The effects of infant orthopaedics on speech and language development in children with unilateral cleft lip and palate
A scientific essay in Medical Sciences
DOCTORAL THESIS defended in public on 12th of November 2002
This thesis describes an investigation into the effects of infant orthopaedics (IO) on the speech and language development in children with complete unilateral cleft lip and palate (UCLP). The study was performed within the framework of the three-centre prospective randomised clinical trial Dutchcleft.
Chapter 1 is a general introduction to the thesis. It describes the controversy regarding the relevance of intra-oral infant orthopaedics, a treatment that is used throughout the world in managing babies with cleft lip and palate. Proponents of IO believe that it enhances feeding, narrows the cleft, results in better facial aesthetics, and leads to better speech development. Opponents state that the treatment is expensive, inhibits maxillary growth, and increases the incidence of dental caries. Most of these statements are based on case reports, clinical observation, intuition, and personal preference. In order to provide the discussion on the value of IO with sound, scientifically based arguments, the Dutchcleft study started in 1993. It comprises four research areas: (1) general aspects, (2) surgical and orthodontic aspects, (3) speech and language development, and (4) a cost-effectiveness analysis. A total number of 54 children with complete UCLP participated in the trial. At trial entrance, which was within two weeks after birth, the babies were assigned to one of two groups by means of a computerised allocation procedure. One group (n=27) received IO in the first 12 months of life (IO group), and the other group (n=27) was not treated with IO (non-IO group). The IO group received a passive maxillary orthopaedic appliance that was made of compound soft and hard acrylic. Parents were instructed to have their baby wear the appliance 24 hours a day, removing it only for cleaning. Every six weeks, the appliance was adjusted by grinding to guide the maxillary segments into the right position. In all babies included in the trial, the lip was closed surgically according to the Millard technique at 18 weeks of age. The palate was closed in two stages, with soft palate repair (modified Von Langenbeck procedure) at 12 months of age. Hard palate closure is delayed until approximately nine years of age. After surgical closure of the soft palate, the appliance was no longer used. The evaluation of the effects of IO on speech and language development was planned as a three-year project funded by the Dutch National Health Service Board. The speech and language evaluation started at age one and continued at 6-month intervals until age 3. Due to a lengthy inclusion period, only part of the children participating in the full trial could be included in the speech and language assessment. The first speech evaluation started just before soft palate closure at the age of 12 months, when the appliance was still in situ in the IO group. At that age, all children were using pre-lexical utterances predominantly.
Chapter 2 focuses on the effects of IO on pre-lexical development at age 1 and 1.5. It was hypothesised that during the application of IO, the child would be able to develop more normal sensori-motor patterns in babbling, because the appliance creates an artificial alveolar ridge and covers the cleft in the hard palate. Consequently, we expected to find an enhanced use of alveolar sounds and high pressure sounds in babbling of IO babies. The development of pre-lexical sound play was assessed by means of a classification system which is based on the infant’s speech production capacities. The system classifies vocalisations for phonation and articulatory movement. In addition to this classification, consonant-like elements (contoids) were analysed for place and manner of articulation by two trained listeners. No differences were found in the vocalisations of the two groups as far as phonatory characteristics and articulatory movement were concerned. The assessment of contoid usage showed that, at the age of 1, babies in the IO group used statistically significant more alveolar contoid sounds in their babbling than the non-IO babies. In normal speech development, alveolar sounds are predominant at this age. So, compared to the non-IO group, the IO children used a phonetic repertoire that was closer to normal than the non-IO children at age one. At 1.5 years of age, when the IO group did no longer use the orthopaedic appliance, there were no apparent differences in characteristics of babbling between the two groups. This seemed to suggest that the effects of IO were temporary and present only during the period in which the IO is applied.
In chapter 3, the phonological development of the groups between age 2 and 3 was described. Since the babbling period is demonstrated to be related to the development of later speech, a beneficial effect of IO treatment in the pre-lexical period may also influence the phonological development in meaningful speech. In this light, it was hypothesised that the better opportunities to practice and establish normal speech motor patterns in the IO group would result in a phonological development that is closer to normal than when treated without IO. The phonological skills of the children were analysed by means of a system for the assessment of phonological development of Dutch children (Fonologische Analyse van het Nederlands: FAN). FAN is based on non-linear phonology and makes use of hierarchical relations between phonological features. It also accounts for the order in which children usually acquire the system of contrastive features. The phonological analysis in this thesis included the number of acquired consonants, order of phonological development, use of phonological processes, and occurrence of nasal escape. No differences were found between the groups in the use of phonological processes at either age. Furthermore, the groups did not differ in the order of phonological development at age 2. At 2.5 years of age, however, the order of phonological development of most IO children was normal or delayed, whereas most non-IO children followed an abnormal developmental pattern. At 3 years of age, the IO group had acquired more initial consonants than the non-IO group. In conclusion, the children who were treated with IO during their first year of life followed a more normal path of phonological development between 2 and 3 years of age.
At the age of 2.5, the speech of the two cleft lip and palate groups (IO, n=10; non-IO, n=10), and the speech of a non-cleft control group (n=8) was evaluated in two perceptual evaluations. Chapter 4 describes an experiment with 16 lay listeners who blindly assessed speech intelligibility. Intelligibility is an important characteristic of speech that reflects the effectiveness of communication. The development of less deviant sensori-motor patterns for speech in the pre-lexical period, and the more normal phonological development in meaningful speech may lead to better intelligibility in the children who were treated with IO. In order to test this hypothesis, the listeners first performed a write-down task, in which they indicated in normal spelling what they had understood of the utterance. After the transcription of the sample, the listeners rated the intelligibility of the speech sample on a ten-point equal-appearing-interval (EAI) scale. From the write-down task, the percentage of words correctly understood was calculated. Before the ratings of intelligibility were used in the analysis, their reliability was assessed first. Both the intra- and interrater reliability were very high. Statistical analysis showed that the ratings in this experiment did not correlate perfectly with the percentage of words correctly understood. This was manifest in the results: the IO group obtained higher intelligibility ratings when compared to the non-IO group, however, this was not reflected in a better write-down score for this group. In the write-down task, the listeners did not better understand the speech of the IO children. In this chapter, the intelligibility ratings for both cleft groups were also compared to the ratings obtained by the control group of non-cleft peers. It appeared that the difference in judged intelligibility between the IO and non-IO group was not statistically significant when tested against the non-cleft control group. Therefore, it is not clear whether IO actually improved the intelligibility of the speech.
The second experiment that was carried out to assess the speech at age 2.5 was a more comprehensive perceptual evaluation with a panel of five expert listeners (see chapter 5). In this experiment too, the speech of the two cleft groups (IO, n=10; non-IO, n=10) was compared to a non-cleft control group (n=8) of the same age. It was expected that treatment with IO would positively influence speech quality because IO facilitates the development of more normal speech patterns. In order to assess all relevant characteristics of cleft palate speech, the panel of expert listeners blindly judged 13 specific speech aspects on seven-point EAI scales. Furthermore, they indicated their total impression of the children’s speech quality on a ten-point scale. Finally, they rated the number of speech therapy sessions that in their opinion were needed in the year following the assessment. The intra- and interrater reliability was good for 12 of the rating scales. Three scales: ‘fronting’, ‘nasal snort’, and ‘nasal realisation’ were excluded from analysis because of low intra- or interrater reliability. Statistically significant differences between the non-cleft group and the cleft groups were present on all 12 scales except for the characteristics ‘palatalisation’ and ‘lateralisation’. These speech characteristics were equally present in all three groups. All other speech errors were scored as being most distinct in the two groups with cleft. The noncleft children’s speech obtained the highest scores for ‘intelligibility’ and ‘correctness of articulation’. There was only one speech aspect that distinguished the IO group from the non-IO group, i.e., the scale ‘intelligibility’. The IO group obtained significantly higher intelligibility ratings than the non-IO group, indicating that the itelligibility of the speech in the IO group was judged to be superior by the expert listeners. Similar results were presented in Chapter 4, in which the IO children also obtained higher ratings for their speech intelligibility scored by lay listeners.
Chapter 6 focuses on the language development. Speech and language development are closely inter-related, and children with better speech may develop language skills more easily through mediation of more adequate feedback. Therefore, language development was hypothesised to benefit from IO treatment. At the age of 2, 2.5, and 3 years language development was evaluated in 12 children (6 IO and 6 non-IO). Eleven of these children (6 IO and 5 non-IO) were also assessed in a follow-up at age 6. Receptive language skills were assessed by means of the Dutch version of the Reynell Developmental Language Scales. Expressive language skills of the toddlers were evaluated by calculating mean length of utterance (MLU) and mean length of longest utterances (MLLU). In the 6-year-olds the expressive language skills were measured by means of standardised Dutch language tests (Taaltests voor Kinderen and Schlichting test). The results showed no differences in receptive language skills between the IO and non-IO group. The expressive language measures MLU and MLLU, however, were influenced by IO treatment. At age 2.5 and 3 years, the children who were treated with IO in their first year of life produced statistically significant longer utterances than the non-IO group. In the follow-up at 6 years, the difference in expressive language between the two groups was no longer significant. Hence, in this relatively small group, IO treatment did not prove to have long lasting effects on language development.
The next chapter (chapter 7) discusses the relationship between phonetic characteristics in the pre-lexical period and development of speech and language at age 2.5 years in the light of IO treatment. Such a relationship may have important implications for therapeutic management and speech intervention, and it may help to understand the effects of IO on speech and language development. In this evaluation only the two groups of children with UCLP were involved. Phonetic characteristics that were included as predictors were the use of alveolar contoids and oral plosives in babbling at age 1 and 1.5. Outcome variables were measured at 2.5 years and comprised two aspects of phonological development (i.e., the number of acquired consonants and the complexity of the phonological system of contrasts), expert judgements on four rating scales (i.e., intelligibility, palatalisation, lateralisation, and backing), and the mean length of utterance (MLU). The results indicated that a higher use of oral plosives in babbling of 1.5-year-old children with UCLP was associated with better intelligibility and higher MLU at age 2.5. Furthermore, a higher occurrence of alveolar contoids in babbling at age 1.5 predicted less compensatory articulation (i.e., less palatalisation, less lateralisation, and less backing) in speech at age 2.5. It could not be established in this investigation whether the correlation between alveolar contoids in babbling at age 1.5 and less compensatory articulation at age 2.5 was directly related to treatment with the infant orthopaedic.
Subsequently, in chapter 8 the cost-effectiveness analysis of IO treatment compared to no such treatment focusing on the speech of the children at the age of 2.5 years is described. The measure for effectiveness on speech was the ‘total impression of speech quality’ rated by the panel of expert listeners (see Chapter 5). There was a statistically significant beneficial effect of IO on this effect measure. Since a statistically significant difference may not be synonymous with a clinically important change, the clinical importance was determined by calculating the magnitude or the size of the effect produced by the intervention. The effect size in this investigation showed that the effect of IO on the ‘total impression of speech quality’ was large, indicating that the improvement may be considered as a clinically important change. Obviously, the costs for treatment by the orthodontist in the IO group were significantly higher than in the non-IO group. For both groups the mean cost were related to the mean rating for ‘total impression of speech quality’. The resulting cost-effectiveness for IO compared to non-IO was €1,041 for 1.34 point speech quality improvement. Relative to the costs that are spent on the comprehensive treatment of children with cleft lip and palate, the financial investment that is necessary to obtain this improvement seems limited. Moreover, since it is plausible that children in the IO group would need less intensive speech therapy, part of the costs for IO treatment may be outweighed by costs prevented for speech therapy in later years. Thus it was concluded that from the perspective of speech development the costeffectiveness of IO over non-IO seems acceptable.
Finally, in chapter 9, the general discussion, additional methodological considerations are given regarding the evaluation of speech and language in this study. It describes a number of difficulties related to the lack of standardised measuring instruments in the study of speech pathology. Subsequently, implications for therapeutic management are indicated. This section concludes with the statement that IO treatment may be regarded as a useful component of the treatment of infants with unilateral cleft lip and palate, as far as speech and language development is concerned. The general discussion ends with suggestions for further research.