all the covariates listed in (4). Model selections will choose the appropriate random terms and variance-covariance structure. 4. We will evaluate the responses and characteristics of those participants who do not complete treatment and compare to those who do complete treatment. 5. Withdrawal rates from study and from treatment will be compared across treatment groups using Fisher’s exact test. Reasons for withdrawal will be described for each treatment arm. 6. Compliance data will be described for each treatment arm using summary statistics.

4.11 Interim Monitoring Rule and Adjustment of Sample Size An independent Data and Safety Monitoring Board (DSMB) will be formed, including 2 speech-language pathologists, one otolaryngologist, one ethicist, one biostatistician, and NIDCD representation and approval to periodically review the study. The initial meeting will be held prior to the start of patient recruitment. At that time, the DSMB will decide what information it will need to monitor the study, whether it wants to have the treatment groups identified and whether an interim monitoring rule should be employed. The DSMB will meet at least bi-annually thereafter through the end of patient follow-up to review the study performance.

We propose that interim efficacy analyses for the randomized clinical trial be performed using an alpha spending function approach to control the overall Type I error probability to be 0.05. This approach was selected because it allows flexibility in the timing and number of interim analyses and these need not be pre-specified. We will use approximate O’Brien-Fleming boundaries so that early interim analyses will be conservative. It is anticipated that the DSMB may request up to two annual interim analyses in addition to the planned end-of-study analysis. If these occur at times when 33.3% and 66.7% of the study information has been accumulated, the alpha levels for these tests will be alpha equals 0.0002 (first interim analysis), alpha equals 0.012 (second interim analysis), and alpha equals 0.0463 (final look). The monitoring boundaries allow early stopping for rejection of the null or the alternative hypothesis (futility). East 5.2 was used to estimate the sample size based on these parameters. The target sample size, after adjustment for the interim monitoring rule, is 59 per group or 118 total.

4.12 Contingency Plans If voice therapy sessions are not tolerated because of behavioral or logistical problems, attempts will first be made to at least maximize the number of sessions that the patient can tolerate. If voice therapy cannot be tolerated at all, then it will be stopped, after an exit interview in which verbal and written instructions regarding maximizing vocal hygiene and improving resonance are given to the subject and parent. The number of sessions successfully completed prior to exit will be recorded for data analysis and reasons for exit will be noted descriptively. However, the subject will continue follow-up.

4.13. Participating Clinical Sites Three centers are planned for enrollment: Massachusetts Eye and Ear Infirmary, Drexel University College of Medicine, and Medical College of Wisconsin. A fourth site was added (The Hospital for Sick Children) to aid in recruitment efforts (02/2012).

The Massachusetts site attracts patients with pediatric vocal dysfunction through the Pediatric Airway, Voice, and Swallowing Center, a joint program between the Massachusetts Eye and Ear Infirmary and the Massachusetts General Hospital. Patients who attend this clinic have the advantage of being seen in one setting by three pediatric subspecialists (otolaryngology, pulmonology, and gastroenterology), as well as the supporting services in one visit. Approximately 60-80 children with vocal fold nodules are seen each year at this site. The investigator at this site is Dr. Christopher Hartnick, M.D. M.S.Epi.

The Medical College of Wisconsin (MCW) site is located within Children’s Hospital of Wisconsin (CHW) which is a regional referral center and the only free-standing children’s hospital in the state of Wisconsin.

Downloaded From: https://jamanetwork.com/ on 10/02/2021 Patients with voice disorders are evaluated in the Voice Clinic at CHW, which is part of the Airway, Digestive and Voice Center at CHW; a joint program between CHW and MCW. Patients who attend this clinic have the advantage of being seen in one setting by three pediatric subspecialists as well as the supporting services in one visit. Approximately 60-80 children with vocal fold nodules are seen each ear at this site. The chief investigator at this site will be Dr. Joseph E. Kerschner, M.D.

Drexel University/ The American Institute for Voice and Ear Research attracts voice patients from around the world. Patients seen at this center are evaluated by one of four laryngologists, and by speech- language pathologists, singing voice specialists, acting voice specialists, and collaborating specialists in related disciplines (neurology, pulmonology, gastroenterology, and others. In addition, the center includes a comprehensive clinical voice laboratory and an experienced laryngeal electromyographer. Approximately 1,058 voice patients are seen annually, accounting for 4,243 visits for voice care (2008 Data). Approximately 10-15 children from ages 6 to 10 are treated annually for vocal fold masses. The chief investigator at this site will be Dr. Robert Sataloff, although three of the four laryngologists at this center treat children.

The Centre for Paediatric Voice and Laryngeal Function is a specialty clinic designed for the care of children with voice disorders. The Centre is the only publicly funded clinic fully dedicated to pediatric voice care and research within the province of Ontario. The clinic is situated within the Hospital for Sick Children, an academic, pediatric referral centre in Toronto. Patients are referred to the clinic from the community, provincially and nationally. The clinic has a dedicated partnership of otolaryngology and speech language pathology to provide the most effective assessment and management of voice disorders. Since 2004, the clinic has assessed and managed more than 500 children with a variety of voice disorders. Approximately 50 children with vocal fold nodules are seen each year. The clinical director is Dr. Paolo Campisi, MD.

Each clinical site is actively involved in the recruitment, evaluation, and treatment of the study participants. Each of the sites has at minimum a Principal Investigator (PI), Clinical Research Coordinator, and two Speech Language Pathologists, one treating SLP who provides both therapy arms and a non-treating SLP who collects secondary data. These individuals from each site provide the expertise necessary for the successful completion of the study protocol.

The multicenter format is chosen for several reasons. First, multiple centers accrue patients more rapidly than a single center. Second, the use of multiple centers widens the potential generalizability of the study’s results. One advantage of the multicenter format of the study is that it helps assure racial and ethnic diversity of the study population. Third, multiple investigators bring a wide range of skills to the study.

The MCW/CHW site will bring expertise in having performed previous clinical trials, significant expertise and publications in this particular area of laryngeal pathology, and a well-organized research team with a dedicated research nurse and research advanced practice nurse with significant experience in conducting clinical research investigations. This expertise, along with the expertise of Dr. Kerschner as a funded NIH investigator will significantly assist the PI in study design, patient accrual, data analysis and coordination and completion of the proposal as described.

4.14 Timeframe Between the three clinical sites, based on review of patient databases over the past 24 months, 200 patients between the ages of 6 to 10 years of age are seen on an annual basis and roughly two-thirds of those would be eligible for study inclusion. These children with vocal nodules and dysphonia are seen in the three recruiting centers “voice” practices as part of the normal clinical volume of patients presenting for treatment. Utilizing a rather conservative enrollment rate of 25% of eligible participants, which has been achieved in other clinical trials at each site, this will allow for 30-35 patients to be enrolled on a yearly basis.

Downloaded From: https://jamanetwork.com/ on 10/02/2021 Due to lower than expected recruitment in the first 6 months of the recruitment period, a new site has been added: The Hospital For Sick Children. The addition of the site will not affect the sample size nor the number of subjects expected to be recruited on a yearly basis.(02/2012)

This study will last 5 years and consist of three time periods: a six-month planning period, a 36-month recruitment and follow up period, and a 12-month close-out and analysis period.

4.15 Future Research The present study, which is best cast as a Phase II trial, will establish critical groundwork for future Phase III trials. Specifically, future studies will expand the subject base in terms of numbers, geographic location, race, and ethnicity to determine the stability of results in the expanded pool. Moreover, results from the present study will be used to gain preliminary insights into mechanisms that may mediate benefits from voice therapy as well as factors that may predict outcome. Those insights will provide a platform for systematic assessment of the issues in studies that are specifically designed to examine them.

5.0 HUMAN SUBJECTS

5.1 Protection of Human Subjects Equipoise exists. There is an inherent state of equipoise, as the impact of voice therapy has not been clearly established in children with vocal fold nodules in prospective trials. In addition, the approaches in each of the potential treatment arms are all variations used in current practice, in which there is no single established standard of care [11, 38, 39]. Regardless, as indicated in 4.8.4, precautions will be taken to assess potential clinician bias through a Fidelity Check protocol. Clinicians expressing a bias will be excluded from participation. Moreover, training will emphasize the potential equal value of both interventions being assessed, and fidelity checks will ensure the correct delivery of the treatments. Finally, our past experience is that out of concerns for good patient care as well as the clinic’s wider reputation, clinical “instinct” will augment training emphases so that each patient will receive excellent care, independent of the specific intervention.

Subject risk and discomfort is minimal. Anticipated subject discomforts include coping with the challenges of attending and adhering to voice therapy teaching. Since voice therapy is a frequent therapeutic choice for nodules, however, such discomforts are within the realm of what they would experience outside of the study. If voice therapy proves more beneficial than the office instructions alone/placebo treatment arm, then subjects in the control group will experience a delay in treatment. If voice therapy shows no difference from office instructions alone/placebo, however, then subjects in the active intervention groups will have been exposed to the risks of those treatments without benefit. No treatment arms involve invasive procedures. Of note, children’s voices may get worse whether they receive treatment or not. Patients and parents will be informed of this possibility and will be provided information on alternative options for clinical care.

Participants will be given ample opportunity to defer enrollment or withdraw. Interested candidates are given written information describing the rationale, the treatment arms, the expectations for follow up testing, the potential risks and benefits, and the unknown nature of the outcome of the study. This summary also indicates that the subject may withdraw from testing at any time without penalty. Subjects have ample opportunity to ask questions of the participating investigators and research coordinators.

Informed consent will be obtained. The informed consent form for this study is a parent proxy form that describes the goals of the study, the two potential assignment groups, the anticipated risks and benefits, and their ability to withdraw from the study as desired. Participants will affirm their willingness to participate by affixing their signature to the Informed Consent form and Assent Form, if applicable for the clinical site.

Subject confidentiality will be protected. Data are collected in a database where subjects are identified only by number; no data are stored with an associated patient name. Code sheets that link the study number to patient name are kept in locked filing cabinets. Any electronic data are stored in password-

Downloaded From: https://jamanetwork.com/ on 10/02/2021 protected files on password-protected computers. Only study investigators have access to the passwords. No data linked to a subject is released without written permission from the participant.

5.2 Monitoring Serious Adverse Events

5.2.1 Role of the Local Site Investigator in Adverse Medical Event Monitoring

The local site investigator will be responsible for following serious adverse event reporting requirements: a. Reviewing the accuracy and completeness of all serious adverse events reported b. Complying with study policies as well as local IRB policies for reporting serious adverse events c. Reporting to the IRB safety issues reported to the site by the principal investigator and d. Closely monitoring research volunteers for any new Adverse Serious Adverse Events (SAEs). e. Reporting all SAEs to the Study Chair’s Office within 24 hours.

5.2.2 Definitions Serious Adverse Events (SAEs) as defined by the International Conference on Harmonization for Clinical Safety Data Management are untoward medical occurrences that: a. Result in death; b. Are life-threatening; c. Require inpatient hospitalization or prolongation of existing hospitalization; d. Result in persistent or significant disability or incapacity; e. Result in a congenital anomaly/birth defect; or f. Result in any other condition that, based upon medical judgment, may jeopardize the subject and require medical or surgical treatment to prevent one of the above outcomes.

5.3 SAE monitoring and reporting Study participants will be monitored at each clinic visit and telephone contact for SAEs. All SAEs including both those related to the study intervention and those not related to the intervention will be collected and recorded on the appropriate study event forms. Active monitoring of SAEs will begin as soon as a study participant signs the Informed Consent and will continue through end-of-study for each participant.

All SAEs require expedited notification. Expedited notification is defined as the completion and submission of the SAE form within 72 clock-hours of a site investigator’s initial awareness of an SAE. The principal investigator will be responsible for evaluating all SAEs for patient safety concerns in a timely fashion. All related and unexpected events (IND Safety Reports) will be reported to the FDA within 15 calendar days.

5.4 Minimization of Expectation Bias Masking of the type of intervention is performed whenever feasible in this study. Regarding equipoise, please see 5.1 above.

5.3 Safety

5.31 Adverse Effects Associated with Voice Therapy Adverse effects associated with voice therapy are measured in terms of 1) number of hours of school- related activity missed over the initial 8 weeks of enrollment in the study, 2) presence/absence of failure of school examination, 3) number of hours of work-related activity missed over the initial 8 (or 12) weeks of enrollment in the study, and 4) presence/absence of loss of parent employment.

The number of hours of missed school and work is analyzed as a continuous variable. It will be tested using a two-sided t-test between voice therapy involved groups and office instruction alone groups.

The presence/absence of children with a failed examination/loss of parental employment is analyzed as a binary variable, according to the proportion of total children with each event. Given the low number of

Downloaded From: https://jamanetwork.com/ on 10/02/2021 anticipated major events in these categories, as well as the desire to avoid false positive results from multiple comparisons, no formal statistical test is planned. If, however, the rate is higher than anticipated (>10%), then the Fisher’s exact test is used to test the null hypothesis, in comparing subjects attending voice therapy sessions to those who were managed with office instructions alone.

5.3.2 Study Conduct Monitoring and Oversight The groups charged with centrally monitoring the various aspects of the study will be: the Executive Committee (EC) and the Data and Safety Monitoring Board (DSMB). Both committees meet at study start-up and annually thereafter. The EC will also have quarterly conference calls. The DSMB may elect to meet more frequently if it deems necessary.

The Executive Committee is the management and decision-making body for the operational aspects of the study. One of its major responsibilities is to monitor the performance of the participating medical centers. The EC considers the need for protocol modifications. The Executive Committee also reviews and approves all manuscripts and abstracts emanating from the study. Typically the EC is composed of the original study Planning Committee. It is chaired by the study PI.

The Data and Safety Monitoring Board (DSMB) will review the progress of the study, including patient intake, completeness of follow-up, data quality, protocol deviations and safety. The DSMB may also choose to implement a formal interim monitoring rule to monitor efficacy. The DSMB will review any protocol modifications recommended by the EC. The DSMB will make recommendations to NIDCD as to whether the study should continue or be terminated. Interim unblinded progress reports will be provided to the DSMB by the study biostatistician.

The Study Group, which consists of all participating investigators, will meet prior to the start of recruitment to be trained on study procedures and annually thereafter to discuss the progress of the study and any problems encountered during the conduct of the trial. Finally, the study protocol and progress must be reviewed annually by the IRB at each site in order for that site to participate in the study.

Site Performance. The Study Chair’s Office and Hines CSPCC evaluate recruitment and retention performance monthly. Other performance problems such as protocol deviations, poor data quality, missing or overdue data, and reasons for withdrawal from study will also be tracked. A monthly study conference call for all study will be held to review study recruitment, data quality and protocol adherence.

5.4 Data Collection and Management

Data Collection. A manual of procedures and case report forms will be developed jointly by the Study Chair’s Office and the Hines Data Management Center. The site study coordinator will collect all data. After a patient signs an informed consent form to participate in the study, the site coordinator will create a 2 files for the participant: one will contain the consent forms and any identifying documents which has the participants name and another that will contain all relevant research documents including source documents, and any other information pertinent to the study that will contain only the participant’s study identification number assigned at screening. The case report forms will be completed at each visit through the Cooperative Studies Program electronic data capture system (EDC). This is a FISMA- compliant system. The system includes a Share Point portal which serves as a central communications points for study documents. REDCap is used as the front end for data entry. Data are stored in an SQL database and transferred to SAS for statistical analysis. Access to the system is controlled by the Hines Data Center. The data center personnel, including the Biostatistician, Statistical Programmer, Database Programmer and Data Coordinator at Hines CSPCC, will have access to read the data.

Data Quality Assurance. Extensive data checks, including missing values, out-of range entries, and consistency between variables, both within and across forms, will be built into the computerized data capture system developed for this project. There will be two levels of checking. The first level will be done at the time the data is entered onto the case report form. These checks will automatically appear on the screen at the time of entry. After submission of the form into the data base, then second level checks

Downloaded From: https://jamanetwork.com/ on 10/02/2021 against other data already captured for that patient will be done. These checks will generate data discrepancies, which will be stored in the EDC system for the site coordinator to access and rectify. Any revisions to the data made by the study site personnel will be recorded in a system audit trail. The Hines CSPCC will monitor completeness and timeliness of the data discrepancy resolution made by site coordinators. The Hines CSPCC Data Coordinator will also periodically request copies of source documents for randomly selected case report forms to check accuracy of data entry.

Data Confidentiality and Data Security. The Hines Data Center will not receive any directly-identifiable data on study patients and therefore will not have the ability to link study data received from the study sites back to study patients.

Data Access. Data will be made available through a web-site at the Principal Investigator’s institution following all NIH guidelines for federally-funded clinical trials to allow for direct public access of data and independent verification of results.

5.5 Inclusion of females, minorities and children in research Females, minorities, and children ages 6-10 will be included in the research. It is anticipated that fewer girls than boys will be enrolled due to known higher rates of nodule occurrence in boys than girls. Every attempt will be made to recruit minorities according to their distribution in the study sites (Boston, Milwaukee, and Philadelphia). Interim analysis includes determining whether females and minorities are adequately represented according to these figures. If enrollment for either group is not adequate, then steps will be taken to promote increased enrollment of the under-represented groups. Such steps include: 1) The PIs and research coordinators at the participating sites contact regional pediatricians and post information that specifies that female or minority children with vocal fold nodules are especially desired; 2) Female patients are targeted via associated groups which are associated with increased vocal use, such as cheerleading and sports teams; 3) Minority patients are targeted via pediatrician clinics that exist in more ethnic regions of the study center cities. Groups and clinics are contacted by the study investigators via phone and email.

Downloaded From: https://jamanetwork.com/ on 10/02/2021 References 1. Berry, D., et al., New Indications of an optimal glottal half-width in vocal production. Journal of Speech, Language, and Hearing Research, 2001. 44: p. 29-37. 2. Sapienza, C.M., B.H. Ruddy, and S. Baker, Laryngeal Structure and Function in the Pediatric Larynx. Language Speech and Hearing Services in Schools, 2004. 35: p. 299-307. 3. Bureau, C. [cited; Available from: http://quickfacts.census.gov/qfd/states/25000.html. 4. Ramig, L.O. and K. Verdolini, Treatment efficacy: voice disorders. J Speech Lang Hear Res, 1998. 41(1): p. S101-16. 5. Verdolini, K. and L.O. Ramig, Review: occupational risks for voice problems. Logoped Phoniatr Vocol, 2001. 26(1): p. 37-46. 6. Ruben, R.J., Valedictory--why pediatric otorhinolaryngology is important. Int J Pediatr Otorhinolaryngol, 2003. 67 Suppl 1: p. S53-61. 7. Akif Kilic, M., et al., The prevalence of vocal fold nodules in school age children. Int J Pediatr Otorhinolaryngol, 2004. 68(4): p. 409-12. 8. Mandell, D.L., et al., Lack of association between esophageal biopsy, bronchoalveolar lavage, and endoscopy findings in hoarse children. Arch Otolaryngol Head Neck Surg, 2004. 130(11): p. 1293-7. 9. Merati, A.L., et al., Pediatric Voice-Related Quality of Life: findings in healthy children and in common laryngeal disorders. Ann Otol Rhinol Laryngol, 2008. 117(4): p. 259-62. 10. Green, G., Psycho-behavioral characteristics of children with vocal nodules: WPBIC ratings. J Speech Hear Disord, 1989. 54(3): p. 306-12. 11. Allen, M.S., J.M. Pettit, and J.C. Sherblom, Management of vocal nodules: a regional survey of otolaryngologists and speech-language pathologists. J Speech Hear Res, 1991. 34(2): p. 229-35. 12. Andrews, M.L. and A.C. Summers, Voice Treatment for Children and Adolescents. 2001, San Diego: Singular. 13. Wilson, D.K., Voice Problems of Children. 3rd ed. 1987, Baltimore: Williams and Wilkins. 14. Nilson, H. and C.R. Schneiderman, Classroom program for the prevention of vocal abuses and hoarseness in elementary school children. Language, Speech, and Hearing services in Schools, 1983. 14: p. 121-127. 15. Flynn, P.T., M.L. Andrews, and B. Cabot, Using Your Voice Safely and Well. 1990, Tuscon, Az: Communication Skill Builders. 16. Boone, D.R., The Boone Voice Program for Children, in A Vocal Hygeine Program for School Age Children, J.V. Cook, D.J. Palaski, and W.R. Hanson, Editors. 1993, Pro-eD: Austin, Texas. p. 21- 26. 17. Trani, M., et al., Voice therapy in pediatric functional dysphonia: a prospective study. Int J Pediatr Otorhinolaryngol, 2007. 71(3): p. 379-84. 18. Moran, M.J. and E. Zylla-Jones, Vocal Hygeine Activities for Children: a resource manual. 1998, San Diego: Singular. 19. Nienkerke-Springer, A., A. McAllister, and J. Sundberg, Effects of family therapy on children's voices. J Voice, 2005. 19(1): p. 103-13. 20. Bonilla, S., et al., School absenteeism in children with asthma in a Los Angeles inner city school. J Pediatr, 2005. 147(6): p. 802-6. 21. Eaton, D.K., N. Brener, and L.K. Kann, Associations of health risk behaviors with school absenteeism. Does having permission for the absence make a difference? J Sch Health, 2008. 78(4): p. 223-9. 22. Kearney, C.A., School absenteeism and school refusal behavior in youth: a contemporary review. Clin Psychol Rev, 2008. 28(3): p. 451-71. 23. Boseley, M.E., et al., Validation of the Pediatric Voice-Related Quality-of-Life survey. Arch Otolaryngol Head Neck Surg, 2006. 132(7): p. 717-20. 24. Shah, R.K., H.A. Feldman, and R.C. Nuss, A grading scale for pediatric vocal fold nodules. Otolaryngol Head Neck Surg, 2007. 136(2): p. 193-7. 25. Sander, E.K., Arguments Against Agressive Pursuit of Voice Therapy in Children. Lang Speech Hear Serv Sch, 1989. 20: p. 94-101. 26. Van Riper, T. and L. Emerick, Introduction to Speech Pathology and Audiology. 1984, Englewood, NJ: Prentiss-Hall.

Downloaded From: https://jamanetwork.com/ on 10/02/2021 27. McNamara, A.P. and C.K. Perry, Vocal Abuse Prevention Practices. Lang Speech Hear Serv Sch, 1994. 25: p. 105-111. 28. Carding, P.N., S. Roulstone, and K. Northstone, The prevalence of childhood dysphonia: a cross- sectional study. J Voice, 2006. 20(4): p. 623-30. 29. Maddern, B.R., T.F. Campbell, and S. Stool, Pediatric voice disorders. Otolaryngol Clin North Am, 1991. 24(5): p. 1125-40. 30. De Bodt, M.S., et al., Evolution of vocal fold nodules from childhood to adolescence. J Voice, 2007. 21(2): p. 151-6. 31. Connor, N.P., et al., Attitudes of children with dysphonia. J Voice, 2008. 22(2): p. 197-209. 32. Association, A.S.-L.-H. Incidence and Prevalence of Communication Disorders and Hearing Loss in Children - 2008 Edition. 2008 [cited; Available from: http://www.asha.org/members/research/reports/children.htm. 33. Building the Legacy: Idea 2004, US Department of Education. 34. Mori, K., Vocal fold nodules in children: preferable therapy. Int J Pediatr Otorhinolaryngol, 1999. 49 Suppl 1: p. S303-6. 35. Moran, M. and A. Pentz, Otolaryngologists opinions of voice therapy for vocal nodules in children. Language Speech and Hearing Services in Schools, 1987(18): p. 172-178. 36. Benjamin, B. and G. Croxson, Vocal nodules in children. Ann Otol Rhinol Laryngol, 1987. 96(5): p. 530-3. 37. Gray, S.D., E. Hammond, and D.F. Hanson, Benign pathologic responses of the larynx. Ann Otol Rhinol Laryngol, 1995. 104(1): p. 13-8. 38. Wetmore, R., Management of Pediatric Voice Disorders. Arch Otolaryngol, 2005. 131: p. 72. 39. Wohl, D.L., Nonsurgical management of pediatric vocal fold nodules. Arch Otolaryngol Head Neck Surg, 2005. 131(1): p. 68-70; discussion 71-2. 40. Silverman, E.M., Incidence of chronic hoarseness among school-age children. J Speech Hear Disord, 1975. 40(2): p. 211-5. 41. McAllister, A., et al., Relations between voice range profiles and physiological and perceptual voice characteristics in ten-year-old children. J Voice, 1994. 8(3): p. 230-9. 42. Campisi, P., et al., MultiDimensional Voice Program analysis in children with vocal cord nodules. J Otolaryngol, 2000. 29(5): p. 302-8. 43. Shah, R.K., et al., Pediatric vocal nodules: correlation with perceptual voice analysis. Int J Pediatr Otorhinolaryngol, 2005. 69(7): p. 903-9. 44. Hersan, R. and M. Behlau, Behavioral management of pediatric dysphonia. Otolaryngol Clin North Am, 2000. 33(5): p. 1097-110. 45. Jiang, J.J. and I.R. Titze, Measurement of vocal fold intraglottal pressure and impact stress. J Voice, 1994. 8(2): p. 132-44. 46. Titze, I.R., Principles of Sound Production. 1994, Englewood Cliffs, N.J.: Prentice-Hall Inc. 47. Roy, N. and D.M. Bless, Personality traits and psychological factors in voice pathology: a foundation for future research. J Speech Lang Hear Res, 2000. 43(3): p. 737-48. 48. Roy, N., D.M. Bless, and D. Heisey, Personality and voice disorders: a multitrait-multidisorder analysis. J Voice, 2000. 14(4): p. 521-48. 49. Verdolini, K. and C.A. Rosen. Lessac-Madsen Resonant Voice Therapy-KIDS. in Seminar, Department of Speech and Hearing sciences. 2004. University of Washington, Seattle. 50. Verdolini-Marston, K., et al., Preliminary study of two methods of treatment for laryngeal nodules. J Voice, 1995. 9(1): p. 74-85. 51. Deal, R.E., B. McClain, and J.F. Sudderth, Identification, evaluation, therapy, and follow-up for children with vocal nodules in a public school setting. J Speech Hear Disord, 1976. 41(3): p. 390- 7. 52. Houghton, F., M. Gleeson, and K. Kelleher, The use of primary/national school absenteeism as a proxy retrospective child health status measure in an environmental pollution investigation. Public Health, 2003. 117(6): p. 417-23. 53. Hartnick, C.J., Pediatric Otolaryngologist treatment of Vocal Fold Nodules. 2008: Boston. 54. Boseley, M.E. and C.J. Hartnick, Assessing the outcome of surgery to correct velopharyngeal insufficiency with the pediatric voice outcomes survey. Int J Pediatr Otorhinolaryngol, 2004. 68(11): p. 1429-33.

Downloaded From: https://jamanetwork.com/ on 10/02/2021 55. Campbell, D.T. and D.W. Fiske, Convergent and discriminant validation by the multitrait- multimethod matrix. Psychological Bulletin, 1959. 56: p. 81-105. 56. Trochim, W.M. Research Methods Knowledge Base. 2000 [cited 2004. 57. Blumin, J.H., et al., The impact of gender and age on voice related quality of life in children: normative data. Int J Pediatr Otorhinolaryngol, 2008. 72(2): p. 229-34. 58. Zur, K.B., et al., Pediatric Voice Handicap Index (pVHI): a new tool for evaluating pediatric dysphonia. Int J Pediatr Otorhinolaryngol, 2007. 71(1): p. 77-82. 59. Hartnick, C.J., Validation of a pediatric voice quality-of-life instrument: the pediatric voice outcome survey. Arch Otolaryngol Head Neck Surg, 2002. 128(8): p. 919-22. 60. Bach, K.K., et al., Validity and reliability of the glottal function index. Arch Otolaryngol Head Neck Surg, 2005. 131(11): p. 961-4. 61. Samlan, R., Voice Analysis, in Otolaryngology-Head and Neck Surgery, C. Cummings, Editor. 2008, Elsevier. 62. Raat, H., et al., Reliability and validity of the short form of the child health questionnaire for parents (CHQ-PF28) in large random school based and general population samples. J Epidemiol Community Health, 2005. 59(1): p. 75-82. 63. Kempster, G.B., et al., Consensus auditory-perceptual evaluation of voice: development of a standardized clinical protocol. Am J Speech Lang Pathol, 2009. 18(2): p. 124-32. 64. Gerratt, B.R. and J. Kreiman, Measuring vocal quality with speech synthesis. J Acoust Soc Am, 2001. 110(5 Pt 1): p. 2560-6. 65. Kreiman, J. and B.R. Gerratt, Perception of aperiodicity in pathological voice. J Acoust Soc Am, 2005. 117(4 Pt 1): p. 2201-11. 66. Kreiman, J., B.R. Gerratt, and K. Precoda, Listener experience and perception of voice quality. J Speech Hear Res, 1990. 33(1): p. 103-15. 67. Kreiman, J., B.R. Gerratt, and M. Ito, When and why listeners disagree in voice quality assessment tasks. J Acoust Soc Am, 2007. 122(4): p. 2354-64. 68. Titze, I.R., The physics of small-amplitude oscillation of the vocal folds. J Acoust Soc Am, 1988. 83(4): p. 1536-52. 69. Verdolini, K., et al., Laryngeal adduction in resonant voice. J Voice, 1998. 12(3): p. 315-27. 70. Shama, K., A. Krishna, and N. Cholayya, Study of harmonics-to-noise ratio and critical band energy spectrum of speech as acoustic indicators of laryngeal and voice pathology. EURASIP Journal on Applied Processing, 2007. 1: p. 50. 71. Hartnick, C.J., M. Volk, and M. Cunningham, Establishing normative voice-related quality of life scores within the pediatric otolaryngology population. Arch Otolaryngol Head Neck Surg, 2003. 129(10): p. 1090-3. 72. Mirasola, K.L., et al., Self-reported voice-related quality of life in adolescents with paradoxical vocal fold dysfunction. J Voice, 2008. 22(3): p. 373-8. 73. Hartnick, C.J., D.A. Zurakowski, and K. Haver, Validation of the Pediatric Cough Questionnaire. ENT Journal, 2008. In press. 74. Hartnick, C.J., et al., Final validation of the Pediatric Tracheotomy Health Status Instrument (PTHSI). Otolaryngol Head Neck Surg, 2002. 126(3): p. 228-33. 75. Verdolini, K., Case study: Resonant Voice Therapy, in Voice Therapy: Clinical Studies, J.C. Stemple, Editor. 2000, Singular: San Diego. p. 46-62. 76. Peterson, K.L., et al., Comparison of aerodynamic and electroglottographic parameters in evaluating clinically relevant voicing patterns. Ann Otol Rhinol Laryngol, 1994. 103(5 Pt 1): p. 335-46. 77. Kelly, M.H. and S. Martin, Domain-general abilities applied to domain-specific tasks: sensitivity to probabilities in perception, cognition, and language. Lingua, 1994. 92: p. 105-140. 78. Aslin, R.N., J.R. Saffran, and E.L. Newport, Statistical Learning in linguistic and nonlinguistic domains, in The Emergence of Language, B. MacWhinney, Editor. 1999, Lawrence Erlbaum: Mahwah, N.J. 79. Gierut, J.A., Phonological complexity and language learnability. Am J Speech Lang Pathol, 2007. 16(1): p. 6-17. 80. Forrest, K., Are oral-motor exercises useful in the treatment of phonological/articulatory disorders? Semin Speech Lang, 2002. 23(1): p. 15-26.

Downloaded From: https://jamanetwork.com/ on 10/02/2021 81. Cohen, N.L., S.B. Waltzman, and S.G. Fisher, A prospective, randomized study of cochlear implants. The Department of Veterans Affairs Cochlear Implant Study Group. N Engl J Med, 1993. 328(4): p. 233-7. 82. Larson, V.D., et al., Efficacy of 3 commonly used hearing aid circuits: A crossover trial. NIDCD/VA Hearing Aid Clinical Trial Group. JAMA, 2000. 284(14): p. 1806-13. 83. Bratt, G.W., NIDCD/VA hearing aid study. J Am Acad Audiol, 2007. 18(4): p. 272-3. 84. Verdolini, K., ed. Resonant voice therapy . Voice therapy: Clinical studies, ed. J.C. Stemple. 2000, Singular: San Diego. 85. Piaget, J., Psychology of Intelligence. 1950: Taylor and Francis, elibrary, 2003. 86. Moher, D., K.F. Schulz, and D. Altman, The CONSORT Statement: revised recommendations for improving the quality of reports of parallel-group randomized trials 2001. Explore (NY), 2005. 1(1): p. 40-5. 87. Plint, A.C., et al., Does the CONSORT checklist improve the quality of reports of randomised controlled trials? A systematic review. Med J Aust, 2006. 185(5): p. 263-7. 88. Hartnick, C.J. and S. Zeitels, Pediatric Videostroboscopy. Int J Ped Otolaryng, 2005. 69: p. 215- 219. 89. NIH Revitalization Act of 1993, in PL 103-43. 90. Pedersen, M., A. Beranova, and S. Moller, Dysphonia: medical treatment and a medical voice hygiene advice approach. A prospective randomised pilot study. Eur Arch Otorhinolaryngol, 2004. 261(6): p. 312-5. 91. Verdolini-Marston, K., M. Sandage, and I.R. Titze, Effect of hydration treatments on laryngeal nodules and polyps and related voice measures. J Voice, 1994. 8(1): p. 30-47. 92. Gottlob, I., M. Awan, and F. Proudlock, The role of compliance in 2 vs 6 hours of patching in children with amblyopia. Arch Ophthalmol, 2004. 122(3): p. 422-3; author reply 424-5. 93. Wulf, G., B. Lauterbach, and T. Toole, The learning advantages of an external focus of attention in golf. Res Q Exerc Sport, 1999. 70(2): p. 120-6. 94. Wulf, G., et al., Attentional focus in complex skill learning. Res Q Exerc Sport, 2000. 71(3): p. 229-39. 95. Stemple, J.C., et al., Efficacy of vocal function exercises as a method of improving voice production. J Voice, 1994. 8(3): p. 271-8. 96. Varni, J.W., C.A. Limbers, and T.M. Burwinkle, Parent proxy-report of their children's health- related quality of life: an analysis of 13,878 parents' reliability and validity across age subgroups using the PedsQL 4.0 Generic Core Scales. Health Qual Life Outcomes, 2007. 5: p. 2. 97. Rothenberg, M., Interpolating subglottal pressure from oral pressure. J Speech Hear Disord, 1982. 47(2): p. 219-23. 98. Smitheran, J.R. and T.J. Hixon, A clinical method for estimating laryngeal airway resistance during vowel production. J Speech Hear Disord, 1981. 46(2): p. 138-46. 99. Verdolini, K., et al., Investigation of vocal fold impact stress in human subjects. J Voice, 1999. 13(2): p. 184-202. 100. Holmberg, E.B., et al., Aerodynamic and acoustic voice measurements of patients with vocal nodules: variation in baseline and changes across voice therapy. J Voice, 2003. 17(3): p. 269-82. 101. Holmberg, E.B., et al., Efficacy of a behaviorally based voice therapy protocol for vocal nodules. J Voice, 2001. 15(3): p. 395-412.

The principal investigator (CJH) wishes to acknowledge the hard work in the preparation of the proposal (specifically regarding the background work sections and the literature review with regards to identifying various models for study design) of two of his fellows, Dr. Mark Boseley and Dr. Jennifer Shin.

Downloaded From: https://jamanetwork.com/ on 10/02/2021 Statistical Analysis Plan of

“A Randomized Controlled Trial of Voice Therapy on Children with Vocal Nodules”

A. Hypothesis

The primary objective of this study is to compare the impact of voice therapy versus office instructions alone on changes in the validated Pediatric Voice-Related Quality of Life Instrument (PVRQOL) in children age 6-10 years old with apparent vocal fold nodules.

The secondary objective is to assess the agreement between the primary outcome measure, the parent-indicated PVRQOL score voice quality rating with phonation threshold pressure, harmonics-to- noise ratio, and nodule ratings.

The tertiary objective is to examine mechanisms for possible benefits shown by voice therapy or the control intervention, including patient demographics, disease characteristics, voicing modality, and patient compliance.

B. Interim monitoring rule

DSMB requested one interim analysis in addition to the planned end-of-study analysis. We use approximate O’Brien-Fleming boundaries and control the overall Type I error probability to be 0.05. If these occur at times when 50% of the study information has been accumulated, the alpha levels for these tests will be alpha equals 0.003 (interim analysis), and alpha equals 0.0470 (final look). We will not correct for investigation-wide alpha error across remaining secondary measures, as they are considered exploratory. The monitoring boundaries allow early stopping for rejection of the null or the alternative hypothesis (futility). East 5.2 was used to estimate the sample size based on these parameters. The target sample size, after adjustment for the interim monitoring rule, is not changed.

C. Primary outcome analysis

C.1 Primary analysis for primary outcome

This analysis will use intention to treat principles, i.e. all randomized patients will be included in the analysis even if their treatment was discontinued early. Patients will continue to be followed for the 12- week follow-up period if treatment is discontinued. The t-test for two independent samples will be used to compare mean changes from baseline in PVRQOL scores in the two treatment groups. The rationale for the use of the t-test is that the study was powered based on clinical data using this statistic. If a patient does not complete follow-up, then multiple imputation procedures will be used, i.e. a regression model will be developed based on patients with complete data to identify patient characteristics which are associated with the change in PVQOL. The missing data are filled in multiple times to generate multiple complete data sets. The multiple complete data sets are analyzed using t-test. The results from the multiple complete data sets are combined to produce inferential results.

C.2 Secondary analysis for primary outcome

Mixed-model Analysis of Variance (ANOVA) will be used, with time and treatment group as independent variables, to model the set of responses over time and treatment differences will be assessed by the time by treatment interaction. When missing is non-ignorable, Pattern mixture models will be used to account for missing data.

C.3 Non-parametric analysis as needed

Downloaded From: https://jamanetwork.com/ on 10/02/2021 If initial data analysis shows that PVRQOL scores do not follow a normal distribution, then the data is analyzed with non-parametric methods instead. The differences between baseline measurements and 12- week measurements are calculated, and the medians compared. The null hypotheses are then tested using the Wilcoxon rank sum test. Two analyses will be done. The first will assign the highest (worst) rank for each patient who does not complete follow-up. The second analysis will be based on patients who complete follow-up.

D. Secondary outcome measures

Two parallel analyses for the primary outcome measure, t-tests for two independent samples and mixed-model ANOVAs as described for the PVRQOL will be done for the following:

Nodule size

Confidence level of overall judgment of CAPE-V

Phonation threshold pressure

Harmonics to noise ratio

Nodule Contour diagnosed, overall judgment of CAPE-V will be checked the treatment difference at baseline and last follow up by using Chi square test.

Additional analyses about possible mechanisms mediating changes in the primary outcome

We will conduct several additional analyses of the primary outcome measure.

1. We will compare mean changes among those participants who completed the full treatment protocol. 2. We will conduct a multiple linear regression analysis to evaluate the magnitude of the treatment effects after adjustment for participant characteristics such as age, gender, and baseline vocal nodule grade, harmful vocal habits, degree of resonant voice, and gastroenterology symptomatology score. 3. We will conduct a longitudinal analysis using mixed-model techniques comparing the trend of PVRQOL across the different treatment arms (interaction between treatment group and time) after adjusting all the covariates listed in (4). Model selections will choose the appropriate random terms and variance-covariance structure. 4. We will evaluate the responses and characteristics of those participants who do not complete treatment and compare to those who do complete treatment. 5. Withdrawal rates from study and from treatment will be compared across treatment groups using Fisher’s exact test. Reasons for withdrawal will be described for each treatment arm. 6. Compliance data will be described for each treatment arm using summary statistics.

The primary analysis of primary outcome will be performed twice, interim (50% study subjects completed the study) and final look. All the others will be done at final look.

F. Progress report

It will be distributed to NIDCD office, DSMB members and Hines study team annually. The study chair’s office personnel will receive a copy that does not include treatment outcome tables.

The structure of tables is listed as following.

Downloaded From: https://jamanetwork.com/ on 10/02/2021

1. Accrual monthly history & target by site

VOICE THERAPY STUDY ACCRUAL BY MONTH YEAR 2011

Participating Site Jan Feb March April May June July Aug Sep Oct Nov Dec MEEI MCW DUM Total Cum. Total

Cum. Target % of target

2. Randomization activity

Table 2a Included and Excluded Subjects by Site Screened and consented Randomized Excluded N N % N % Site MEEI MCW DUM Total

Downloaded From: https://jamanetwork.com/ on 10/02/2021 Table 2b Reasons for Screening Exclusions by Site Site MEEI MCW DUM Reason N % N % N % Total Age < 6 or >10 years Baseline PVRQOL scores are > 87.5 Dysphonia duration prior to randomization <12 weeks Hearing in better ear 35 dB or lower Children who cannot tolerate an awake video-stroboscopic examination Children with non-standard nodules or masses that are not unequivocally nodules Yellow or blue color suggestive of a mucous- or fluid-filled component Pink friable lesions at the posterior aspect of the true vocal folds Watery edematous expansion without a distinct protrusive mass Children who have previously received voice therapy, except articulation or speech therapy History of developmental delay History or presence of cognitive disorder History or presence of behavioral disorder History or presence of neurological disorder History or presence of articulation disorder History or presence of language disorder Prior knowledge of voice therapy principles through previous interventions Caregivers inability to commit to the 8 week voice therapy schedule Children whose caregivers cannot complete the primary endpoint Caregivers who are non-English speaking Caregivers who are unable to commit to the 3-month follow-up schedule Caregivers feel that the diagnostic information gained by a planned examination under anesthesia is not valuable for the severity of presenting symptoms Children with significant confounders of voice-related quality of life Children with vocal fold paralysis Children with a neurologic disorder of the larynx (dystonias, tics, tremors, etc) Children with an acute upper respiratory tract infection Not willing to be randomized

Table 2c Randomization by Each Stratum Site Gender Age Treatment A(n) Treatment B (n) MEEI Boy 6-7 8-10 Girl 6-7 8-10 WCM Boy 6-7 8-10 Girl 6-7 8-10 DUM Boy 6-7 8-10 Girl 6-7 8-10

Downloaded From: https://jamanetwork.com/ on 10/02/2021 3. Case history of study subjects by treatment arm

Table 3a.1 Time of Voice Problem Being Noticed by Treatment Arm A B

N Mean StDev Min Max N Mean StDev Min Max p-value Time of voice problem being noticed

Table 3a.2 Voice Problem History by Treatment Arm Total A B p-

N N % N % value Voice problem being present more than 12 weeks Yes No How did voice problem begin Suddenly Gradually Other Vocal symptom (will list 10 most frequent None symptoms) Hoarse voice quality Effortful/strained speaking Pain in the throat Raspy/scratchy voice quality Voice tired easily Frequent dry throat Weak/breathy voice quality Nasal voice quality Frequent throat- clearing/coughing Voice too low Lump in throat sensation Voice too high Frequent sore throats Trouble speaking loud or soft Other Problem changed over time Improved Same Worse Voice worse during certain seasons No variation Fall Spring Winter Summer N/A Voice worse at certain times of day Yes No validation Surgery for voice therapy Yes No Other Voice Therapy Yes No

Downloaded From: https://jamanetwork.com/ on 10/02/2021 Table 3b Development History by Treatment Total A B

N N % N % p-value Movement Yes No Communication Yes No Eating Yes No Hearing Yes No

Table 3c Medication Usage by Treatment Arm Medication Total A B N % N %

Table 3d Environment by Treatment Arm A B

N Mean StDev Min Max N Mean StDev Min Max p-value Air conditioning or dry heat Mold or damp envionment Chemicals Cigarette smoke(secondary smoke) Dust Environmental pollution Other

Downloaded From: https://jamanetwork.com/ on 10/02/2021

Table 3e Family Environment by Treatment Arm Total A B

N N % N % p-value Siblings Yes No Parents living together Yes No Other family member has voice problem Yes No Home environment Quiet Moderate Noisy Noisy

Table 3f Other Information by Treatment Arm A B

N Mean StDev Min Max N Mean StDev Min Max p-value # of 8-oz glasses of water daily Month-breathing Sweating Dry environments Spicy, acidic food or beverage Aggressive throat clearing unrelated to eating or drinking Aggressive throat clearing related to eating or drinking Aggressive coughing unrelated to eating or drinking Aggressive coughing related to eating or drinking Aggressive voice use in play or otherwise

Downloaded From: https://jamanetwork.com/ on 10/02/2021

Table 3g Rating of Activities by Treatment Arm* A B

N Mean StDev Min Max N Mean StDev Min Max p-value Imitate Any diagnosed speech- language learning difficulties Talk loudly Yell or cheer Scream Other family members yell or cheer Cry Sing Perform vocally Participate in sports Note: * 0 = never; 1=very rarely; 2 = rarely; 3 = sometimes; 4 = often; 5 = always

4. Baseline characteristics Table 4a Demographics by Treatment Arm

Total A B N N % N % p-value Gender Male Female Ethnicity Hispanic or Latino origin Not of Latino or Hispanic origin Race American Indian or Alaskan Native Asian African-American White No response Native English language Non-English Native bilingual

Downloaded From: https://jamanetwork.com/ on 10/02/2021

Table 4b Baseline Vocal Nodules Diagnose by Treatment Arm

Total A B N N % N % p-value Nodule Size Small Moderate Laryngologist 1 Large No Small Moderate Laryngologist 2 Large No Small Moderate Laryngologist 3 Large No Nodule Contour Discrete Laryngologist 1 Sessile Discrete Laryngologist 2 Sessile Discrete Laryngologist 3 Sessile

Downloaded From: https://jamanetwork.com/ on 10/02/2021

Table 4c Baseline PVRQOL by Treatment Arm

A B p- N Mean StDev Min Max N Mean StDev Min Max value Has trouble speaking loudly or being heard in noisy situations. Runs out of air and needs to take frequent breaths when talking Doesn’t know what will come out when begins speaking Anxious or frustrated (because of his or her voice) Depressed (because of his or her voice) Has trouble using the telephone or speaking with friends in person Has trouble doing his or her job or schoolwork (because of his or her voice) Avoids going out socially (because of his or her voice) Has to repeat himself or herself to be understood has become less outgoing (because of his or her voice) PVRQOL Score

Downloaded From: https://jamanetwork.com/ on 10/02/2021

Table 4d Baseline CAPE-V by Treatment Arm Total A B N N % N % p-value Overall Severity Consistent Mildly Deviant Severely Deviant Intermittent Moderately Deviant

Roughness Consistent Mildly Deviant Severely Deviant Intermittent Moderately Deviant

Breathness Consistent Mildly Deviant Severely Deviant Intermittent Moderately Deviant

Strain Consistent Mildly Deviant Severely Deviant Intermittent Moderately Deviant

Pitch Consistent Mildly Deviant Severely Deviant Intermittent Moderately Deviant

Loudness Consistent Mildly Deviant Severely Deviant Intermittent Moderately Deviant

Downloaded From: https://jamanetwork.com/ on 10/02/2021

Table 4e Baseline Aerodynamic Analysis by Treatment Arm A B N Mean StDev Min Max N Mean StDev Min Max p-value Trial 1 Trial 2 Trial 3 Trial 4 Trial 5 SPL Mean Trial 1 Trial 2

Estimated Trial 3 Phonatory Trial 4 Subglottic Trial 5 Pressure (cm H2O) Mean Trial 1 Trial 2 Trial 3 Trial 4 Trial 5 Air Flow Rate (L/sec) Mean

Table 4f Baseline Acoustic Recording by Treatment Arm A B N Mean StDev Min Max N Mean StDev Min Max p-value vowel Trial 1

Trial 2

Trial 3

Mean

Harmonics-to- Noise

5. Completeness of data collection

Table 5a Number of Completed Forms by Visit Form 0: Form 1: Form 2 : Form 3: Form 4: Form 5: Form 6: Form 7: Form 8:

N N N N N N N N N Visit 0 1 s-1

Downloaded From: https://jamanetwork.com/ on 10/02/2021 Form 0: Form 1: Form 2 : Form 3: Form 4: Form 5: Form 6: Form 7: Form 8:

N N N N N N N N N s-2 s-3 s-4 s-5 s-6 s-7 s-8 2 3

Form18: Form19: Form20: Form 10: Form 11: Form 13: Form 14: Form15: Form 16: Form 17: N N N N N N N N N N Visit 0 1 s-1 s-2 s-3 s-4 s-5 s-6 s-7 s-8 2 3

Downloaded From: https://jamanetwork.com/ on 10/02/2021

Table 5b Missing Baseline Forms by Site Form 8: Audiology Form 2 Form :Vocal 3: Form Form 5: Form 6: Total Form 0: Form 1: nodules CAPE- 4: Acoustic Case Form 7: Randomized Demographic Randomization diagnoses V PTP recording history PVRQOL N N % N % N % N % N % N % N % N % N % Site MEEI

MCW

DUM

Total

Table 5c Missing Visit 2 Forms by Site

Form 10 Form11: Form 13: Form14:

Form 2: Form 3 : Form 4: Form 5: Form 7: N % N % N % N % N % N % N % N % N % Site

MEEI

MCW

DUM

Total

Downloaded From: https://jamanetwork.com/ on 10/02/2021

Table 5d Missing Visit 3 Forms by Site

Form 10 Form11: Form 13: Form14:

Form 2: Form 3 : Form 4: Form 5: Form 7: N % N % N % N % N % N % N % N % N % Site

MEEI

MCW

DUM

Total

Table 5e Missing Forms by Session by Site

Session 6: Session 7: Session 8:

Session 1: Session 2: Session 3: Session 4: Session 5: N % N % N % N % N % N % N % N % Site

MEEI

MCW

DUM

Total

Downloaded From: https://jamanetwork.com/ on 10/02/2021

6. Protocol deviation Table 6a Protocol Deviation by Site

Deviation Inclusion/Exclusion Procedure not SAE not reported appropriately to violated completed IRB Other Total N % N % N % N % N Site

MEEI

MCW

DUM

Total

Table 6b Details of Protocol Deviations Site Subject Number Treatment Arm Type of Deviation Description of Deviation Date

Downloaded From: https://jamanetwork.com/ on 10/02/2021 Site Subject Number Treatment Arm Type of Deviation Description of Deviation Date

7. Withdrawals Table 7a Current Status of Randomized Subjects A B Total N % N % N % Current participating Completed study

Withdrew from Follow-up Withdrew from treatment but continued follow-up

Table 7b.1 Complete Visits Total A B

N N % N % Visit

Downloaded From: https://jamanetwork.com/ on 10/02/2021

Table 7b.2 Complete Sessions Session 1 2 3 4 5 6 7 8

N N N N N N N N

Table 7c Current Status of Randomized Subjects Visit

1.0 2.0 3.0 Withdraw from follow-up N % N % N %

Table 7d Withdrawals from Study Treatment Session 1 2 3 4 5 6 7 8 Total N % N % N % N % N % N % N % N % N Withdraw from treatment

Table 7e Withdrawal Reason by Treatment Arm

A B N % N % P-value Lost to follow up Withdrew consent due to lack of benefit Withdrew consent due to AEs Withdrew consent due to other reason

Table 7f Withdrawal Reason by Site MEEI MCW DUM

N % N % N % Lost to follow up Withdrew consent due to lack of benefit Withdrew consent due to AEs Withdrew consent due to other reason

Downloaded From: https://jamanetwork.com/ on 10/02/2021

Table 7g Missed Visits by Treatment

Total A B N N % N % Visit 1 2 3 Session 1 2 3 4 5 6 7 8

Table 7h Reasons for Missed Visit N % Reason for Missed Visit

Unable to contact

Missed this one

Other

Total

8. Compliance and Fidelity check

Table 8a Compliance by Treatment Arm A B

N Mean StDev Min Max N Mean StDev Min Max p-value Compliance

Downloaded From: https://jamanetwork.com/ on 10/02/2021

Table 8b Clinician Fidelity Check with Therapy Procedure Reviewer Reviewer Reviewer Total 1 2 3 N N % N % N % Does the clinician follow the outlined treatment plan Yes and address goals for this particular session No number appropriately, according to the child’s skill level?

Does the clinician use Biomechanical instructions? Yes No Does the clinician use Extrinsic rewards for good Yes vocal behavior? No Does the clinician refer to this treatment program Yes and related activities as “voice therapy?” No Does the clinician provide additional homework Yes other than work in line with the principles of the No treatment program? Not observable Summarize expectations of the child prior to the Yes next session not observable No

Reviewer 1 Reviewer 2 Reviewer 3 St St St N Mean Dev Min Max N Mean Dev Min Max N Mean Dev Min Max Engage the child with appropriate enthusiasm Interact with the child in an age- appropriate manner Provide instructions using language specific to the treatment program where applicable Summarize expectations of the child prior to the next session Activities are functional Goals are perceptually salient Numerous repetitions of target behaviors are used

Downloaded From: https://jamanetwork.com/ on 10/02/2021 Variable practice is incorporated Rate—ask the child to evaluate vocal productions Show—model targets and non- targets for the child Better—ask the child to use negative practice Mold—use physical manipulations to improve the child’s performance

Table 8c Clinician Fidelity Check with Hygiene Procedure Reviewer Reviewer Reviewer Total 1 2 3 N N % N % N % Does the clinician follow the outlined treatment plan Yes and address goals for this particular session No number appropriately, according to the child’s skill level?

Reviewer 1 Reviewer 2 Reviewer 3 N Mean StDev Min Max N Mean StDev Min Max N Mean StDev Min Max Engage the child with appropriate enthusiasm Interact with the child in an age- appropriate manner

Downloaded From: https://jamanetwork.com/ on 10/02/2021 Provide instructions using language specific to the treatment program where applicable Summarize expectations of the child prior to the next session Activities are functional Goals are perceptually salient Numerous repetitions of target behaviors are used Variable practice is incorporated Rate—ask the child to evaluate vocal productions Show—model targets and non-targets for the child Better—ask the child to use negative practice Mold—use physical manipulations to improve the child’s performance

9. Outcome measures at end point or changes from baseline (p-values will be reported only at interim and final looks.)

Table 9a Vocal Nodules Diagnosis by Treatment Arm

A Total B p- N N Mean Std N Mean Std value

Downloaded From: https://jamanetwork.com/ on 10/02/2021 Nodule Size Baseline Visit 3 Laryngologist 1 Change Baseline Laryngologist 2 Visit 3 Change Baseline Laryngologist 3 Visit 3 Change Baseline Average Visit 3 Change P- N N % N % value Nodule Contour Discrete Laryngologist 1 Sessile Discrete Laryngologist 2 Sessile Discrete Laryngologist 3 Sessile

Table 9b Each item of PVRQOL by Treatment Arm A B p- N Mean StDev Min Max N Mean StDev Min Max value Has trouble speaking loudly or being heard in noisy situations. Runs out of air and needs to take frequent breaths when talking Doesn’t know what will come out when begins speaking Anxious or frustrated (because of his or her voice) Depressed (because of his or her voice) Has trouble using the telephone or speaking with friends in person Has trouble doing his or her job or schoolwork (because of his or her voice) Avoids going out socially (because of his or her voice)

Downloaded From: https://jamanetwork.com/ on 10/02/2021 Has to repeat himself or herself to be understood has become less outgoing (because of his or her voice)

Table 9c Change of PVRQOL by Treatment Arm

A B N Mean StDev Min Max N Mean StDev Min Max p-value Baseline Visit 3

PVRQOL Change

Table 9d Visit 3 CAPE-V by Treatment Arm Total A B N N % N % p-value Overall Severity Consistent Mildly Deviant Severely Deviant Intermittent Moderately Deviant

Roughness Consistent Mildly Deviant Severely Deviant Intermittent Moderately Deviant

Breathness Consistent Mildly Deviant Severely Deviant Intermittent Moderately Deviant

Strain Consistent Mildly Deviant Severely Deviant Intermittent Moderately Deviant

Pitch Consistent

Downloaded From: https://jamanetwork.com/ on 10/02/2021 Mildly Deviant Severely Deviant Intermittent Moderately Deviant

Loudness Consistent Mildly Deviant Severely Deviant Intermittent Moderately Deviant

Table 9e Overall judgment of CAPE-V by Treatment Arm Total A B N N % N % p-value Overall judgment Yes Before is better No

Table 9f Confident level of overall judgment of CAPE-V by Treatment Arm A B N Mean StDev Min Max N Mean StDev Min Max p-value Confidence level

Table 9g Aerodynamic Analysis by Treatment Arm A B N Mean StDev Min Max N Mean StDev Min Max p-value Baseline Visit 3

SPL Change Estimated Baseline Phonatory Visit 3 Subglottic Pressure (cm H2O) Change Baseline Visit 3 Air Flow Rate (L/sec) Change

Table 9h Acoustic Recording by Treatment Arm

Downloaded From: https://jamanetwork.com/ on 10/02/2021 A B N Mean StDev Min Max N Mean StDev Min Max p-value Vowel Trial 1

Trial 2

Trial 3

Mean

Baseline Visit 3 Harmonics- to-Noise Change

10. Safety Table 10a Adverse Events by Treatment Arm A B Total Total

Adverse Unique Adverse Unique Events Subjects % Events Subjects % Adverse events Behavioral changes

Other Total

Table 10 b Study Related Adverse Events by Treatment Arm A B Total Total

Adverse Unique Adverse Unique Events Subjects % Events Subjects % Adverse events Behavioral changes

Other Total

Table 10 c Serious Adverse Events by Treatment Arm by Site Serious Adverse Events

A B N % N % Site MEEI MCW DUM Total

Downloaded From: https://jamanetwork.com/ on 10/02/2021

Table 10 d Serious Adverse Events by Treatment Arm by Reason Serious Adverse Events A B N % N %

Breach of patient confidentiality Unblinding of participants and parents Life threatening or fatal Total

Table 10 e Details of Serious Adverse Events Site: Subject: Treatment Arm: Date of Onset: Date of Report: Study Related: Reason(s) for SAE: Diagnosis: Resolved: SAE Narrative:

Downloaded From: https://jamanetwork.com/ on 10/02/2021 CONSORT FLOWCHART

Enrollment

Assessed for eligibility (n=148)

Excluded (n=34) • Not meeting inclusion criteria (n=33) • Declined to participate (n=0) • Other reasons (n=1) Randomized (n=114)

Allocation

Allocated to AIV (n=57) Allocated to MVA (n=57) • Received allocated intervention (n=47 ) • Received allocated intervention (n= 52)

Follow-Up

Withdrew from study before end of intervention (Visit 2)(n=5) • Lost to follow-up (n=2) Withdrew from study before end of intervention • Withdrew consent for lack of perceived (Visit 2) (n=9) benefit (n=1) • Lost to follow-up (n=2) • Withdrew consent for other reasons (n=2) • Withdrew consent for other reasons (n=7) Withdrew after intervention, before 3 month Withdrew after intervention, before 3 month follow up (Visit 3) (n=4) follow up (Visit 3) (n=1) • Lost to follow up (n=3) • Lost to follow-up (n=1) • Withdrew consent for other reasons (n=1) Excluded due to ineligibility (n=1) Excluded due to ineligibility (n=1) • Misrandomization (n=1) • Misrandomization (n=1)

Analysis

Included in primary outcome analysis (n=56) Included in primary outcome analysis (n=56) • Completed study (n=47) • Completed study (n=46) • Analysed with imputed values (n=9) • Analysed with imputed values (n=10) • Ineligible for Study (n=1) © 2017 American Medical Association. All rights reserved.

Downloaded From: https://jamanetwork.com/ on 10/02/2021 Checklist of items to include when reporting a randomized trial (56-58)

PAPER SECTION Reported And topic Item Description on page # TITLE & ABSTRACT 1 How participants were allocated to interventions (e.g., "random

allocation", "randomized", or "randomly assigned"). INTRODUCTION

Background 2 Scientific background and explanation of rationale. METHODS Eligibility criteria for participants and the settings and locations where the 3 Participants data were collected. Precise details of the interventions intended for each group and how and Interventions 4 when they were actually administered. Objectives 5 Specific objectives and hypotheses. Clearly defined primary and secondary outcome measures and, when Outcomes 6 applicable, any methods used to enhance the quality of measurements (e.g., multiple observations, training of assessors). How sample size was determined and, when applicable, explanation of Sample size 7 any interim analyses and stopping rules. Randomization -- Method used to generate the random allocation sequence, including 8 Sequence generation details of any restriction (e.g., blocking, stratification). Method used to implement the random allocation sequence (e.g., Randomization -- 9 numbered containers or central telephone), clarifying whether the Allocation concealment sequence was concealed until interventions were assigned. Randomization -- Who generated the allocation sequence, who enrolled participants, and 10 Implementation who assigned participants to their groups. Whether or not participants, those administering the interventions, and Blinding (masking) 11 those assessing the outcomes were blinded to group assignment. When relevant, how the success of blinding was evaluated. Statistical methods used to compare groups for primary outcome(s); Statistical methods 12 Methods for additional analyses, such as subgroup analyses and adjusted analyses. RESULTS Flow of participants through each stage (a diagram is strongly recommended). Specifically, for each group report the numbers of participants randomly assigned, receiving intended treatment, Participant flow 13 completing the study protocol, and analyzed for the primary outcome. Describe protocol deviations from study as planned, together with reasons. Recruitment 14 Dates defining the periods of recruitment and follow-up. Baseline data 15 Baseline demographic and clinical characteristics of each group. Number of participants (denominator) in each group included in each Numbers analyzed 16 analysis and whether the analysis was by "intention-to-treat" . State the results in absolute numbers when feasible (e.g., 10/20, not 50%). For each primary and secondary outcome, a summary of results for each Outcomes and estimation 17 group, and the estimated effect size and its precision (e.g., 95% confidence interval). Address multiplicity by reporting any other analyses performed, including Ancillary analyses 18 subgroup analyses and adjusted analyses, indicating those pre-specified and those exploratory. Adverse events 19 All important adverse events or side effects in each intervention group. DISCUSSION Interpretation of the results, taking into account study hypotheses, Interpretation 20 sources of potential bias or imprecision and the dangers associated with multiplicity of analyses and outcomes. Generalizability 21 Generalizability (external validity) of the trial findings. Overall evidence 22 General interpretation of the results in the context of current evidence.

Downloaded From: https://jamanetwork.com/ on 10/02/2021 eTable 1. Eligibility Criteria Inclusion Criteria 6-10 years of age Presence of vocal fold nodules (bilateral, localized, benign, superficial growths with protrusion on the medial surface of the true vocal folds at the junction of their anterior and medial thirds) Voice-related quality of life < 87.5 (0=worst; 100=best) Dysphonia duration > 12 weeks, to ensure that vocal dysfunction is chronic in nature Hearing in good ear better than 35 dB Agreement by informed consent from the caregivers, and informed assent from the child participant, with anticipated commitment to compliance throughout the follow-up period of one month Exclusion Criteria Vocal fold lesions that are not nodules (e.g., unilateral polyps, cysts, papilloma, granuloma) a Lesions with the following characteristics: (a) yellow or blue color suggestive of a mucous-or fluid- filled component, (b) pink friable lesions at the posterior aspect of the true vocal folds, (c) water edematous expansion without a distinct protrusive mass a Presence of vocal fold paralysis, irregular epithelium, potential malignancy, or airway obstruction requiring urgent or operative intervention a Prior history of having undergone voice therapy Inability to tolerate an awake video stroboscopic examination History of laryngeal surgery Children with whom voice therapy would be problematic: (a) developmental delays, (b) cognitive disorder, (c) behavioral disorder, (d) neurologic disorder, (e) articulation disorder, (f) phonological disorder, (g) language disorder, (h) language disorder, (i) expressive or receptive language delay, (j) specific language impairment, (l) central auditory processing disorder, (m) fluency disorder Prior knowledge of voice therapy principles through previous interventions (with the exception of articulation or speech therapy) Inability to commit to 12 weeks of therapy Inability of caregivers to complete PVRQOL, either due to time constraints or because of a language barrier a Defining characteristics of these exclusion criteria were determined and corroborated by evaluations of three senior otolaryngologists of each participant’s videostroboscopic examination.

Downloaded From: https://jamanetwork.com/ on 10/02/2021 eTable 2. Description of treatment protocols Session Description of Adventures in Voice (AIV) Protocol Session 1 Describe basic fundamentals of vocal anatomy and normal voice production as it relates to vocal fold nodule pathology. Depictions of smooth vocal fold edges necessary for optimal voice production, as well as normal vibratory patterns, should be included. Also provide explanation of phonotrauma and its effect on the vocal folds. Provide vocal hygiene education and establish personalized vocal hygiene goals. Voice Super Power 1: water and steam usage. Discuss the importance of systemic and topical hydration on vocal production. Participants may draw, color, cut pictures from magazines, select images, or otherwise depict this healing power. Identify healthy vs. unhealthy sounds. Modify loud voice behaviors with softer voice use and provide other acceptable alternatives (e.g., non-voicing body gestures) Voice Super Power 2: no voice, quiet voice, inside voice Voice Super Power 3: loud body, but not loud voice Session 2 Train participants to recognize and apply resonant voice, focusing on easy oral vibrations and forward placement Voice Super Power 4: easy, forward vibrations Voice Super Power 5: secret signal (resonant forward placement of /m/ in hummed yes/no response) Session 3 Review and apply Voice Super Powers previously trained. Identify self-correction strategies Voice Super Power 6: pull-outs (“pulling-out” out of “old” vocal patterns into “new” vocal patterns) Session 4 Review and apply Voice Super Powers previously trained. Identify appropriate ways to project voice to create loud voice productions Voice Super Power 7: loud clear voice Session 5 Review and apply Voice Super Powers previously trained Review and further apply proper loud voice use Session 6 Review and apply Voice Super Powers previously trained Train participant to produce loud voice with background noise Voice Super Power 8: loud voice in noise Session 7 Review and apply Voice Super Powers previously trained Train participant to produce loud voice in context of various emotions Voice Super Power 9: Let go! Freedom! Session 8 Apply voice use and Voice Super Powers in real-life Apply specific voice use and individualized activities for each individual participant, chosen from a closed set Discuss application and troubleshooting of Voice Super Powers outside of the clinical setting Voice Super Power 10: “If I get sick or hoarse, I can …”

Downloaded From: https://jamanetwork.com/ on 10/02/2021

Session Description of My Voice Adventure (MVA) Protocol Session 1 Explain normal vocal mechanism and voice production Set daily hydration goals, based on each child’s individual weight Train the participant to complete homework sheets Establish a reward system for returning homework sheets at every session Session 2 Collect homework sheets Discuss abusive voice behaviors Address ways to avoid abusive voice behaviors Session 3 Collect homework sheets Discuss potentially abusive voice situations and environments Address ways to avoid or adapt to abusive voice situations Session 4 Collect homework sheets Discuss “adequate” versus “inadequate” vocal quality Session 5 Collect homework sheets Review potentially abusive voice behaviors and abusive voice situations Session 6 Collect homework sheets Review importance of hydration Review “adequate” versus “inadequate” vocal quality

Downloaded From: https://jamanetwork.com/ on 10/02/2021 eFigure 1.

Downloaded From: https://jamanetwork.com/ on 10/02/2021