Cleft Palate ± Lip (CP±L) is one of the most well-known congenital anomalies worldwide. Children and adolescences with CP±L are at high risk of developing speech and communication difficulties due to a range of cleft-related etiologies. Even with timely or early repair of the palate, children with CP±L may still experience cleft-related speech difficulties with a notable proportion of children having continuing velopharyngeal insufficiency (VPI). VPI refers to the presence of incomplete velopharyngeal closure during speech production. Non-articulation parameters of VPI include hypernasality, nasal airflow errors, and grimace, with hypernasality being the hallmark characteristic of VPI. Speech error patterns attributed to continuing VPI may involve what is termed as active compensatory articulatory gestures. For instance, children with cleft palate speech/VPI may have difficulty building up sufficient oral pressure to produce oral targets and may attempt to achieve closure at the level of the glottis resulting in replacement of oral pressure targets with glottal stops or fricatives. The loss of distinctive features and the ability to make meaningful contrasts potentially impacts on global outcomes such as speech understandability and acceptability adversely. In fact, there was still no sufficient evidence to support the efficacy of specific intervention approaches or techniques. Additionally, there are only very few studies that have systematically compared different intervention approaches. Consensus on the effectiveness of different approaches remains inconclusive.

It is of vital importance to implement evidence-based speech intervention to achieve the optimal speech and resonance outcomes. Historically, speech intervention approaches can be categorized into motor-phonetic (MP) (+/- principles of motor learning) and linguistic-phonological (LP). Motor-phonetic approaches attempt to modify the phonetic errors individually by explicit instruction of the articulatory movement and extensive amount of practice. Motor-phonetic approaches can be undertaken with or without the application of principles of motor learning, which emphasize intensive practice, optimal task difficulty, and variable practice conditions for better retention and generalization. However, the generalization effect was solely evident for the targeted individual phonemes. Increasingly, cleft palate speech/VPI speech disorder is being viewed as a phonological impairment rather than simply an articulation disorder. Linguistic-phonological approaches, on the other hand, refer to phonological rule-based interventions, which attributes the child's speech sound errors to a disordered internal phonological system. Such approaches, target multiple speech sounds or sound classes, and assumes that there would be generalization across phonemes according to their distinctive features. Examples of LP approaches include the minimal pair approach, maximal oppositions and treatment of the empty set, multiple oppositions and Metaphon. Current evidence has suggested that LP approach resulted in a more superior generalization effects in terms of the speech outcomes within the same class and the overall percentage of speech accuracy. Nevertheless, it is crucial to recognize that these LP approaches still predominantly regard speech as a linear sequence of segments and speech sound errors as a set of rules. Target selection tends to be focused on individual phonemic errors in the children's phonetic repertoire and selected specific phonemes for treatment.

In contrast, non-linear phonological (NLP) intervention seeks to capture the complexity of speech sounds organized in a hierarchical structure. It is common to observe a range of phonemes or even classes of phonemes being replaced with a single glottal stop or fricative in cleft palate speech. This results in an extensive collapse of meaningful contrasts in the individual's speech, impacting negatively on speech understandability. As such, target selection in cleft speech intervention would be a critical component for successful treatment outcomes. NLP intervention views a word as a multi-tiered structure and emphasizes the broader nature of the phonological system. With a holistic viewpoint of the speech sound system, the target selection and the treatment direction based on NLP framework would be different. Different research had investigated the use of non-linear phonology in speech intervention. Some of them had revealed the positive outcomes. It is believed that speech intervention guided by non-linear phonology framework would facilitate wider changes in the phonological system of children with cleft palate speech / VPI speech disorder, resulting in a more significant generalization effect and increased overall speech intelligibility. Therefore, the current study would like to investigate the effectiveness of different types speech intervention approaches in children and adolescences with CP±L.

Introduction of the study Cleft Palate with/without Lip (CP±L) Cleft Palate with or without Lip (CP±L) is one of the most well-known congenital anomalies worldwide. Epidemiological studies indicate that it affects approximately 1 in every 600 newborn babies, with rate of occurrences reported at 1.33 to 2.23 per 1000 live and stillbirths in the Chinese population. CP±L can present as syndromic or non-syndromic. Children with CP±L are at high risk of developing speech and communication difficulties due to a range of cleft-related etiologies. Some studies have also showed that children with CP±L may have language and literacy difficulties. Research showed that this group of children require speech-language intervention at different stages across their lifespan. Speech articulation, intelligibility, velopharyngeal competence and psychosocial states have been some of the recognized outcome measurements in this field.

Cleft Palate Speech/ Velopharyngeal Insufficiency Even with timely or early repair of the palate, children with CP+L may still experience cleft-related speech difficulties with a notable proportion of children having continuing velopharyngeal insufficiency (VPI). VPI refers to the presence of incomplete velopharyngeal closure during speech production. Non-articulation parameters of VPI include hypernasality, nasal airflow errors, and grimace, with hypernasality being the hallmark characteristic of VPI. Speech error patterns attributed to continuing VPI may involve what is termed as active compensatory articulatory gestures. For instance, children with cleft palate speech/VPI may have difficulty building up sufficient oral pressure to produce oral targets and may attempt to achieve closure at the level of the glottis resulting in replacement of oral pressure targets with glottal stops or fricatives. The loss of distinctive features and the ability to make meaningful contrasts potentially impacts on global outcomes such as speech understandability and acceptability adversely. Another possible articulation outcome of VPI is weak/nasalized consonants or complete replacement of oral pressure targets with their homorganic nasal counterparts e.g., /t/ replaced with /n/. Speech articulation outcomes unrelated to VPI also include lateralization errors and/or backing errors. A possible etiology of this is the presence of an oronasal fistula. In the latter context for instance, the child attempts to avoid the oronasal fistula by moving the tongue posterior to the fistula.

Direct Speech Intervention Current treatment approaches and the evidence In fact, treatment or treatment efficacy studies in cleft palate/VPI remain limited. A systematic review by Bessell et al. (2013) concluded that there was no sufficient evidence to support the efficacy of specific intervention approaches or techniques. There are only very few studies that have systematically compared different intervention approaches. Consensus on the effectiveness of different approaches remains inconclusive. The heterogeneity of the treatment approaches was also revealed by a subsequent exploratory study by Williams et al. (2021), which identified 49 distinct intervention methods reported by Speech Therapists in the UK, including strategies targeting very young children, such as complexity approaches, cued articulation, and cycles approaches. A more recent systematic review and meta-analysis by Sand et al (2021) also reiterated the low overall quality of evidence in treatment studies addressing cleft/VPI speech disorders, urging the need for interventions that consider the global benefits for clients. Treatment outcomes may include the overall speech intelligibility and communicative participation in everyday settings. With that said, a standardized outcome set is also advocated for future studies to enable a more holistic analysis of the treatment outcomes.

It is of vital importance to implement evidence-based speech intervention to achieve the optimal speech and resonance outcomes. Research has shown that active errors including lateralization, palatalization of fricatives and glottal articulation are habituated and active compensatory articulatory patterns that can be readjusted responsively by speech therapy intervention. On the contrary, passive errors, such as weak and/or nasalized consonants due to VPI, require surgical intervention. The use of glottal articulation may also reflect VPI and both surgical intervention and speech intervention are needed. Historically, speech intervention approaches can be categorized into motor-phonetic (MP) (+/- principles of motor learning) and linguistic-phonological (LP). Motor-phonetic approaches attempt to modify the phonetic errors individually by explicit instruction of the articulatory movement and extensive amount of practice. Motor-phonetic approaches can be undertaken with or without the application of principles of motor learning, which emphasize intensive practice, optimal task difficulty, and variable practice conditions for better retention and generalization. A case study by Hanley et al (2023) demonstrated that MP approach incorporating PML was found to be effective in improving the accuracy of both active and passive cleft speech errors in children with cleft palate. Generalization was also seen in untreated items after 8 weeks of face-to-face twice-weekly sessions. However, the necessity of conducting treatment on the passive cleft type errors remains questionable. Moreover, the generalization effect was solely evident for the targeted individual phonemes.

Increasingly, cleft palate speech/VPI speech disorder is being viewed as a phonological impairment rather than simply an articulation disorder. Howard et al. (2019) also stressed the phonological consequences of mechanical/structural constraints such as VPI in children with CP±L. Linguistic-phonological approaches, on the other hand, refer to phonological rule-based interventions, which attributes the child's speech sound errors to a disordered internal phonological system. Such approaches, target multiple speech sounds or sound classes, and assumes that there would be generalization across phonemes according to their distinctive features. Examples of LP approaches include the minimal pair approach, maximal oppositions and treatment of the empty set, multiple oppositions and Metaphon. A case study by Anderson et al (2024) employed the multiple oppositions approach to treat non-oral and passive CSCs. It revealed that the multiple oppositions intervention, as an LP approach, was found to be effective in improving the speech accuracy and intelligibility at conversational level, with sustained results also observed during the maintenance phase. The results suggested that the LP approach brought about a system-wide change of the phonological system in children with CP±L. Despite these findings, the limitations of small sample sizes in studies warrant caution in interpreting the generalizability of the results. Another randomized control trial with a larger sample size by Alighieri et al (2025) compared the effectiveness of MP and LP approaches (i.e. modified Metaphon approach). They found that while both MP and LP approaches could lead to generalization, LP approach resulted in a more superior generalization effects in terms of the speech outcomes within the same class and the overall percentage of speech accuracy. Nevertheless, it is crucial to recognize that these LP approaches still predominantly regard speech as a linear sequence of segments and speech sound errors as a set of rules. Target selection tends to be focused on individual phonemic errors in the children's phonetic repertoire and selected specific phonemes for treatment.

An Alternative Treatment Framework Application of non-linear phonology Phonological intervention approaches described above are often implemented under a linear paradigm, focusing on isolated phonemes. In contrast, non-linear phonological intervention, developed by Bernhardt and Stemberger (1994), seeks to capture the complexity of speech sounds organized in a hierarchical structure. It is common to observe a range of phonemes or even classes of phonemes being replaced with a single glottal stop or fricative in cleft palate speech. This consequently results in an extensive/ significant collapse of meaningful contrasts in the individual's speech, thereby VPI impacting negatively on speech understandability. As such, the premise of speech intervention would be to maximize overall understandability, acceptability and intelligibility. Target selection in cleft speech intervention would, therefore, be a critical component for successful treatment outcomes. Non-linear phonological intervention views a word as a multi-tiered structure. It emphasizes the broader nature of the phonological system and it believes that the system is more complex than speech sounds alon. With a holistic viewpoint of the speech sound system, the target selection and the treatment direction based on non-linear phonological framework would be different. Different research had investigated the use of non-linear phonology in speech intervention. Some of them had revealed the positive outcomes. It is believed that speech intervention guided by non-linear phonology framework would facilitate wider changes in the phonological system of children with cleft palate speech / VPI speech disorder. Subsequently, a more significant generalization effect across sound classes would be expected so that the overall speech intelligibility would be maximized. Yet, the application of non-linear phonology in the field of cleft/VPI population is yet to be adopted widely.

Research Questions

Indeed, SLPs reported a high degree of variability in selecting treatment approaches in their practice for children with CP±L. Such discrepancies may also be attributed to the lack of consensus in target selection and subsequent treatment directions. In view of this, the present study aims to conduct a randomized controlled trial to investigate the effectiveness of non-linear phonological intervention approach in improving speech outcomes in Cantonese-speaking children with CP±L and compare the effectiveness of LP approaches based on linear and non-linear phonology. The study will also further examine the phonological skills and functional impact on the individuals with CP±L after speech intervention. The followings are the proposed research questions:

(i) Is non-linear phonological intervention effective in improving speech outcomes in Cantonese-speaking children and adolescences with CP±L? Ho: Non-linear phonological intervention has no effect on improving speech outcomes in Cantonese-speaking children and adolescences with CP±L.

H1: Non-linear phonological intervention is effective in improving speech outcomes in Cantonese-speaking children and adolescences with CP±L.

(ii) Is a speech intervention approach based on non-linear phonological framework more effective in improving speech outcomes than the one based on linear phonology in Cantonese-speaking children and adolescences with CP±L? Ho: A speech intervention approach based on a non-linear phonological framework is not more effective in improving speech outcomes than one based on linear phonology in Cantonese-speaking children and adolescences with CP±L.

H1: A speech intervention approach based on a non-linear phonological framework is more effective in improving speech outcomes than one based on linear phonology in Cantonese-speaking children and adolescences with CP±L.

(iii) Are the phonological skills improved after phonological intervention in both groups? Ho: Phonological skills do not show significant improvement after phonological intervention in both groups of participants.

H1: Phonological skills show significant improvement after phonological intervention in both groups of participants.

(iv) Are the patient-reported outcomes also improved with the speech outcomes followed by the treatment? Ho: The patient-reported outcomes are not improved with the enhanced articulatory accuracy after treatment.

H1: The patient-reported outcomes are improved with the enhanced articulatory accuracy after treatment.

Inclusion Criteria:

• (i) age of 4 to 17;
• (ii) primary repair of the cleft palate±lip done prior to 24 months old;
• (iii) presence of at least one CSCs type errors which must include non-oral CSCs and/or posterior CSCs;
• (iv) language proficiency in Cantonese.

Exclusion Criteria:

• (i) a cognitive impairment;
• (ii) a hearing impairment based on pure-tone audiometry;
• (iii) known diagnosis of language difficulties/disorders;
• (iv) and/or other medical conditions that would affect their abilities to participate in the speech tasks independently.