Predicting Communication-Related Quality of Life in Children With Velopharyngeal Dysfunction

Abstract

Objective:

The purpose of this study was to determine factors predicting communication-related quality of life in children with cleft palate with or without cleft lip (CP ± L) or congenital velopharyngeal insufficiency (VPI).

Design:

Cross-sectional design.

Setting:

Outpatient pediatric craniofacial anomalies clinic.

Patients:

Two-hundred and seventy-one children <18 years of age, diagnosed with CP ± L or congenital VPI.

Main Outcome Measure:

Velopharyngeal Insufficiency Quality of Life (VELO) scores as well as articulation, speech intelligibility, resonance, and voice outcomes were extracted from the electronic medical record in order to characterize communication-related quality of life. Total VELO score, as well as speech limitations, situational difficulty, emotional impact, perception by others and swallow subscores, were collected. Demographic factors, economic deprivation, medical history, and history of palatoplasty/pharyngoplasty were considered. Linear regression identified predictors of VELO total and subscores.

Results:

Poorer speech intelligibility and increased hypernasality significantly predicted lower (worse) total VELO scores (P < .01), as well as more severe speech limitations (P < .01), situational difficulty (P < .01), emotional impact (P < .01), and perception by others (P < .01). The presence of glottal stop substitutions also predicted more severe emotional impact (P < .01). A comorbid behavioral health condition predicted significantly lower total VELO scores (P < .01), and children living in neighborhoods experiencing greater socioeconomic deprivation presented with worse swallow section scores (P < .01).

Conclusions:

Hypernasality and decreased speech intelligibility primarily influenced the communication-related quality of life in children with velopharyngeal dysfunction. Additional research is warranted to determine how intervention can best mitigate VPI-related challenges in order to improve social interaction and quality of life.

Introduction

Cleft palate with or without cleft lip (CP ± L) is associated with a variety of communication difficulties.^1–3 These include language delays,^4–6 speech sound errors,^7,8 hypernasal resonance,⁹ and voice problems.^10,11 These communication differences can constrain social interaction, reduce engagement with others, and negatively impact the way interactional partners respond to an individual.^12,13 In fact, communication differences can limit a child’s participation within their social world and thus subvert their overall quality of life.¹⁴ Previous research has sought to describe how specific communication differences might influence communication-related quality of life,¹⁵ which is defined as the extent to which an individual’s speech supports meaningful communication and participation in daily life.^16–18

Velopharyngeal dysfunction (VPD) underlies some of the communication differences observed in this population.^19,20 Typical velopharyngeal function involves the contraction of muscles (ie, levator veli palatini, superior pharyngeal constrictor, palatopharyngeus) to create a sphincter or valve which separates the oral and nasal cavities.²¹ This separation is crucial for speech and swallow.²² Velopharyngeal dysfunction is an umbrella term that refers to various etiologies of inadequate closure between these 2 cavities, including congenital velopharyngeal insufficiency (VPI) which may result from structural deficits other than cleft.²³ Possible symptoms of VPD include hypernasal resonance, nasal air emission, deficits in speech articulation, and nasal regurgitation of liquid or food.^9,20,24,25

Speech differences associated with VPD not only hinder communication, they attract negative attention from interactional partners.^26,27 Patient-reported outcomes measures have been designed specifically to assess the influence of VPD on communication-related quality of life. One example is the VPI Effects on Life Outcomes Instrument (VELO). The VELO is a 23-item questionnaire (26 items when completed by parent proxy) aimed at assessing the influence of communication differences on children with VPD.^17,26,28 The VELO produces a total score as well as 5 subscores measuring speech limitations, swallow problems, situational difficulty, emotional impact, and perception by others. These subsections assess potential symptoms of VPD and their psychosocial implications.¹⁷ This item was created based on focus groups of clinicians, patients, and their parents.^17,26 The VELO has been employed to assess communication in a variety of contexts²⁹ and has been translated into multiple languages.^30–33

Previous studies examined the relationship between the VELO and speech function. In 48 children with CP ± L, parent-reported VELO scores were reported to correlate with perceptual speech outcomes as measured by the Pittsburgh Weighted Speech Scale (PWSS) completed by speech-language pathologists (SLP).¹⁴ The PWSS, however, accounted for only 26% of the variance in VELO scores,¹⁴ suggesting this relationship was weak. Another investigation reported that VELO scores were correlated with clinician ratings of intelligibility in 29 patients with CP ± L.²⁹ Yet more specific speech outcomes were not examined. Perhaps the most comprehensive investigation examined the Dutch version of the VELO in 25 patients with CP ± L.³⁴ Total scores were associated with speech intelligibility, speech acceptability, passive cleft speech characteristics, and velopharyngeal composite scores (combined hypernasality, nasal emission, and nasal turbulence ratings). Moreover, total VELO scores predicted the need for speech therapy.³⁴ Further research is needed to understand how these findings extend to the English version of the VELO. Considering that speech outcomes can be heterogeneous in children with CP ± L,^35,36 it is also important to consider a larger patient population.

Speech-language pathologist evaluations for children with CP ± L or congenital VPI focus on multiple aspects of speech which could drive VELO scores.³⁷ These include subsystems such as articulation and resonance as well as overall intelligibility.³⁸ Nonoral compensatory errors as well as obligatory errors associated with VPD may also drive VELO outcomes. For example, nonoral compensatory errors such as productions of glottal stops, pharyngeal fricatives, and/or nasal fricatives are more prevalent in the population with CP ± L when compared to individuals without those conditions,^7,39–41 and these can reduce speech intelligibility and acceptability and attract negative attention. Thus, the presence of nonoral compensatory errors, or reduced performance on formal articulation testing, might be associated with poorer VELO scores. Alternatively, it could be that more global measures of intelligibility predict VELO outcomes.²⁹ Intelligibility is a multifactorial construct reflecting how easily speech can be understood and is often assessed perceptually by clinicians or parents.^42–44 Multiple factors contribute to intelligibility including articulation, prosody, and even vocal volume.⁴⁵ It is possible that articulation test performance and intelligibility might influence VELO subscores differently.

Resonance disorders could potentially drive communication-related quality of life scores. Hypernasality or excessive nasal acoustic energy is not only associated with compensatory articulation errors, it may draw negative attention in and of itself.⁴⁶ Hypernasality is often evaluated by clinicians using auditory-perceptual measurement,⁴⁷ and nasalance—the ratio of nasal acoustic energy and total nasal and oral acoustic energy during an utterance—is assessed using the Nasometer.²⁵ It would be valuable to understand how hypernasality influences VELO scores and whether this relationship is better captured by auditory-perceptual assessment or Nasometer assessments. Vocal pathologies can influence an individual’s ability to manipulate vocal volume causing speech to sound more effortful.⁴⁸ As soft vocal volume⁴⁹ or rough voice quality can affect speech intelligibility,^50–52 it is possible that dysphonia could influence VELO outcomes.

There may be indirect factors that influence VELO scores. Palatoplasty type, for example, could influence surgical outcomes and therefore a child’s ability to build oral air pressure and close the velopharyngeal port.⁵³ The need for revision surgery could be associated with quality of life as patients experience VPD for longer periods of time and undergo multiple surgical interventions.⁵⁴ As the VELO includes a section probing nasal regurgitation, it could be that a history of feeding difficulties or dysphagia influence this subsection. It is likely that the presence of a genetic syndrome might influence VELO scores as medical complexity can influence quality of life.⁵⁵

It is also important to consider a patient’s behavioral health profile.⁵⁶ Numerous studies have shown that children with CP ± L or congenital VPI present with elevated rates of anxiety and other behavioral health diagnoses.^57–60 Furthermore, they may experience significant teasing or bullying due to speech or orofacial differences.^61,62 Behavioral health not only influences quality of life but it could influence the manner in which children communicate with others. As such, this factor should be considered when examining predictors of quality of life measures.

It is possible that economic deprivation or social economic status could influence communication-related quality of life.⁶³ Multiple studies have documented that children who experience economic deprivation perform more poorly on standardized language tests⁶⁴ and may experience deficits in vocabulary or phonology.⁶⁵ The effects of economic deprivation extend far beyond speech outcomes,⁶⁶ as low socioeconomic status, is tied to increased physical illness, poorer healthcare, behavioral health problems, lack of educational resources, and food insecurity.^67–70 Recent studies suggest that the effects of economic deprivation extend to children with CP ± L,^71,72 yet little is known about how this factor might influence quality of life measures in this population.

Although it is clear that children with CP ± L or congenital VPI are at risk for a range of communication challenges,⁷³ the specific factors which drive communication-related quality of life are unclear. As such, this study employed the VELO assessment to quantify the influence of speech differences in children with CP ± L or congenital VPI on various aspects of quality of life. Speech intelligibility, articulation, resonance, voice quality, medical comorbidities, history of palatoplasty/ pharyngoplasty, syndrome status, behavioral health profile, and economic deprivation were considered. It was hypothesized that reduced speech intelligibility, poorer articulation, and increased hypernasality would be associated with decreased communication-related quality of life. It was also hypothesized that behavioral health diagnoses or economic deprivation would be associated with worse VELO scores.

Methods

This cross sectional study examined children presenting to an outpatient craniofacial anomalies clinic with cleft palate or congenital VPI. All study procedures were approved by the relevant institutional review board. The VELO scores and other speech, voice, resonance, demographic, and medical history data were extracted from the electronic medical record (EMR) in order to characterize the impact of speech difficulties associated with CP ± L or congenital VPI.

Population

Patients were identified from the EMR of the craniofacial anomalies clinic associated with a large tertiary children’s hospital. Children were required to (a) be under the age of 18, (b) be diagnosed with cleft palate or congenital VPI by a plastic surgeon and SLP, (c) have undergone a multidisciplinary craniofacial team evaluation between the years of 2021 and 2023, and (d) have completed the VELO during the aforementioned evaluation. Children were excluded if they were nonverbal, presented with sensorineural hearing loss, spoke a language other than English in the home, presented with head or neck cancer, or presented with cognitive delays which prevented completion of the VELO. Hearing status was determined based on audiologist evaluations completed the same day as the VELO. Cognitive delays were identified based on parent report and the medical record using primary care or developmental pediatrician evaluations.

Measures

The following data were extracted from the EMR: VELO scores, patient demographics (age, sex), cleft-related diagnosis, syndromic diagnoses, medical comorbidities (ie, dysphagia, sleep apnea), Grade, Roughness, Breathiness, Asthenia and Strain (GRBAS) ratings of voice quality, behavioral health history, Nasometer scores, auditory-perceptual ratings of hypernasality, ratings of overall speech intelligibility, cleft-related articulatory errors, socioeconomic status, and cleft repair history. All evaluations were performed on the same day.

Velopharyngeal Insufficiency Quality of Life (VELO).

The VELO is a 23-item questionnaire aimed at assessing communication-related quality of life in children with VPD.^17,26,28 Patients are asked to complete the VELO based on their experience over the past 4 weeks. Each item presents the patient with a statement (eg, “air comes out of my nose when I talk”), which they rate on a 5-point Likert scale (never = 0, almost never = 1, sometimes = 2, often = 3, almost always = 4) based on how often they experience the symptom or situation described. The VELO produces 5 subscores and a total score. Subscores include sections measuring speech limitations, swallow problems, situational difficulty, emotional impact, and perception by others. The parent proxy form includes a sixth section which measures caregiver impact. For each section, items are summed to produce raw scores which are converted to standard scores.^17,26,28 Standard total and subscores range from 0 to 100, with greater values indicating better quality of life.^17,26,28 In the current study, the VELO was completed during speech evaluations with an SLP. For children <8, a parent completed the VELO form by proxy. Patients ≥8, completed the VELO independently.

Patient demographics.

Patient age (in years), sex, and 9-digit zip code of home address at the time of evaluation were collected, and cleft-related diagnoses were extracted. This included either Veau classification of cleft palate or congenital VPI.

Medical history.

History of palatoplasty/pharyngoplasty and diagnosis of genetic syndrome were extracted. Specifically, date and type of initial palatoplasty were gathered, as was any history of secondary surgery (palatoplasty/pharyngoplasty) following initial palatoplasty. Palatoplasty/pharyngoplasty procedures were performed by one of 3 plastic surgeons. For analysis, children were coded as either those who had a syndromic diagnosis or those who did not. A history of diagnosed dysphagia or sleep apnea was extracted from the medical record. These diagnoses were coded as either present or not for analysis. It was recorded whether patients had undergone a Videofluoroscopic Swallow Study within the last 5 years.

Behavioral health.

Any behavioral health diagnoses (ie, anxiety, depression, and posttraumatic stress disorder) were extracted from the medical record (present or not). Behavioral health diagnoses were made by psychiatrists or primary care physicians. For statistical analysis, behavioral health diagnoses were collapsed to increase statistical power.

Auditory-Perceptual Evaluation of Resonance.

Ratings of hypernasality were performed by one of 8 SLPs specialized in the treatment of craniofacial anomalies using the Cleft Audit Protocol for Speech–Augmented–Americleft Modification (CAPS-A-AM) scale.^74,75 Hypernasality was rated on a 5-point Likert scale ranging from balanced (0) to severely hypernasal resonance (4). Hyponasality was rated on a 3-point Likert scale, ranging from none (0) to markedly hyponasal (2).⁷⁴ All SLPs received formal training on rating resonance. Consensus meetings were intermittently held to ensure interrater reliability throughout the study period. At the time of data extraction, the most recent consensus meeting produced an interclass correlation coefficient (2-way, random effects) of .87 (P = .011). In keeping with past study, this clinic requires a minimum of .75 for ratings to be considered in “good” reliability range.⁷⁶

Nasometer Scores.

Nasometry data were collected from the EMR. These data were originally collected using the Nasometer II (Model 6400, Pentax Medical). The picture-cued portion of the MacKay-Kummer SNAP-R test protocol was used.⁷⁷ Data collected on the bilabial plosive phrase (“pick up the book, pick up the baby, pick up the pie”) was used for analysis. This sentence was chosen because past study suggests results from this stimuli are correlated with the other stimuli loaded with oral pressure sounds,^25,78,79 and many children were too young to reliably produce a reading passage.

Auditory-Perceptual Voice Assessment.

The GRBAS⁸⁰ ratings were performed by SLPs on the day of evaluation. All listed parameters of voice were rated on a 4-point Likert scale with greater numbers indicting more severely impacted voice quality. Consistent with past research, only overall grade of voice quality was considered for analysis.^81,82

Speech Intelligibility.

Intelligibility was rated using a 4-point Likert scale employed in previous investigations.^44,73,83 Ratings ranged from normal (0), mildly impaired (1), moderately impaired (2), and severely impaired (3).⁴⁴ The Intelligibility in Context Scale (ICS) was also collected. The ICS is a 7-item tool where parents rate their child’s intelligibility over the last month.⁸⁴ Each item is rated on a 5-point Likert scale, with higher numbers indicating better intelligibility.⁸⁴

Articulation.

Children recited the American English Sentence Sample.⁸⁵ If no articulation deficits were detected by the evaluating SLP—no further evaluation was performed. If any difference or deficit was suspected, or if children were too young to repeat the American English Sample sentences, articulation was formally tested using either the Goldman-Fristoe test of articulation or the Iowa Pressure Articulation Test, which were scored according to published protocols.^86,87 In addition to formal articulation testing, SLPs determined whether children presented with articulatory modifications commonly associated with VPD. These included glottal stops, pharyngeal fricatives, and nasal fricatives.^7,88–90 Each type of articulation error was rated as either present or not.

Area Deprivation Index (ADI).

Patient 9-digit zip codes were converted into ADI scores. The ADI measures socioeconomic disadvantage on a neighborhood (or census block) level.⁹¹ This composite measure considers income, housing quality, employment, and educational opportunities in a given area.⁹¹ Country-level ADI scores were utilized, which measure socioeconomic advantage on a 100-point scale with higher numbers indicating greater levels of disadvantage.⁹¹

Statistics

Descriptive statistics and frequencies were calculated for all measures. Linear regression was used to identify predictors of VELO total and subscores. To identify factors that should be included in the regression model, univariate analyses were performed. Factors significantly associated with VELO total or subscores at the alpha .01 level were then regressed on VELO scores to determine predictors of this outcome. Total score and speech-related subscores (speech limitation, situational difficulty, emotional impact, and perception by others) were significantly associated with intelligibility rating, presence of glottal stops (present, none), auditory-perceptual ratings of hypernasality (CAPS-A-AM scores), syndrome status (syndrome diagnosed, none), behavioral health diagnosis (yes/no), palate repair type (straight line, Furlow double opposing Z-plasty, unknown, none), and ADI score. Thus these factors were regressed on total VELO score and all speech-related subscores (in separate models). Factors regressed on swallow problems subscore were history of dysphagia (yes/no), auditory-perceptual ratings of hypernasality, intelligibility rating, syndrome status, and ADI score. Age was included as a covariate in all models. Auditory-perceptual ratings of hypernasality and Nasometer scores were both associated with VELO scores; however, these measures were also correlated with each other and thus both factors could not be included in regression models. Auditory-perceptual ratings of resonance were more strongly associated with VELO scores; thus this measure was included in the model. Similarly, ICS scores and clinician ratings of intelligibility were correlated. Clinician ratings were more strongly associated with VELO scores; therefore, these ratings were selected for regression models. Post hoc analyses were performed using Tukey’s Honest Significant Difference tests, odds ratios, or Pearson’s correlations as appropriate. Bonferroni corrections were employed to protect against type-1 error. Alpha was set at .01 for determining significance.

Results

Two-hundred and seventy-one children with CP ± L or congenital VPI (mean age = 10.3, SD = 4.2) were included in this study. Demographic data, medical history, and distribution of speech, voice, and swallow outcomes are presented in Table 1.

Table 1.

Patient Demographic, Speech, Medical History, and VELO Data.

Factor	N	%	Factor	N	%	Factor	N	%
Age			VELO (mean/SD)			Nasometer (mean/SD)
Mean/SD	10.3	4.2	Total score	75.31	20.8	Bilabial plosive	17.12	11.1
Sex			Speech limitations	72.86	21.1	Alveolar	18.68	12.4
Male	162	59.8%	Swallow problems	94.73	12.5	Velars	18.49	12.4
Female	109	40.2%	Situational difficulty	72.52	25.8	Sibilants	22.32	15.0
Cleft-related diagnosis			Emotional impact	86.41	19.3	Nasals	55.94	33.8
Congenital VPI	31	11.4%	Perception by others	89.59	17.4	Zoo passage	17.47	11.5
Veau I	10	3.7%	Caregiver burden	82.97	22.5	Hypernasality
Veau II	54	19.9%	GRBAS			Balanced resonance	125	46.1%
Veau III	71	26.2%	Normal voice	224	82.7%	Slight	90	33.2%
Veau IV	48	17.7%	Mild	137	50.6%	Mild	33	12.2%
Submucous	57	21.0%	Moderate	8	3.0%	Moderate	15	5.5%
Syndromes/other diagnoses			Voice complaint			Severe	8	3.0%
None	220	81.2%	Yes	31	11.4%	Hyponasality
22q11.2 Deletion	18	6.6%	No	240	88.6%	Absent	233	86.0%
Opitz	3	1.1%	History of dysphagia			Mild	31	11.4%
Pierre Robin sequence	12	4.4%	Yes	84	31.0%	Marked	6	2.2%
Hemifacial microsomia	4	1.5%	No	187	69.0%	Intelligibility
Fetal alcohol syndrome	3	1.1%	VFSS in last 5 years			Normal	125	46.1%
Stickler syndrome	3	1.1%	Yes	22	8.1%	Mild	97	35.8%
Other	8	3.0%	No	249	91.9%	Moderate	38	14.0%
Malocclusion			Behavioral health			Severe	10	3.7%
Class 1	34	12.5%	None	216	79.7%	Intelligibility in Context Scale (N = 118)
Class 2	23	8.5%	ADHD	17	6.3%	(mean/SD)	28.3	5.5
Class 3	70	25.8%	Anxiety	18	6.6%	Glottal Stops
None	127	46.9%	Depression	8	3.0%	Present	19	7.0%
Palate surgery type			OCD	1	0.4%	None	252	93.0%
None	59	21.8%	Autism spectrum	10	3.7%	Pharyngeal fricatives
Straight-line	93	34.3%	ODD	1	0.4%	Present	10	3.7%
Furlow Z-plasty	86	31.7%	Behavioral Health Treatment			None	258	95.2%
Unknown	31	11.4%	Yes	39	14.4%	Nasal fricatives
Age at palate surgery			None	232	85.6%	Present	10	3.7%
Months (mean/SD)	13.1	1.2	Sleep apnea			None	256	94.5%
Second palate surgery			Yes	98	36.2%	Articulation testing method
Yes	69	25.5%	No	171	63.1%	IPAT	124	45.8%
No	202	74.5%	Area Deprivation Index			GFTA	45	16.6%
Palate fistulas observed			1–69	209	77.1%	Sentence repetition	102	37.6%
Yes	30	11.1%	>70	62	22.9%	Articulation scores (mean/SD)
Labial alveolar	5	1.8%	ADI (mean/SD)	52.8	20.4	IPAT (mean/SD)	37.7	9.9
None	223	82.3%				GFTA (mean/SD)	64.8	29.2

Abbreviations: ADHD, attention deficit hyperactivity disorder; OCD, obsessive compulsive disorder; ODD, oppositional defiance disorder; VPI, velopharyngeal insufficiency; VELO, Velopharyngeal Insufficiency Quality of Life; IPAT, Iowa Pressure Articulation Test; GFTA, Goldman Fristoe Test of Articulation; VFSS, videofluoroscopic swallow study; GRBAS, Grade, Roughness, Breathiness, Asthenia and Strain.

Total Score.

The regression model accounted for 50.8% of variability in total VELO score (P < .001, r2 = .50). Intelligibility ratings significantly predicted total score (β = −7.70, t = −4.5, P < .001), as children with normal intelligibility (M = 83.8, SD = 16.2) scored significantly higher than those with mildly impaired intelligibility (P < .001, M = 72.2, SD = 21.4). Scores significantly decreased again for children with moderately impaired intelligibility (P < .001, M = 58.8, SD = 19.2), and remained in similar range for those with severely impaired intelligibility (M = 62.7, SD = 21.9). Distribution of VELO scores across intelligibility ratings is presented in Figure 1. Hypernasality ratings also predicted total score (β = −4.89, t = −3.3, P < .001), as children rated as having balanced (M = 82.0, SD = 16.3) or slightly hypernasal resonance (M = 81.8, SD = 14.3) scored in statistically similar range (P = .98); however, scores were significantly lower for children rated as having mildly hypernasal resonance (P < .001, M = 72.3, SD = 21.7). Scores significantly decreased a second time for children with moderate hypernasality (P < .001, M = 62.3, SD = 23.2) and remained in statistically similar range for those with severely hypernasal speech (P = .78, M = 61.7, SD = 23.1). Distribution of VELO scores across hypernasality ratings is presented in Figure 2. Finally, behavioral health status predicted total VELO score (β = −7.71, t = −2.7, P = .007), as children with behavioral health diagnoses presented with significantly lower total VELO scores (M = 67.8, SD = 23.1) than children without any such diagnosis (P = .003, M = 77.2, SD = 19.8). No other significant predictors were identified.

Figure 1.

Means and standard deviations for Velopharyngeal Insufficiency Quality of Life (VELO) total and subscores across speech intelligibility ratings.

Figure 2.

Means and standard deviations for Velopharyngeal Insufficiency Quality of Life (VELO) total and subscores across hypernasality ratings.

Speech Limitations Subscore.

The regression model accounted for 59% of the variability in emotional impact scores (P < .001, r2 = .59). Intelligibility ratings significantly predicted speech limitations score (β = −10.18, t = −1.6, P < .001), as children rated as having normal intelligibility (M = 82.5, SD = 14.9) scored significantly higher than children rated as having mildly impaired intelligibility (P < .001, M = 70.3, SD = 20.7). Scores significantly decreased again for children with moderately impaired intelligibility (P < .001, M = 52.1, SD = 19.2) and remained in statistically similar range for those rated as having severely impaired intelligibility (P = .149, M = 54.6, SD = 24.4). Additionally, hypernasality ratings significantly predicted speech limitations score (β = −4.45, t = −3.2, P = .001). Children with balanced resonance (M = 80.18, SD = 16.5) or slightly hypernasal resonance (M = 81.2, SD = 15.8) presented with significantly higher scores than those with mild hypernasality (P < .01, M = 70.57, SD = 21.8). Scores significantly decreased again for children with moderate hypernasality (P < .01, M = 57.6, SD = 22.1), and remained in similar range for those with severe hypernasality (P = .79, M = 55.05, SD = 21.1). No other significant predictors were identified.

Swallow Problems Subscore.

Regression accounted for 7.2% of variability in swallowing problems score (P < .001, r2 = .07). The ADI significantly predicted swallow problem score (β = −0.93, t = −2.5, P = .01), as children living in areas with greater socioeconomic disadvantage presented with significantly lower scores (ADI ≥ 70 M = 90.4, SD = 19.1) than those from more affluent areas (P = .002, ADI ≤ 69 M = 95.9, SD = 9.5). No other significant predictors were identified.

Situational Difficulty Subscore.

The regression model accounted for 66% of variability in situational difficulty scores (P < .001, r² = .66). Intelligibility ratings significantly predicted situational difficulty score (β = −16.43, t = −10.1, P < .001), as children with normal intelligibility (M = 86.8, SD = 15.3) scored significantly higher than children rated as having mildly impaired intelligibility (P < .001, M = 68.9, SD = 23.1). Scores significantly decreased again for children with moderately impaired intelligibility (P < .001, M = 43.1, SD = 21.8) and remained in statistically similar range for those rated as having severely impaired intelligibility (P = .25, M = 48.0, SD = 34.0). Furthermore, hypernasality ratings significantly predicted situational difficulty score (β = −4.43, t = −2.9, P = .004). Children with balanced resonance (M = 81.3, SD = 20.3) presented with statistically similar scores as those with slight hypernasality (P = .23, M = 76.8, SD = 19.6). Scores were significantly lower for children with mildly hypernasal speech (P < .01, M = 67.9, SD = 26.9). Scores significantly decreased again for children with moderate hypernasality (P < .01, M = 57.2, SD = 23.4) and remained in statistically similar range for those with severe hypernasality (P = .62, M = 54.0, SD = 35.7).

Emotional Impact Subscore.

The regression model accounted for 47% of the variability in emotional impact score (P < .001, r² = .47). Intelligibility ratings significantly predicted emotional impact score (β = −5.56, t = −3.63, P < .001). Children rated as having normal intelligibility (M = 93.46, SD = 12.0) scored significantly higher than children rated as having mildly impaired intelligibility (P < .001, M = 85.13, SD = 17.4). Scores significantly decreased again for children with moderately impaired intelligibility (P < .001, M = 68.59, SD = 26.4) and remained in statistically similar range for those rated as having severely impaired intelligibility (P = .149, M = 77.55, SD = 21.45). Glottal stops also significantly predicted emotional impact score (β = −14.8, t = −3.13, P = .002). Children with glottal stops presented with significantly lower emotional impact scores (M = 60.92, SD = 29.1) than children without glottal stops (P < .001, M = 88.25, SD = 16.9). Hypernasality ratings also significantly predicted emotional impact score (β = −3.17, t = −2.39, P = .017). Children with balanced resonance (M = 92.0, SD = 11.9) presented with significantly higher scores than those with slight hypernasality (P < .01, M = 84.24, SD = 22.4). Scores significantly decreased again for children with mild hypernasality (P < .01, M = 74.48, SD = 21.37), and remained in statistically similar range for those with moderate (P = .79, M = 75.9, SD = 28.4) and severe hypernasality (P = .71, M = 82.5, SD = 21.4). No other significant predictors were identified.

Perception by Others Subscore.

The regression model accounted for 27% of the variability in emotional impact score (P < .001, r² = .27). Intelligibility ratings significantly predicted perception by others score (β = −7.52, t = −5.9, P < .001). Children rated as having normal intelligibility (M = 96.08, SD = 9.2) or mildly impaired intelligibility (M = 89.7, SD = 13.9) scored significantly higher than children rated as having moderately impaired intelligibility (P < .001, M = 43.1, SD = 21.8) and remained in statistically similar range for those rated as having severely impaired intelligibility (P = .41, M = 48.0, SD = 34.0). Moreover, hypernasality ratings significantly predicted perception by others score (β = −3.11, t = −2.6, P = .009) because children with balanced resonance (M = 94.1, SD = 20.3) scored significantly higher than children perceived as having any degree of hypernasality (P < .001, M = 57.2, SD = 23.4). No other significant predictors of perception by other score were identified.

Discussion

Children with cleft palate, with or without cleft lip, (CP ± L) present with a range of communication challenges.³⁹ These children are at particular risk for VPD which may hinder communication, attract negative attention, and thus diminish quality of life.^9,46 Specific factors which drive the impact of communication difficulties in children with CP + L and/or congenital VPI are not well understood. This study employed the VELO assessment to quantify the influence of speech difficulties on communication-related quality of life in children with CP ± L or congenital VPI. It was hypothesized that reduced speech intelligibility, poorer articulation, and increased hypernasality would be associated with worse communication-related quality of life. This hypothesis was partially confirmed as more severe speech intelligibility and auditory-perceptual ratings of hypernasality significantly predicted worse total VELO scores as well as the speech limitations, situational difficulty, emotional impact, and perception by others subscores. Regarding articulation, presence of glottal stop errors predicted more severe emotional impact subscores, however, no other significant relationships between speech modifications or formal articulation testing were observed. It was also hypothesized that a behavioral health diagnosis or economic deprivation would be associated with VELO scores. These hypotheses were partially confirmed. The presence of a behavioral health condition predicted more severe total VELO scores, but the influence of socioeconomic deprivation was limited to the swallow subscore. This occurred because greater socioeconomic deprivation was associated with increased nasal regurgitation symptoms. These findings offer insight into factors that impact communication-related quality of life in children with CP ± L or congenital VPI.

Speech intelligibility is a global measure of how well an individual’s verbal communication is understood by others. The speech behaviors probed on the VELO focus on an individual’s ability to be understood in various contexts. For example, the situational difficulty portion of the VELO specifically probes how easily various communication partners understand the child. The findings of this study emphasize that a child’s overall ability to be understood influences their ability to communicate and likelihood of attracting negative attention. It was not surprising, therefore, that poorer speech intelligibility ratings were associated with worse VELO scores.

It was interesting that speech intelligibility rather than articulation testing significantly predicted VELO scores. This likely occurred because overall speech intelligibility, rather than individual speech sound errors, interfered most with communication. VPI symptoms, such as hypernasality and audible nasal emissions, are not usually captured on standardized articulation testing but these patterns can impact intelligibility. One exception was that glottal stop substitutions predicted worse emotional impact scores. Glottal stops occur when an individual obstructs airflow on the level of the glottis in an attempt to build oral air pressure⁹² and/or compensate for insufficient oral air pressure due to poor velopharyngeal port closure.⁹³ Glottal stops can be noticeable and attract negative attention, especially when they co-occur with hypernasality.⁹³ In the current study, children who produced glottal stops reported more frustration secondary to their speech differences. These findings highlight the importance of designing SLP intervention programs to enhance overall intelligibility in spontaneous communication and to minimize the production of modifications that detract from communication and interfere with speech intelligibility.

As expected, hypernasality influenced VELO scores. Hypernasality is prevalent in children with VPD, and the speech limitations section of the VELO specifically probes excessive nasal resonance. Only one question on the speech limitations section specifically addresses “air coming out the nose,” suggesting hypernasality was associated with other factors including communication-related weakness, inadequate oral pressure, fatigue, or difficulty coordinating airflow.⁹⁴ Children with hypernasality may have made laryngeal, articulatory, or respiratory compensations to ensure that they were understood, such as overarticulating, increasing vocal effort, or adjusting speech rate. These compensations could demand increased physical and mental effort.⁹⁵ Future research may determine the influence of fatigue or weakness on nasality and general speech intelligibility.

Increased hypernasality was also associated with more severe perception-by-others scores and greater emotional impact scores. Increased nasal resonance negatively influenced the way children’s speech was perceived by others, and children and their parents were aware of this influence. These findings reflect past study reporting that hypernasality draws negative attention from peers^46,96 and that from 21% to 47% of children with CP ± L have been bullied because of their speech difficulties.^97–99 Treatment teams should be cognizant of the fact that children with hypernasality may experience negative emotion sequela and should consider the negative emotional impact of hypernasality when designing treatment plans to support children with CP ± L or congenital VPI.

The presence of a behavioral health diagnosis was associated with more severe total VELO scores. Behavioral health comorbidities are relatively prevalent in the population with CP ± L,⁵⁷ and children experiencing anxiety or depression symptoms would likely be particularly sensitive to the psychosocial influence of speech differences. Behavioral health comorbidities could make it more difficult for children and their families to pursue speech interventions. It should be noted, however, that no significant differences were observed for any single subscore, suggesting that the influence of behavioral health comorbidities was general rather than specific.

It was concerning that children facing economic deprivation scored more poorly on the swallow section of the VELO. This section primarily probes nasal regurgitation, suggesting children from less affluent areas were more likely to experience this symptom. Socioeconomic status and race influence cleft-related care—with children from racial minority groups or poorer families experiencing more delays for palatolplasty.¹⁰⁰ Additionally, recent study suggests that children experiencing socioeconomic deprivation are more likely to experience postsurgical complications following revision palatoplasty.¹⁰¹ This finding should be interpreted with caution, however, considering that the regression model for the swallow portion of the VELO only accounted for 7.2% of variability in scores. Still, possible effects of economic deprivation on the quality of life of children with CP ± L warrant additional investigation. It should be noted that a history of dysphagia was not associated with increased swallow scores. This finding is logical given that the swallow section of the VELO probes nasal regurgitation rather than deglutition or aspiration symptoms.

It is noteworthy that surgical technique was not significantly associated with VELO outcomes. This may be because surgical skill rather than technique primarily drives fistula prevalence¹⁰² and perhaps indirectly speech outcomes. As the current study examined data from a single center with 3 plastic surgeons, it may be that surgeon skill level and surgical decision making were fairly homogenous. Future work should examine patient-reported outcomes globally, as it is possible that data from a single location reflect specific geographical, cultural, or practice patterns.

There are some factors which warrant consideration when interpreting the current data. Since the current study was conducted at a single site, future work may examine data from multiple institutions in order to reflect children with CP ± L across the US. A strength of this study lies in the examination of clinical practice data from a large sample of children with CP ± L. Thus, these data reflected patterns found in actual clinical practice. Although the study design prevented the calculation of individual reliability measures, experienced SLPs performed regular reliability and consensus meetings to ensure accuracy of ratings. Additionally, the regression models performed in this study only partially explained the variation in VELO scores; therefore, there were clearly additional influential factors outside the scope of this study which warrant future investigation. Likewise, it should be remembered that neighborhood measures of socioeconomic status such as the ADI are not perfect measures of economic deprivation. Future study should include individual proxies of socioeconomic privilege.⁷²

Conclusion

Poorer speech intelligibility and increased hypernasality significantly predicted more severe total VELO scores as well as worse subscores on speech limitations, situational difficulty, emotional impact, and perception by others. No significant relationships between speech errors or formal articulation testing were observed, but the production of glottal stop substitutions predicted more severe emotional impact subscores. This would suggest that in children with CP ± L, communication-related quality of life is driven by highly salient aspects of speech such as hypernasality, presence of glottal stops, and global measures of intelligibility. The presence of a behavioral health condition predicted more severe total VELO scores, and children living in neighborhoods experiencing greater socioeconomic deprivation presented with poorer scores on the swallow section of the VELO. These findings highlight factors that impact communication-related quality of life in children with CP ± L or congenital VPI and emphasize the importance of addressing VPI-related communication challenges to improve patient quality of life.

Data Availability

Due to ethical and privacy considerations, data cannot be made available.