Abstract
Objectives: (1) To define adequate speech material and normative data for Dutch children with a new Nasometer, the "NasalView System". (2) To define the minimal amount of required speech tasks. Methods: 55 children (30 normal and 25 velo-pharyngeal insufficient), aged 4–11 are included. All children had to read or repeat two Dutch passages [one with a normal amount of nasal consonants (normal passage) and one with none (denasal passage)]. (1) Out of group means and standard deviations (S.D.) "pathological nasalance boundaries" [mean±2×S.D.], in combination with the sensitivity, specificity and positive predictive values for both passages, are computed. (2) With ANOVA all sentences within each passage are tested for significant differences in nasalance. Results: (1) The pathological boundaries are 28.6–41.4% (mean: 35.0) and 21.4–34.7% (mean: 28.1) for the normal and denasal passage, respectively. For the normal passage a sensitivity of 96%, a specificity of 93% and a positive predictive value of 92% is computed. For the denasal passage these parameters are 96%, 95% and 96%, respectively. (2) Within the normal passage only the third sentence is significantly different in nasalance, compared to the entire passage (31.2% vs. 35.0%). Within the denasal passage the second and fifth sentence are significantly different (23.8% and 24.8% vs. 28.1%). However, the individual non-significantly different sentences show a higher variation in nasalance compared to the entire passages.
E-mail address: [email protected]
doi:10.1016/S0531-5131(03)01083-5
Click here for the PDF version
Contents
1. Introduction
Hypernasal speech is a major symptom in children with cleft palate abnormalities or with other etiologies of velopharyngeal insufficiency. Nasalance, as measured by an acoustic Nasometer, has been defined as the ratio between 'nasal' and 'whole voice' (nasal+oral) sound pressure levels, and this ratio is logically expected to be increased in case of velopharyngeal insufficiency. It appears to be an interesting method for quantifying nasality. In 1998 the Nasalance Acquisition System (NasalView) became available, with editing facilities for separate measures on selected parts of the speech material [1, 2, 3, 4, 5, 6].
The objective of this study is:
- (1) To elaborate adequate speech material for young children, and to investigate the minimal amount of required speech tasks for a reliable estimation of the average nasalance, considering that in young children the duration of cooperation may be limited.
- (2) To define normative values, as well as sensitivity and specificity percentages.
2. Subjects and methods
2.1. Subjects
Thirty normal children and 25 children with objectivated velo-pharyngeal insufficiency, all Dutch speaking.
2.2. Material
Speech material was phonetically selected as follows:
- (1) a passage consisting of eight short and simple "normal" sentences with a proportion of nasal sounds (11.67%) representative for Dutch language (11.63%), and
- (2) a passage consisting of ten short and simple sentences containing no nasal sounds at all (the "denasal" passage).
Both passages are suitable for young children (very simple words and syntax) as they can easily be repeated in case the child is not able to read them. Editing facilities allow, when necessary, the removal of the speech of parent or speech therapist before analysis. Table 1 shows a phonetic transcription of these passages.
3. Results
- The normal limits (±2 standard deviations) for nasalance percentages are 28.6–41.4% (mean: 35.0) and 21.4–34.7% (mean: 28.1), for the normal and denasal passage, respectively.
- Within the normal passage only the third sentence significantly differs in nasalance, when compared to the entire passage (31.2% vs. 35.0%). Within the denasal passage the second and fifth sentence significantly differ (23.8% and 24.8% vs. 28.1%) when compared to the entire passage. However, inter-subject variability increases when the speech sample is limited to a part of the material, although it remains satisfactory, even with one single sentence (standard deviation increases by about 20%).
- Nasometry clearly discriminates normal children from children with known hypernasality (Figs. 1 and 2). For the normal passage a sensitivity of 96%, a specificity of 93% and a positive predictive value of 92% is computed. For the denasal passage these parameters are 96%, 95% and 96%, respectively.
4. Conclusions
The NasalView System seems to be reliable and quantifies valid nasalance values when nasality is evaluated. Editing facilities and the choice of adequate speech material is essential when dealing with young children.
Nasometry clearly discriminates normal children from children with known velpharyngeal insufficiency.
Within both passages high levels of sensitivity, specificity and positive predictive values are obtained.
When necessary (difficult child), the speech material could be reduced to one of the sentences of the normal passage, with the exception of the third one. For the denasal passage, the speech material could also be reduced to one of the sentences, but accepted the second and the fifth ones. The same sentence needs to be used for comparison (e.g. after treatment). Due to larger interindividual variation of nasalance percentages, the use of single sentences is to be limited to particular situations.
The denasal passage discriminates slightly better for hypernasal speech.