Electropalatography for articulation disorders associated with cleft palate
Abstract
This is a protocol for a Cochrane Review (Intervention). The objectives are as follows:
To assess the effectiveness of electropalatography (EPG) in improving the treatment of articulation errors in individuals with repaired cleft palate.
Background
Cleft palate is the most common birth defect and the most common congenital deformity of the face (Kummer 2001). Its prevalence varies from 1:500 to 1:2000, depending on racial background and gender (Marazita 2004). Cleft palate occurs "when the roof of the mouth has not joined completely...it can range from just an opening at the back of the soft palate to a nearly complete separation of the roof of the mouth (soft and hard palate)" (CLAPA 2001). Cleft palate results from abnormal facial development during the fourth to twelve weeks of the gestation period. The cleft may occur on one side (unilateral) or on both sides (bilateral). Cleft palate is usually closed surgically by the age of 18 months (Seagle 2004; Watson 2001). However, the timing of surgical intervention varies between centres.
Cleft palate and syndromes
Cleft palate may be associated with other congenital anomalies, and may occur as a part of a well‐defined syndrome such as velocardiofacial syndrome. To date, there are more than 400 syndromes which include cleft palate as a feature (Winter 2000) and this accounts for about 30% of the cleft cases. Individuals born with cleft palate are more likely to have cognitive deficits than normal children, with an even higher risk in children born with a syndrome (Endriga 1999), possibly due to aberrant brain development (Nopoulos 2002) and other factors. Cognitive problems associated with various types of learning disabilities can add to the detrimental effect of palatal cleft on speech acquisition.
Compensatory misarticulations
The impact of palatal cleft on speech acquisition is apparent early, at the stage of pre‐speech vocalisations and the influence may continue even after surgical intervention (Peterson 2006). Normal articulation is expected in 25% of preschoolers with repaired cleft palate who receive routine care by a cleft palate team, but "a significant number of individuals will continue to demonstrate problems with articulation in adolescence" (Peterson 2001). Compensatory misarticulations are one of the prominent articulation problems observed in individuals born with cleft palate. The prevalence of compensatory misarticulations in English‐speaking children born with cleft palate varies from 22% to 28% (Dalston 1992; Hardin‐Jones 2005; Peterson 1990) and a recent study reported a prevalence of 28.5% in Greek‐speaking children (Paliobei 2005). Compensatory misarticulations are learned articulation deviations and most of them are errors in place of articulation (Peterson 2006). For example, the sound /t/ is backed from the alveolar region to the velar region, hence, listeners perceived a /k/ sound. They are believed to result from strategies developed by individuals with cleft palate as a response to inadequate intraoral pressure for normal articulation caused by the structural deficit. Once learned, these abnormal learned patterns can persist due to habituation (Peterson 2001). Since compensatory misarticulations are due to abnormal learning, they are under the speaker's control and, therefore, can respond to speech therapy (Kummer 2001).
Speech therapy
It is estimated that between 50‐75% of individuals born with cleft palate require speech intervention at some point in their lives (Peterson 2001; Witzel 1991). Based on a comprehensive speech assessment, speech‐language therapists (SLTs) apply phonetic and phonological approaches and principles of behaviour modification in speech therapy (Sell 2001). The SLTs teach directly the target speech sounds to the individuals, with the aid of different techniques and tools such as tactile cues, the use of mirror, and diagram of the oral cavity. Once the target production is established, the sound is practiced and stabilized in a hierarchical progression of speech contexts, from vowels through words, phrases, sentences to spontaneous speech (Peterson 2006). The SLTs also work with the individuals on their speech sound discrimination and self‐monitoring skills in order to establish internal mechanisms for correct target selection and perceptual‐motor self‐monitoring (Sell 2001). However, not all individuals with compensatory errors respond well to this standard treatment approach; posterior articulation pattern persists in some individuals despite of years of speech therapy (Sell 2001). Especially in adolescents and adults, modification is more likely to occur if some type of biofeedback therapy, such as speech therapy using electropalatography (EPG), is provided.
Using EPG for speech assessment and therapy
EPG has been used to assess lingual articulation and to treat articulatory disorders in individuals with cleft palate through the use of visual feedback during speech therapy. It is a computer‐based instrument which gives information on the location and timing of the tongue's contact with the hard palate during speech (Hardcastle 1991; Hardcastle 1997). Each patient undergoing EPG assessment or therapy has to wear a custom‐made artificial dental plate which is moulded to fit the speaker's hard palate. The dental plate is embedded with 62 electrodes on the lingual surface. When the electrodes are contacted by the tongue, a signal is sent to an external processing unit through lead‐out wires and real‐time visual feedback of the location and time of tongue‐palate contacts is shown on a computer monitor.
EPG has been used in speech therapy with individuals who have persistent articulation problems and are unresponsive to standard therapy procedures (Stengelhofen 1990). The real‐time visual feedback provided by EPG has also been employed for treating abnormal posterior articulations in individuals with cleft palate. Although there exists a substantial literature about speech intervention using EPG, there has not been a systematic review of the effectiveness of EPG, hence, it remains difficult for SLTs to make decisions about whether or not to use EPG in their clinical practice.
This review aims to assess the effectiveness of EPG in speech therapy for managing articulation errors for individuals with articulation disorders related to cleft palate.
Objectives
To assess the effectiveness of electropalatography (EPG) in improving the treatment of articulation errors in individuals with repaired cleft palate.
Methods
Criteria for considering studies for this review
Types of studies
Randomised controlled studies in which participants have been randomly allocated to an experimental and a control group by, for example, using a random number table; or quasi‐randomised controlled studies where, for example, participants have been alternately allocated to an experimental and a control group.
Types of participants
Individuals with articulation problems associated with cleft palate. Included will be individuals with recognized syndromes (e.g. velocardiofacial syndrome, Treacher Collins syndromes, etc). There will be no restriction on the age range.
Those with learning disability associated with severely limited expressive/spoken language (as judged by SLTs using informal or standardised language test) or severe hearing impairment will be excluded from the review purely on the assumption that a certain level of cognitive abilities of the individuals is required in order for them to benefit from EPG therapy.
Types of interventions
Interventions using EPG compared to no treatment, delayed treatment, standard treatment, or alternative treatment techniques (e.g. treatment using ultrasound).
There will be no restriction on the frequency, intensity and duration of speech intervention.
Types of outcome measures
The primary outcome will be the measure of correct articulation of speech sounds targeted in therapy, which could be assessed by:
* Standardised articulation tests (e.g. Edinburgh Articulation Test);
* Other perceptual evaluation of articulation (e.g. Cleft Audit Protocol for Speech, percentage of consonants correct);
* Articulatory accuracy based on instrumental measures such as electropalatography (EPG);
* Measures of speech intelligibility (e.g. by means of a perceptual rating scale, or by calculating the number of words correctly transcribed by examiner).
Secondary outcomes will include:
* Measures of listener acceptability (e.g. by using a perceptual rating scale);
* Participants' perceptions of impact of wearing the EPG plate on (i) speech production, (ii) tongue movement, (iii) sensation in the mouth, (iv) self‐perceived appearance, (v) gagging, and (vi) saliva;
* Adverse effects.
Outcome measures of interest will not form part of the inclusion criteria.
Search methods for identification of studies
Electronic searches
Searches will be run on the following databases:
-
The Cochrane Central Register of Controlled Trials (CENTRAL)
-
MEDLINE
-
EMBASE
-
ERIC
-
PsycINFO
-
CINAHL
-
Linguistics and Language Behaviour Abstracts (LLBA)
-
Allied and Complementary Medicine (AMED)
-
Latin American Health Sciences Literature (LILACS)
-
Japana Centra Revuo Medicina
-
The National Research Register
-
ClinicalTrials.gov
-
Dissertation Abstracts
The following search terms will be used to identify articles.Appropriate MeSH terms will be used where available.. No language or date restrictions will be applied.
1. cleft palate
2. electropalatograph*
3. EPG
4. palatograph*
5. palatomet*
6. 2 or 3 or 4 or 5
7. 1 and 6
Other searches
The following relevant journals will be handsearched from 1980 onwards. The Master List of Journals (http://www.cochrane.us/masterlist.asp) was checked to ensure that they had not been handsearched by The Cochrane Collaboration.
-
Clinical Linguistics and Phonetics
-
Cleft Palate Journal/ Cleft Palate‐Craniofacial Journal
-
International Journal of Language and Communication Disorders
Reference lists of all articles that are identified as relevant will be reviewed.
Colleagues and researchers will be contacted to identify other possible published and unpublished studies such as technical or research reports, conference papers, and different types of dissertations.
There will be no language restrictions; translations will be sought when necessary.
Data collection and analysis
Selection of trials
Two reviewers (FG and AL) will conduct the search independently. All references generated from the search strategy will be managed using a reference management programme. The reviewers (FG and AL) will independently assess all titles and abstracts against the inclusion criteria. Articles that appear to meet the inclusion criteria will be retrieved for full text version, whereas articles which do not fulfil the inclusion criteria will be discarded. In case it is uncertain whether an article meets the inclusion criteria, the full text version will be obtained. Further details regarding the study will be obtained by contacting the authors if necessary. In the event of disagreement over inclusion of any individual study, FG and AL will seek to reach a consensus by discussion and if unable to do so, the full article will be reviewed by JL. Reviewers will not be blinded to the name(s) of the author(s), institution(s) or publication source at any level of review.
Data management and extraction
A paper data extraction form will be developed and piloted. Any subsequent versions of the form will include revision dates. FG and AL will independently extract the following data from the articles:
1. Participants
-
Number of participants
-
Age
-
Sex
-
Inclusion and/or exclusion criteria
-
Level of severity of articulation disorders
-
Reason(s) for patients being rejected, if applicable
-
Other baseline characteristics of the participants reported e.g. hearing ability
2. Methods
-
type(s) of speech assessment conducted
-
assessment findings (e.g. number and types of articulation errors, such as double articulation, palatalization)
3. Interventions
-
Type of interventions
-
Length of intervention i.e. number of hours of the course of speech therapy
-
Number of therapy session
-
Whether compliance evaluated
4. Outcomes
-
Total numbers and events for each intervention group or risk ratio and its standard error (binary outcomes)
-
Mean and standard deviation (continuous data
Any uncertainty and disagreement will be resolved through discussion and consultation with JL. If additional information is needed, the first author of the study will be contacted.
Risk of bias
FG and AL will independently assess the risk of bias in the included studies under the following six domains. Any disagreements between the two reviewers will be resolved by discussion. A third reviewer (JL) will be consulted if a consensus could not be reached. The assessment of risk of bias will consist two parts: (1) a succinct description, which will include verbatim quotes from reports or correspondence and/or a comment from the reviewer, regarding the procedures conducted to avoid bias; and (2) a judgement of the risk of bias based on the summary in part 1. The judgement criteria for each domain are described below.
A judgement of 'Yes' means that there was low risk of bias in the study; whereas the judgment of 'No' indicates a high risk of bias in the included study. A judgement of 'Unclear' will be given to study if the risk of bias was uncertain due to insufficient information available for the judgement.
1) Sequence generation
The method used to generate the allocation sequence will be described using quotes wherever possible (e.g. "patients were randomly allocated"). Comments such as "probably done", "probably not done" will be added to supplement an ambiguous quote. The included studies will be assigned to one of the following categories:
(A) indicates adequate method was used for randomisation e.g. using computer generated or table of random numbers.
(B) indicates uncertainty about whether appropriate method of randomisation was used.
(C) indicates that inadequate method of was randomisation used e.g. case file number, date of birth, or alternate numbers.
2) Allocation concealment
The included studies will be assigned to one of the following quality criteria whereby:
(A) indicates adequate concealment of allocation e.g. prenumbered or coded identical containers administered serially to participants.
(B) indicates uncertainty about whether the allocation was adequately concealed e.g. the authors did not describe the allocation methods.
(C) indicates that the allocation was not adequately concealed e.g. alternate assignment.
In those studies assessed as B or C, the pre‐treatment assessment and the allocation of participants will be described in the review (Description of Studies) in order to identify differences between intervention and control groups that can be ascertained at baseline.
3) Blinding
Measures used to ensure blinding of (a) key investigators and (b) outcome assessor(s) from knowledge of which intervention a participant had received will be described and evaluated. This domain will be graded as yes (A), unclear (B), and no (C). Assessment will be made for each main outcome (e.g. outcome measured at 6 months post‐therapy, outcome measured at 12 months post‐therapy).
4)Incomplete outcome data
Incomplete outcome data due to attrition (drop‐out) during the study, or exclusions from the analysis.
Protection from attrition bias:
Attrition described, proportion smaller than 10% of assigned patients and difference between the two groups not greater than 10%.
Reason(s) of participants withdrawal explained
Attrition appropriately analysed (e.g. intention‐to‐treat analysis)
The review authors will grade each of these criteria as yes (A), unclear (B), and no (C).
5)Selective reporting bias
Also referred to as within‐study publication bias. Funnel plots (effect size against standard error) will be drawn if sufficient studies are found. An asymmetric funnel indicates a relationship between effect size and study size, which suggests the possibility of either publication bias or a systematic difference between smaller and larger studies. If a relationship is identified, the clinical diversity of the studies will also be examined (Egger 1997).
6) Other sources of bias
Measures used to ensure the groups compared in the study were similar at baseline in terms of severity of articulation problems and other baseline characteristics (e.g. hearing ability) will be described and evaluated using the same scheme ‐ yes (A), unclear (B), and no (C).
Data analysis
The following comparisons will be made:
1. EPG therapy versus No treatment controls
2. EPG therapy versus delayed treatment
3. EPG therapy versus standard speech therapy
4. EPG therapy versus alternative treatment technique
Missing data
Authors will be contacted and asked to supply missing data. In the event that the authors could not be contacted or data could not be supplied, missing data and drop‐outs will be assessed for each included study and the number of participants who were included in the final analysis as a proportion of all participants in each study will be reported. Where known, reasons for missing data will be provided.
Binary data
Binary data may be likely (e.g. "articulation improved versus no change"). The data will be analysed by calculating the risk ratio.
Continuous data
Standardised articulation test results, articulation accuracy based on EPG or perceptual evaluation, judgement of speech intelligibility and listener acceptability will be treated as continuous data. Weighted mean difference (WMD, or the 'difference in means') will be calculated if the outcome measurements in all studies are made on the same scale. Otherwise, standardised mean differences (SMD) will be used to combine studies that measured the same outcome using different methods.
Assessment of Heterogeneity
Clinical heterogeneity will be assessed by examining the types of participants, the severity of articulation errors, interventions, and outcome measures as specified in the criteria for included studies. The consistency across studies will be assessed using the Chi‐squared test for heterogeneity, through visual inspection of forest plots, and by using I2 test (Higgins 2002, Higgins 2003). I2 is a quantity that describes the percentage of variability in point estimates that is due to heterogeneity rather than sampling error. If there is evidence of heterogeneity (i.e. an I2 value of 25% or greater), the possible causes of heterogeneity will be explored by examining the studies and data again and where appropriate conducting subgroup analyses.
Meta‐analysis
Meta‐analysis, using the RevMan 4.2 (Update Software) statistical package, will be carried out if there are sufficient data. A fixed effects model will be used for analysis. If there is statistical heterogeneity which cannot be explained (e.g. by means of subgroups), a random effects meta‐analysis will be used to incorporate heterogeneity among studies. A random effects meta‐analysis model involves an assumption that the effects being estimated in the different studies are not identical, but follow some common distribution.
Subgroup analysis
If possible, subgroup analyses will be conducted to assess the impact of individual's age and intensity of therapy (number and frequency of speech therapy sessions). Interpretation of the findings from multiple subgroup analyses will consider the possibility of false positive findings (Higgins 2005).
Sensitivity analysis
Sensitivity analyses will be performed to assess the robustness of conclusions by examining the impact of study quality. The factors that are considered as important to study quality include randomisation, blinding to outcome assessment, and attrition (Juni 2001). Studies which are categorised as (A) or (B) for these factors in the methodological quality assessment will be included in the analysis.
