ARTICLE IN PRESS

Barriers to Voice Therapy in Dysphonic Children

*,†Anne F. Hseu, *Grant Spencer, *Geralyn Woodnorth, †Sydney Kagan, *,†Kosuke Kawai, and *,†Roger C. Nuss,
*yBoston, Massachusetts

Summary: Background. The most common etiologies of dysphonia in the pediatric population are vocal fold
nodules and muscle tension dysphonia. Vocal therapy is the first line treatment for these disorders in children.
Despite this, not all children undergo therapy. The goal of this study is to examine how factors such as patient
demographics and parental perceptions differ between children that choose to undergo or not to undergo voice
therapy.
Methods. A retrospective review was conducted of all pediatric patients seen at a tertiary voice clinic between
January 2014 and December 2017. Patients were included if diagnosed with vocal fold nodules and/or muscle ten-
sion dysphonia. Patients were divided into groups of children that received voice therapy at our institution and
those that did not. Data include demographics, Consensus Auditory-Perceptual Evaluation of Voice (CAPE-V)
scores and pediatric Voice Handicap Index (pVHI) scores. Distance to therapy site was approximated using
patient zip codes.
Results. Three hundred and forty-six children were included, 224 (65%) boys and 122 (35%) girls. In the 2 years
following initial diagnosis, 74 (21%) children participated in voice therapy at our institution. Patients who under-
went voice therapy were older than those who did not (mean age: 9.1 [SD 3.5] vs 7.6 [SD 3.8] years; P = 0.004).
Patients who received voice therapy were more likely to live closer to the therapy site (mean distance: 15.5 [SD
13.0] vs 24.3 [SD 23.9] miles; P< 0.001). Likelihood of receiving voice therapy did not differ by gender or health
insurance status (private vs public). Patients who underwent voice therapy had significantly greater CAPE-V
Overall Severity scores than those who did not (mean score: 44.6 [SD 19.4] vs 37.4 [SD 18.0]; P = 0.003). Higher
CAPE-V Strain scores were associated with increased likelihood of voice therapy. pVHI scores did not differ
between the two groups.
Conclusion. Older age, shorter distance to therapy site, and increased CAPE-V Overall Severity and Strain
scores were associated with higher likelihood of receiving voice therapy. Gender, insurance status, and pVHI
scores did not affect likelihood of receiving voice therapy. Patients may primarily consider ease of access and
necessity of treatment when considering voice therapy.
Key Words: Dysphonia−Children or pediatrics−Voice−Voice disorders−Vocal cord nodules−Voice therapy.
Abbreviations: MTD, Muscle Tension Dysphonia−CAPE-V, Consensus Auditory-Perceptual Evaluation of
Voice−pVHI, Pediatric Voice Handicap Index.

BACKGROUND In the pediatric group, voice therapy is the first line treat-
Dysphonia in the pediatric population is not an infrequent ment for vocal fold nodules and muscle tension dysphonia.
occurrence. Studies have shown a prevalence as high as 38% Although surgical intervention with careful microflap exci-
in school-age children.1,2 The most common etiologies of sion of nodules is more common in adults, this treatment is
dysphonia in children are vocal fold nodules.2,3 These typically not recommended in children due to the high risk
benign lesions are often related to phonotrauma. Muscle of nodule recurrence. Additionally, the natural history of
tension dysphonia (MTD), or vocal strain, is also a common nodules in children would suggest that improved voice out-
diagnosis that can occur both primarily or secondarily. In comes may be obtained with time, growth of the larynx,
primary cases, MTD can present following an acute illness, and changes in vocal demands alone. Therapy for nodules
intubation, or episode of increased vocal demand. In sec- has been found to have good benefit in children, though, as
ondary cases, MTD can accompany diagnoses such as vocal it may improve voice quality and diminish harmful compen-
fold nodules, vocal fold hypomobility or immobility, or satory voicing.4 Despite this, not all children diagnosed with
other benign vocal fold lesions. nodules or MTD receive therapy.
The goal of this study is to examine the barriers to voice
therapy in the pediatric population. We hope to elucidate
how factors such as patient demographics and parental per-
Accepted for publication January 8, 2021.
Declaration of interest: none
ceptions of vocal quality differ between children that choose
Role of the funding source: no funding to disclose. to undergo or not to undergo voice therapy.
From the *Department of Otolaryngology and Communication Enhancement,
Boston Children’s Hospital, Boston, Massachusetts; and the yDepartment of Otolar-
yngology, Harvard Medical School, Boston, Massachusetts.
Address correspondence and reprint requests to: Dr. Anne F. Hseu, 300 Longwood
Ave, BCH-3129, Boston MA, 02115 E-mail: [email protected]
METHODS
Journal of Voice, Vol. &&, No. &&, pp. &&−&& An IRB-approved retrospective review was conducted of
0892-1997
© 2021 The Voice Foundation. Published by Elsevier Inc. All rights reserved.
pediatric patients (age 2 years − 18 years) seen for initial
https://doi.org/10.1016/j.jvoice.2021.01.008 evaluation at a tertiary voice clinic within Boston Children’s
 ARTICLE IN PRESS
2 Journal of Voice, Vol. &&, No. &&, 2021

Hospital’s Department of Otolaryngology and Communica- and the Hosmer-Lemeshow goodness of fit statistic. Age,
tion Enhancement between January 2014 and December proximity to therapy site, and CAPE-V scores were
2017. As part of their initial visit with the voice clinic, all included in the multivariable regression model. Odds
patients were evaluated by a pediatric otolaryngologist as ratio (OR) and 95% confidence interval (CI) were esti-
well as a speech language pathologist (SLP). Consensus mated. The correlation between CAPE-V and pVHI
Auditory-Perceptual Evaluation of Voice (CAPE-V) scores scores was assessed using Pearson correlation coefficient.
were recorded by the SLP during the patient’s initial evalua- All analyses were performed using SAS version 9.4 (SAS
tion. The CAPE-V tool aims to describe the severity of the Institute, Cary North Carolina).
patient’s dysphonia by rating six attributes of the patient’s
voice (overall severity, roughness, breathiness, strain, pitch,
RESULTS
and loudness) on a 0 − 100 scale, with 0 being considered
346 children were included, 224 (65%) boys and 122
within normal limits.5 While measurements of intra- and
(35%) girls. One hundred and sixty patients had a diag-
inter-rater reliability were not obtained for this study,
nosis of benign vocal fold lesion(s) and secondary MTD,
CAPE-V scores were obtained by one of five SLPs, all with
while 128 patients only had a diagnosis of vocal fold
specialty training in pediatric voice. A pediatric Voice
lesion(s). Fifty-eight patients had a diagnosis of primary
Handicap Index (pVHI) questionnaire was also electroni-
MTD. 289 (84%) patients had a private insurance, while
cally administered prior to each visit. The pVHI is a 23-item
57 patients (16%) patients had only public insurance. In
questionnaire that assesses functional, physical, and emo-
the two years following initial diagnosis, 74 (21%) chil-
tional consequences of pediatric voice disorders on children’
dren participated in voice therapy at our institution.
lives. This voice related quality of life questionnaire rates
Patients who underwent voice therapy were older than
parental perceptions of the child’s voice disturbance on a 92
those who did not (mean age: 9.1 [SD 3.5] vs 7.6 [SD 3.8]
point scale.6
years; P = 0.004; Table 1). Patients who received voice
Patients were initially identified through a departmental
therapy lived closer to the therapy site than those who
voice clinic database and included if they had been diag-
did not receive voice therapy (mean distance: 15.5 [SD
nosed with vocal fold nodules and/or muscle tension dys-
13.0] vs 24.3 [SD 23.9] miles; P < 0.001). Patients who
phonia. Patients were divided into groups of children that
underwent voice therapy had significantly higher CAPE-
received voice therapy at our institution and those that did
V Overall Severity scores than those who did not (mean
not. Patients were excluded from this study if they had voice
score: 44.6 [SD 19.4] vs 37.4 [SD 18.0]; P = 0.003). Like-
clinic visits prior to the start of the study’s date range, if
lihood of receiving voice therapy did not differ by gender
they did not meet the study’s age range at their first visit, or
or health insurance status (private vs public). Older age,
if they did not receive an SLP evaluation as part of their
proximity to therapy site, and CAPE-V score remained
visit.
significantly associated with likelihood of receiving voice
Data including demographics, date of initial presentation,
therapy in the multivariable regression model (Table 2).
date(s) of voice therapy, distance to voice therapy site,
Children older than 8 years of age were twice as likely to
health insurance type (public or private), CAPE-V scores,
receive voice therapy than children younger than 8 years
and pVHI scores were analyzed. Distance to therapy site
of age (adjusted OR 2.09; 95% CI: 1.21, 3.58). Children
was approximated using a centralized location for each zip
who lived close to the therapy site (<20 km) were more
code. Each patient’s insurance type was determined as either
likely to receive voice therapy than those who live far
public or private according to which insurances were listed
from the therapy site (>20 km; adjusted OR 1.91; 95%
at their time of visit. If a patient had both a public and pri-
CI: 1.08, 3.40).
vate insurance listed, then the insurance type was considered
private. Patient insurance type was considered public if the
patient had only public insurances and no private insurances Gender differences
listed. Males were more likely than females to present for eval-
To compare the distribution of characteristics between uation at the voice clinic. Accounting for this difference
patients who underwent voice therapy and those who did in our overall population, we did not see any effect of
not, we used a Chi-square test for categorical variables gender on a patient’s likelihood to receive or defer voice
and a t-test for normally distributed continuous variables. therapy. Males were more likely to present at a younger
For non-normally distributed continuous variables, we age and with more severe dysphonia than females. The
used Wilcoxon rank-sum test. A logistic regression model average age at presentation for males was 7.6 years old,
was used to examine predictors of receiving voice therapy. while average age at presentation for females was
We examined whether age, gender, insurance, proximity 8.5 years old. Males presented with higher CAPE-V
to therapy site, CAPE-V overall severity, and pVHI are Severity and Loudness scores than females (40.4 [SD
predictors of receiving voice therapy. The multivariable 18.59] vs 36.0 [17.97], P= 0.04; 24.0 [SD 19.51] vs 19.1
regression model was built using a backward selection [SD 17.46], P= 0.04). No significant differences were
procedure, with P < 0.05 as the retention criterion. We found in either insurance type or mean distance to ther-
evaluated the fit of the model by examining the C-statistic apy site by gender.
 ARTICLE IN PRESS
Anne F. Hseu, et al Barriers to Voice Therapy 3

TABLE 1.
Patient Characteristics by Receipt of Voice Therapy
Received voice therapy (n=74) Did not receive voice therapy (n=272) P-value
Age (years) 9.1 (3.5) 7.6 (3.8) 0.004
Gender 0.60
Female 28 (37.8%) 94 (34.6%)
Male 46 (62.2%) 178 (65.4%)
Proximity to therapy site (km) 15.5 (13.0) 24.3 (23.9) <0.001
Health insurance 0.95
Private 62 (83.8%) 227 (83.5%)
Public 12 (16.2%) 45 (16.5%)
CAPE-V overall severity 44.6 (19.4) 37.4 (18.0) 0.003
pVHI 25.0 (18.4) 23.9 (14.8) 0.81
Diagnosis 0.43
Both vocal fold lesion(s) and MTD 39 (52.7%) 121 (44.5%)
Vocal fold lesion(s) 25 (33.8%) 103 (37.9%)
Primary MTD 10 (13.5%) 48 (17.6%)
Pediatric Voice Handicap Index (pVHI) scores were available in 87 patients.

pVHI
Complete pVHI data was collected for 25 voice therapy
recipients and 62 nonrecipients, for a total of 87 patients.
No significant differences were seen in total pVHI scores
between the voice therapy recipient and non-recipient
groups. No correlation was seen between mean age at first
presentation and pVHI. There was a moderate correlation
between total pVHI scores and CAPE-V Overall Severity
scores (r=0.54; 95% CI: 0.36, 0.67; P<0.001; Figure 1), indi-
cating a moderate provider-parent agreement in regards to
the severity of the patient’s voice disturbance.

CAPE-V
Patients who underwent voice therapy had significantly
greater CAPE-V Overall Severity scores than those who did
not (mean score: 44.6 [SD 19.4] vs 37.4 [SD 18.0];
P = 0.003). Voice therapy patients were also rated to have
FIGURE 1. Consensus Auditory-Perceptual Evaluation of Voice
higher degrees of strain at first evaluation (43.8 [SD 19.0] vs
(CAPE-V) score and pediatric Voice Handicap Index (pVHI) score
35.7 [SD 18.0], P < 0.001), though no significant differences
(r = 0.54; 95% CI: 0.36, 0.67; P < 0.001).
between the two groups were seen in ratings of roughness,
breathiness, pitch, or loudness.

TABLE 2.
Predictors of Receiving Voice Therapy
Unadjusted OR (95% CI) P-value Adjusted OR (95% CI) P-value
Older age (>8 yr vs. < 8 yr) 1.96 (1.17, 3.30) 0.01 2.09 (1.21, 3.58) 0.01
Shorter distance to therapy site (< 20 km vs. > 20 km) 1.77 (1.01, 3.07) 0.04 1.91 (1.08, 3.40) 0.03
CAPE-V overall severity (>50 vs. <50) 1.72 (1.01, 2.94) 0.04 2.02 (1.15, 3.52) 0.01
Gender (female vs. male) 1.15 (0.68, 1.96) 0.60 −
Insurance (private vs. public) 1.02 (0.51, 2.05) 0.95 −
pVHI 1.00 (0.98, 1.03) 0.81 −
Adjusted Odds Ratio (OR) and 95% confidence interval (CI) were based on the multivariable logistic regression model that includes age, proximity to therapy
site, and CAPE-V overall severity score. C-statistic was 0.64 in the multivariable model.
 ARTICLE IN PRESS
4 Journal of Voice, Vol. &&, No. &&, 2021

DISCUSSION CAPE-V overall severity was associated with higher like-

Vocal fold nodules refer to bilateral fibrovascular thicken- lihood of undergoing voice therapy but parental derived
ings of the membranous true vocal folds. The etiology is perceptions of voice quality from the pVHI questionnaire
thought to be repetitive mechanical stress or phonotrauma, were not. Unfortunately, only 25% of patients had complete
typically related to voice use patterns. In children, the inci- pVHI data that could be used for analysis. This limits our
dence of vocal fold nodules has been reported to be as high capacity to draw conclusions. However, CAPE-V and
as 30% in the school age group.7 Children with siblings, pVHI have been shown to be correlated, so future studies
attention deficit hyperactivity disorder (ADHD), and male with a larger set of pVHI data may be able to elucidate
gender are most likely to be affected but females tend to pVHI’s relationship with the variables analyzed in this
have more progression in adolescence.8 study.12
Conservative management has traditionally been recom- Distance to therapy appears to be a barrier, as patients
mended for vocal fold nodules in the pediatric population. who received voice therapy were more likely to live closer to
This can include voice hygiene which focuses on education the therapy site. The method for determining distance,
about healthy voice care, increasing hydration, and elimina- though, involved an approximation using patient zip codes
tion of vocal abusive behaviors. Voice therapy incorporates that inherently has some degree of inaccuracy. Further-
a variety of behavioral techniques which may include semi- more, it does not account for the availability of public trans-
occluded vocal tract exercises, resonant voice, vocal func- portation options or the patient’s private transportation
tion exercises, biofeedback, and laryngeal massage. options. Several studies have suggested that patients are
Sequelae from vocal fold nodules and MTD can range more likely to pursue voice therapy if teletherapy is offered,
from very minimal to more severe voice and communication which removes the distance barrier.13 Telemedicine has not
outcomes. Patients with minimal sequelae may experience been commonly used in for pediatric voice therapy and little
no interference with their communication or quality of life. research is available regarding teletherapy in the pediatric
In severe cases, however, children can have issues with fre- voice population.14 In Massachusetts, state licensure regula-
quent periods of aphonia, painful phonation, and poor com- tions for speech-language pathologists were modified in
munication, affecting self-esteem and relationships with 2019 to include telepractice. However, insurance coverage
others. Treatments for dysphonia therefore can be very fam- for teletherapy has been inconsistent and as a result, voice
ily driven, as there may be hesitation to dedicate time and teletherapy has not been pursued until recently at Boston
energy to a non-life-threatening pathology. At our institu- Children's Hospital.
tion, the recommendation for therapy comes from a group Due to the COVID-19 pandemic, there has been an
consensus between a patient’s family, pediatric otolaryngol- abrupt transition to virtual visits in all aspects of medicine,
ogist, and speech language pathologist. including therapy. At our institution, speech language path-
Although voice therapy in children has been shown to have ologists have transitioned to teletherapy for patients ranging
good results, we recognize that not all children undergo from 4 to 17 years old with both vocal fold nodules and
therapy.7,9,10,11 In our case series, patient age proved to be a MTD who were initially evaluated in-office. Anecdotal evi-
statistically significant factor associated with receiving voice dence of positive parental feedback and gains made via tele-
therapy. A potential explanation for this may be the percep- therapy provide promising initial data regarding the efficacy
tion that older children are more likely to be able to partici- of voice therapy delivered via videoconferencing software.
pate or have perceptible gains through therapy. Older While audio quality via video conference is inferior to in-
children were not more likely to present with more severe person visits, limited research suggests that voice assessment
dysphonia, as measured by the SLP and CAPE-V. Age of the and treatment may be conducted with the use of a cellular-
patient also did not correlate with the parents’ perception of phone microphone.15 As gains made in voice therapy are
voice deviance, as measured by the pVHI. proven via ongoing research, it is hoped that insurance com-
In our study group, more boys than girls were included, panies will continue to provide consistent coverage for voice
consistent with prior findings that males are more likely to teletherapy after stay-at-home orders are lifted.
be diagnosed with vocal fold nodules in school age. This There were several limitations associated with this study.
trend reverses in the adolescent population. We did not find First, the study was performed in a retrospective manner at
that gender was associated with higher likelihood of receiv- a single tertiary pediatric institution. Also, data are not
ing voice therapy. available regarding the reason behind failure to undergo
Insurance coverage for voice therapy varies greatly by voice therapy. Several possibilities exist including loss to fol-
state, insurance company, as well as individual policies. The low-up, resolution of symptoms prior to therapy, or transi-
majority of patients in our case series had private insurance. tion of care to another institution. Further, some children
There was no significant difference detected in rates of voice and families may have elected to pursue voice therapy at a
therapy based on insurance type. Consideration was given closer facility or with therapy availability that better
to the possibility of patients carrying multiple types of insur- matched their schedule. Only 21% of patients elected to
ance and we recognize that our study is limited by categori- undergo therapy at our institution and our data did not cap-
zation of insurance type into simplified private or public ture those that elected for therapy elsewhere. Future studies
groupings. should take this factor into consideration. Though a 2-year
 ARTICLE IN PRESS
Anne F. Hseu, et al Barriers to Voice Therapy 5

study period was allowed for patients for the onset of ther- 5. Kempster GB, Gerratt BR, Verdolini Abbott K, et al. Consensus audi-
apy, there may also be patients that did start therapy outside tory-perceptual evaluation of voice: development of a standardized
of this time frame. Measurements for both inter- and intra- clinical protocol. Am J speech-language Pathol. 2009;18:124–132.
https://doi.org/10.1044/1058-0360(2008/08-0017.
rater reliability for our CAPE-V ratings are lacking. 6. Zur KB, Cotton S, Kelchner L, et al. Pediatric Voice Handicap Index
While previous studies have established the validity and reli- (pVHI): a new tool for evaluating pediatric dysphonia. Int J Pediatr
ability of the CAPE-V tool, future studies should nonethe- Otorhinolaryngol. 2007. https://doi.org/10.1016/j.ijporl.2006.09.004.
less include measures of rater reliability at their own Published online.
practice.5, 16, 17 Following the example of Prendes et al18, in 7. Tezcaner CZ, Ozgursoy SK, Sati I, et al. Changes after voice therapy
in objective and subjective voice measurements of pediatric patients
order to assess rater reliability, 10% of a study’s graded with vocal nodules. Eur Arch Oto-Rhino-Laryngol. 2009;266:1923–
voice samples should be randomly selected for repeat exami- 1927. https://doi.org/10.1007/s00405-009-1008-6.
nation during data analysis. Lastly, prospective studies 8. Mudd P, Noelke C. Vocal fold nodules in children. Curr Opin Otolar-
across multiple institutions or locales are needed to better yngol Head Neck Surg. 2018;26:426–430. https://doi.org/10.1097/
MOO.0000000000000496.
understand the findings of this study, although this study
9. Trani M, Ghidini A, Bergamini G, et al. Voice therapy in pediatric
did identify younger age and distance to therapy site as bar- functional dysphonia: a prospective study. Int J Pediatr Otorhinolar-
riers. Future research may entail family derived survey data yngol. 2007;71:379–384. https://doi.org/10.1016/j.ijporl.2006.11.002.
to complement our results so that these factors may be 10. Niedzielska G. Acoustic analysis in the diagnosis of voice disorders in
addressed when feasible. children. Int J Pediatr Otorhinolaryngol. 2001;57:189–193. https://doi.
org/10.1016/S0165-5876(00)00411-0.
11. Hartnick C, Ballif C, De Guzman V, et al. Indirect vs direct voice ther-
apy for children with vocal nodules. JAMA Otolaryngol Neck Surg.
CONCLUSION 2018;144:156. https://doi.org/10.1001/jamaoto.2017.2618.
Among pediatric patients diagnosed with vocal fold nodules 12. Johnson K. Comparison of the pediatric voice handicap index with
and muscle tension dysphonia, treatment with voice therapy perceptual voice analysis in pediatric patients with vocal fold lesions.
is not always undertaken. Older age, shorter distance to Arch Otolaryngol Neck Surg. 2011;137:1258. https://doi.org/10.1001/
archoto.2011.193.
therapy site, and increased CAPE-V Overall Severity were
13. Brehm SB, Weinrich B, Kelchner L. A practical discussion of current
associated with increased likelihood of receiving voice ther- approaches to providing voice therapy to children with dysphonia.
apy. Gender, insurance status (private vs public), and pVHI Perspect ASHA Spec Interes Groups. 2018;3:40–46. https://doi.org/
scores did not affect likelihood of receiving voice therapy. 10.1044/persp3.SIG3.40.
Patients may primarily consider ease of access and necessity 14. Doarn CR, Zacharias S, Keck CS, et al. Design and implementation of
an interactive website for pediatric voice therapy—the concept of in-
of treatment when considering voice therapy.
between care: a telehealth model. Telemed e-Health. 2019;25:415–422.
https://doi.org/10.1089/tmj.2018.0108.
15. Grillo EU. An online telepractice model for the prevention of voice dis-
REFERENCES orders in vocally healthy student teachers evaluated by a smartphone
1. Carding PN, Roulstone S, Northstone K, ALSPAC Study Team. The application. Perspect ASHA Spec Interes Groups. 2017;2:63–78.
prevalence of childhood dysphonia: a cross-sectional study. J Voice. https://doi.org/10.1044/persp2.SIG3.63.
2006. https://doi.org/10.1016/j.jvoice.2005.07.004. Published online. 16. Helou LB, Solomon NP, Henry LR, et al. The role of listener experi-
2. Martins RHG, Hidalgo Ribeiro CB, Fernandes de Mello BMZ, et al. ence on Consensus Auditory-perceptual Evaluation of Voice (CAPE-
Dysphonia in children. J Voice. 2012;26:674.e17–674.e20. https://doi. V) ratings of postthyroidectomy voice. Am J speech-language Pathol.
org/10.1016/j.jvoice.2012.03.004. 2010;19:248–258.
3. Akif Kiliç M, Okur E, Yildirim I, et al. The prevalence of vocal fold 17. Zraick RI, Kempster GB, Connor NP, et al. Establishing Validity of
nodules in school age children. Int J Pediatr Otorhinolaryngol. the Consensus Auditory-Perceptual Evaluation of Voice (CAPE-V).
2004;68:409–412. https://doi.org/10.1016/j.ijporl.2003.11.005. Am J speech-language Pathol. 2011;20:14–22.
4. Ongkasuwan J, Friedman EM. Is voice therapy effective in the man- 18. Prendes BL, Yung KC, Likhterov I, et al. Long-term effects of injec-
agement of vocal fold nodules in children? Laryngoscope. tion laryngoplasty with a temporary agent on voice quality and vocal
2013;123:2930–2931. https://doi.org/10.1002/lary.23830. fold position. Laryngoscope. 2012;122:2227–2233.