JCPSLP Vol 18 No. 1Mar 2016

described above. For the purposes of this study, velopharyngeal function was categorised by the researchers into two groups: (a) adequate; or (b) not adequate based on assessment data and clinical observations contained within written speech pathology assessment reports. Written reports indicating that velopharyngeal function was ‘probably adequate’ and ‘borderline’ were categorised as inadequate by the researchers. The assessment data and clinical observations reported by speech pathologists were informed by ratings of resonance (i.e., hyper-nasal or hypo-nasal), and nasal airflow (i.e., nasal emission/turbulence). Middle ear function Tympanometry is a clinical measure of middle ear function, which is used to examine the condition of the middle ear, mobility of the tympanic membrane (or ear drum), and the volume of the ear canal (Hamid & Brookler, 2007). For the purposes of this study, middle ear function was categorised into three groups: Type A tympanometry in at least one ear, suggestive of normal middle ear function; Type B tympanometry, suggestive of middle ear dysfunction and often associated with the presence of middle ear effusion; and Type C tympanometry, suggestive of significant negative middle ear pressure, which may indicate middle ear pathology. Procedure Ethical approval was obtained from The University of Queensland Behavioural and Social Sciences Ethics Review Committee, and the relevant hospital human research ethics committee. After gaining approval to access an established clinical database at the hospital, a search was conducted to identify children who met the inclusion criteria. After identifying all potential participants, information on their demographics as well as on the variables of interest was obtained from the database and medical charts namely: phonological awareness skills, language skills, speech production skills, velopharyngeal function, and middle ear function. Medical charts were reviewed to confirm existing information in the database and retrieve missing data relevant to the variables of interest. Participants with complete datasets were subsequently All analyses were conducted using the Statistical Package for Social Sciences (SPSS) Version 21.0 for Windows. To address the first aim of the study, descriptive statistics were used to explore the phonological awareness skills of participants’ according to the five subtests of the PIPA. A one-way repeated measures ANOVA was conducted to compare scores on these five subtests of the PIPA, with PIPA subtest scores and composite PIPA scores used to explore the phonological awareness skills of participants. Regression analysis was used to address the second aim of the study, namely to investigate the relationship between phonological awareness skills and language standard scores (a continuous variable), and the three ordinal categorical variables, namely speech diagnosis, velopharyngeal function, and middle ear function. Prior to analysis, normality of the dependent variable, phonological awareness, was examined and confirmed. Examination of the data also revealed no outlying data scores. The variables of interest (i.e., language standard scores, speech diagnosis, velopharyngeal function, and middle ear function) were initially selected as possible candidate variables for a multivariate regression model. Bivariate analyses were included in this study. Data analysis

audio-taped. Audiologists completed the assessment of middle ear function. Each assessment is described below. Phonological awareness The Preschool and Primary Inventory of Phonological Awareness (PIPA; Dodd, McIntosh, Teitzel, & Ozanne,, 2000) consists of five subtests normed on Australian children, namely Syllable Segmentation (the ability to identify syllables within a word), Rhyme Awareness (the ability to identify similar sounding word endings), Alliteration Awareness (the ability to identify and produce words beginning with the same sound), Phoneme Isolation (the ability to recognise individual sounds in words), and Phoneme Segmentation (the ability to segment individual sounds of a word). In each subtest of the PIPA there are 12 items. A standard score between 7 and 13 indicates skills within normal limits. For the purposes of this study, the standard scores from each PIPA subtest were added to form a composite score for analysis. Speech production Participants’ speech production skills were assessed using the Diagnostic Evaluation of Articulation and Phonology (DEAP; Dodd, Hua, Crosbie, Holm, & Ozanne, 2002), the Great Ormond Street Speech Assessment (Gos.Sp.Ass.; Sell, Harding, & Grunwell, 1999), or the Cleft Audit Protocol for Speech – Augmented (CAPS-A; John, Sell, Sweeney, Harding-Bell, &Williams, 2006). The Phonology Assessment component of the DEAP was administered with 21 participants. The sentences section of the Gos. Sp.Ass. and/or CAPS-A was administered with the remaining 9 participants. For the purposes of analysis, speech diagnosis was classified into two categorical variables, namely speech difficulties and no speech difficulties. This categorisation was informed by the presence or absence of delayed and/or disordered speech production processes, including consideration of the presence or absence of both articulatory errors and phonological processes. Language The Clinical Evaluation of Language Fundamentals – Preschool 2nd Edition (CELF-P2; Wiig, Secord, & Semel,, 2004) or the Clinical Evaluation of Language Fundamentals 4th Edition (CELF-4; Semel, Wiig, & Secord,, 2003) were used to formally assess participants’ language skills depending on the age of the child at assessment. The age range for the CELF-P is 3;0 to 6;11 years, whereas the CELF-4 is designed for children aged 5;0 to 21;0 years. Due to the clinical nature of the data collected for this study, the decision to use either the CELF-P or the CELF-4 was dependent on the child’s level of language skills and ability to attend to stimuli. Due to variability in attention and compliance, subtests administered with each participant varied; however, each participant completed the minimum number of relevant subtests to derive either a Core Language Score (CLS) or a Receptive Language Index (RLI). The CLS (n = 15) or RLI (n = 15) scores were used for the purposes of statistical analysis in this study. Expressive language scores were not included for analysis as scores may have been affected by reduced intelligibility or articulation errors associated with structural issues in a cleft palate population, particularly with morphemes including high pressure consonants (e.g., /s/ and /z/ in particular). Velopharngeal function Velopharyngeal function was assessed using the Gos.Sp.Ass (Sell et al., 1999) or CAPS-A (John et al., 2006) as

36

JCPSLP Volume 18, Number 1 2016

www.speechpathologyaustralia.org.au