See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/345813376

Differential Diagnosis of Childhood Apraxia of Speech Compared to Other

Speech Sound Disorders: A Systematic Review

Article in Journal of Speech Language and Hearing Research · November 2020

DOI: 10.1044/2020_AJSLP-20-00063

CITATIONS READS

8 1,298

5 authors, including:

Elizabeth Murray Jenya Iuzzini-Seigel

The University of Sydney Marquette University
22 PUBLICATIONS 493 CITATIONS 25 PUBLICATIONS 373 CITATIONS

SEE PROFILE SEE PROFILE

Edwin Maas Hayo Terband

Temple University Utrecht University
57 PUBLICATIONS 1,636 CITATIONS 43 PUBLICATIONS 591 CITATIONS

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

A survey of adults who were diagnosed with CAS as children. View project

New method for diagnosing acquired apraxia of speech - http://goo.gl/J429Ca View project

All content following this page was uploaded by Jenya Iuzzini-Seigel on 13 November 2020.

The user has requested enhancement of the downloaded file.

Note: This document is a post-print of a manuscript that has been accepted for publication at AJSLP. This
version has been peer-reviewed,but has not received final edits from the publisher.

Differential diagnosis of Childhood Apraxia of Speech compared to other Speech Sound Disorders:
A Systematic Review.

Authors: Elizabeth Murray1,2, Jenya Iuzzini-Seigel3, Edwin Maas4, Hayo Terband5, Kirrie J. Ballard1

Affiliation of all authors:

1
The University of Sydney, Sydney, Australia
2
Remarkable Speech + Movement, Sydney, Australia
3
Marquette University, Milwaukee WI, USA
4
Temple University, Philadelphia PA Pennsylvania, USA
5
Utrecht Institute of Linguistics - OTS, Utrecht University, Utrecht, the Netherlands

Corresponding author:

Elizabeth Murray
Faculty of Health Sciences
The University of Sydney
PO Box 170
Lidcombe 1820
Phone: +61 2 9351 9780
Email address: [email protected]

ORCID IDs:

Elizabeth Murray - https://orcid.org/0000-0002-0883-155X

Jenya Iuzzini-Seigel - https://orcid.org/0000-0002-7679-2556

Edwin Maas - https://orcid.org/0000-0003-4452-2196

Hayo Terband - https://orcid.org/0000-0001-7265-3711

Kirrie J. Ballard - https://orcid.org/0000-0002-9917-5390

Keywords: (not in title)

Dyspraxia, Assessment, Sensitivity/specificity, Phonology, Dysarthria

[Type here]

Conflict of Interest Statement:

There are no conflicts of interest. Authors did not review their own manuscripts in this systematic review.

Funding Statement:

There was no funding for this work.

Abstract

Purpose: To determine the discriminative features that might contribute to differentiation of childhood

apraxia of speech (CAS) from other speech sound disorders (SSD).

Method: A comprehensive literature search was conducted for articles or doctoral dissertations that

included ≥1 child with CAS and ≥1 child with SSD. Of 2071 publications screened, 53 met the criteria.

Articles were assessed for (a) study design and risk of bias; (b) participant characteristics and confidence

in diagnosis and (c) discriminative perceptual, acoustic or kinematic measures. A criterion was used to

identify promising studies: AAN study design (Class III+), replicable participant descriptions and

adequate confidence in diagnosis (≥3), and ≥1 discriminative and reliable measure.

Results: Over 75% of studies were retrospective, case-control designs and/or assessed English-speaking

children. Many studies did not fully describe study design and quality. No studies met the Class I

(highest) quality rating according to AAN guidelines. CAS was most compared to speech

delay/phonological disorder. Only 6 studies had diagnostic confidence ratings of 1 (best). Twenty-six

studies reported discriminative perceptual measures, 14 reported discriminative acoustic markers, and 4

reported discriminative kinematic markers. Measures were diverse and only two studies directly

replicated previous findings. Overall 7 studies met the quality criteria and another 8 nearly met the study

criteria to warrant further investigation.

2
[Type here]

Conclusions: There are no studies of the highest diagnostic quality. There are 15 studies that can

contribute to further diagnostic efforts discriminating CAS from other SSD. Future research should utilize

careful diagnostic design, support replication and adhere to standard reporting guidelines.

Plain language summary

Childhood apraxia of speech (CAS) is a motor speech disorder where children have difficulties planning

movement to speak clearly. CAS is hard to differentiate from other speech problems. This review looked

for evidence-based measures to help contribute to determining CAS from other speech problems.

The systematic review reviewed 53 articles found after a thorough search and screening process. Authors

rated studies based on their study quality, descriptions of the children who participated in the studies and

the measures that discriminated CAS from other speech problems.

The results showed no studies met the highest quality rating. Most studies used data collected in the past

and deliberately compared select groups rather than looking across a sample representative of the

population. Many studies did not fully describe study design and quality. CAS was most compared to

speech delay/phonological disorder, a meaning-based speech disorder. Overall 15 studies reported

measures that discriminated groups with enough quality to warrant further testing and research.

3
[Type here]

Children with speech sound disorders (SSDs) have “gaps or simplifications in their speech sound

systems that can make what they say difficult to understand” (Bowen, 2015 p. 3). SSDs are one of the

most prevalent types of communication difficulties of early childhood (Eadie et al., 2015). They

constitute between 30-75% of a pediatric speech-language pathologists’ (SLPs) caseload (ASHA, 2016;

Broomfield & Dodd, 2004; McLeod & Baker, 2014). SSD is an umbrella term for a wide range of

difficulties due to known etiology (e.g., craniofacial anomalies, hearing impairment, down syndrome and

other neurodevelopmental or genetic anomalies), or unknown etiology (IEPMCS, 2012). The focus here is

on one specific SSD - childhood apraxia of speech (CAS; also known as developmental apraxia of speech

and developmental verbal dyspraxia). CAS is a developmental, neurological SSD that affects motor

planning and/or programming (ASHA, 2007). The impairment in children with CAS can be described as a

deficit in transforming phonological codes into motor speech commands (ASHA, 2007; Terband,

Maassen, Guenther, & Brumberg, 2009) essentially the early stages of planning the movements that will

generate the intended speech sounds and sound movement sequences. The outcome of this deficit is

unintelligible, inconsistent and robotic speech due to difficulties timing the articulators and transitioning

from sound to sound and syllable to syllable (ASHA, 2007). Like other SSDs, the impact of CAS is not

only on speech production. It adversely affects social communication and increases the risk of bullying,

mental health concerns, and difficulties with phonological awareness and literacy skill development

(Lewis, Freebairn, Hansen, lyengar, & Taylor, 2004; Murray & Iuzzini-Seigel, 2017; Rusiewicz, Maize,

& Ptakowski, 2018). It also can negatively impact academic and occupational success (Carrigg, Parry,

Baker, Shriberg, & Ballard, 2016; Lewis, Freebairn, Hansen, Taylor, et al., 2004).

An initial assessment battery (Macrae, 2016) is needed to assess a child suspected of any SSD to

determine (a) if they have a SSD, (b) what type of SSD they have and (c) what evidence-based treatments

may be of greatest benefit to them. SLPs have effective means to determine if a child has an SSD

compared to typical development, using speech samples assessed with published developmental norms

and considering functional needs (Storkel, 2019). Where the client matches the population sample,

4
[Type here]

standardized tests are used (e.g., Goldman & Fristoe, 2015). However, differential diagnosis of CAS from

other SSDs remains a clinical challenge.

There are currently no evidence-based diagnostic guidelines, criteria, or markers for reliably

differentiating CAS from a range of other SSDs (Dodd, 2014; Iuzzini-Seigel & Murray, 2017). The key

challenges are the current reliance on perceptually-based feature lists for differential diagnosis that lack

operational definitions of each feature or critical thresholds for degree, frequency, and context of these

features; as well as a lack of a psychometrically robust assessment tool (e.g., ASHA, 2007). This is in the

face of overlapping symptoms across SSDs (McCabe, Rosenthal, & McLeod, 1998; Rupela, Velleman, &

Andrianopoulos, 2016), high rates of comorbidity (Iuzzini-Seigel, 2019; Liégeois & Morgan, 2012) and

variability across and within children over time (Maassen, Nijland, & Terband, 2012). Despite these

challenges, we need effective differential diagnosis to provide appropriate information to families and

provide treatments that target the specific underlying impairment(s).

There are many descriptive and diagnostic studies that have been published since CAS was first

mentioned in the literature by Morley, Court, and Miller (1954). As a starting point, the aim of the current

study is to systematically review this literature to evaluate the rigor of this body of work and determine

any potential markers, procedures or tools that can support future research and clinical efforts to

accurately differentiate children with CAS from other SSDs.

Existing literature reviews on the diagnosis of CAS

There have been three previous literature reviews specifically focused on diagnostic methods

involving children with CAS, all of which focused on published assessment tools. In the first, McCauley

and Strand (2008) assessed six standardized tests of oral non-verbal and speech motor performance for

their reliability and methods of interpreting results. Their threshold of acceptability included acceptable

normative data, clear-cut behavioral standards for decision making, adequate information describing

participants and statistics used, a reliability coefficient (e.g., intra-class correlation coefficient) of 0.90 or

higher for inter-rater reliability and clear methods of content, construct and validity. The Verbal Motor

Production Assessment for Children (VMPAC, Hayden & Square-Storer, 1999) met more criteria than the

5
[Type here]

other tests, yet still failed to meet six of the eight criteria. For example, the VMPAC only contained

norms for typically developing children, lacked statistical examination of construct validity and provided

limited guidance on interpreting the test for therapy planning or measuring change over time. Overall,

none of the tools met the criteria to be recommended for clinical use. Secondly, Gubiani, Pagliarin, and

Keske-Soares (2015) reviewed published assessment tools for CAS used in any journal article for either

initial diagnosis or differential diagnosis purposes to determine tools for Brazilian Portuguese speakers.

They identified 5 assessment tools: VMPAC, Dynamic Evaluation of Motor Speech Skills (DEMSS),

Kaufman Speech Praxis Test, The Orofacial Praxis Test, and the Madison Speech Assessment Protocol

(MSAP) and described each of these without critical appraisal. They concluded that none were adapted

and standardized for use with clients who speak Brazilian Portuguese, but that the DEMSS is suitable for

adaptation as it has a defined protocol and known reliability and validity. Finally, Sayahi and Jalaie

(2016) used a brief systematic search strategy of four databases and identified 13 studies fitting their

search criteria. The authors agreed with McCauley and Strand (2008) that the VMPAC was the most

reliable test to use. They also assessed the frequency of clinical markers in the diagnosis of CAS and

found inconsistency was the most frequently used marker in the studies they reviewed. However, there

were multiple journal articles and tools that were not identified (e.g., the DEMSS was not found in their

search). They also did not critically review the methodology or outcomes of the studies, which is essential

for determining which protocols and assessment tools are sufficiently reliable, valid and sensitive/specific

to be implemented clinically. Thus, there is a need for an extensive review of the peer-reviewed literature

for procedures or markers under evaluation. A review of experimental studies is warranted to determine

which protocols and behavioral markers show promise for sensitive and specific differential diagnosis of

CAS, and which can be implemented both in clinical and research practice with greater confidence.

Methodology and quality rating of differential diagnostic studies

Critical evaluation of the CAS versus SSD diagnostic literature requires examination of

methodology with respect to study design, internal validity, and risk of bias. Study design can be

categorized by levels of evidence (Merlin, Weston, & Tooher, 2009; OCEBM, 2016). Besides systematic

6
[Type here]

reviews (level 1), the best individual diagnostic designs (level 2) are cross-sectional (aka cohort studies

that do not require a longitudinal component) studies. The intention is to assess a large group of people

with a defined set of clinical features (e.g., any speech errors) and further classify participants using data-

driven methods. To qualify as level 2, a study needs to apply a clinical gold standard (reference test) to

compare to the measures being evaluated (index test) to determine measures and diagnosis, with blinding

and consecutive entry into the study; otherwise, it is classified as level 3. Case-control designs where one

group of people with an a priori established diagnosis (e.g., children with CAS) are compared to another

group without that diagnosis (e.g., children with phonological disorder) or with a different diagnosis are

level 4. Finally, mechanism-based studies and early diagnostic yield studies without an established

reference test provide the lowest level of evidence (level 5). These classification systems are aimed at

medical and instrumental diagnostic accuracy studies and the levels of evidence are driven from medical

research. Specifically taking behavioral studies of neurological disorders into consideration, the American

Academy of Neurology (AAN, 2011) developed their own class system for diagnostic studies (reported in

Table 1). We have adopted this system for classifying overall study quality as the broader categories are

better suitable for evaluating behavioral studies rather than instrumental diagnostic studies at this stage.

Nonetheless, the medical levels of evidence provide us direction on how to improve diagnostic methods

in the future.

Regarding internal validity and risk of bias, there is only one critiquing tool to assess a published

study (i.e., Quality Assessment of Diagnostic Accuracy Studies – 2 or QUADAS-2; Whiting et al., 2011)

and one reporting guideline for writing up a diagnostic study for publication (i.e., Standards for Reporting

of Diagnostic Accuracy Studies (i.e., Standards for Reporting of Diagnostic Accuracy Studies or STARD,

Cohen et al., 2016). These again focus on medical instrumental assessments and have only been available

this decade. Both report the need for (a) an index test (i.e., a diagnostic test that is being evaluated) to be

compared to an established reference test with an appropriate interval between tests; (b) blinding of

results between the index and reference tests; (c) inclusion of methods for reducing bias in participant

enrollment, including details of enrollment method (e.g., consecutive, random, or convenience sampling);

7
[Type here]

and (d) statement of whether the data collection was planned prospectively or retrospectively, relative to

when the reference test was completed. The focus regarding index versus reference tests is, importantly,

on establishing whether testing is biased by circularity; that is, whether the index and reference tests are

non-independent because they are based on the same measures. These tools also recommend authors

provide clear descriptions of the tests/measures used and reporting the outcome of the assessment,

indicating who was diagnosed with what disorder following the test and the statistical approach applied.

Ideally, studies provide cross-tabulation of results across tests and diagnostic accuracy (e.g., 95%

confidence intervals, calculation of sensitivity and specificity, odds ratios and/or likelihood ratios).

The AAN (2011) also provides valuable information on critiquing behavioral studies. Here, we

have used the AAN data extraction process to examine each study identified in our systematic search

supplemented with the above risk of bias features from the QUADAS-2 and STARD. To further evaluate

the replicability of study procedures by others, we have also reported on the level of detail provided for

dependent measures and the inter-rater reliability of measurements, given that many measures of speech

production rely on perceptual judgment which is prone to several biases (Kent, 1996). Finally, we also

considered the level of detail on participants’ characteristics, which is crucial for all diagnostic studies

(Whiting et al., 2011).

-------------------------------------

Insert Table 1 about here

-------------------------------------

Participant descriptions are especially relevant for this review considering that (a) the current

gold-standard in diagnosing CAS is expert judgment on presence of a set of perceptually judged speech

features (ASHA, 2007; Murray, McCabe, & Ballard, 2015) and (b) CAS can be comorbid with other

speech and language disorders, which can complicate interpretation of performance (e.g., Murray,

Thomas, & McKechnie, 2019). For these reasons, we evaluated participant descriptions of each

participant group in terms of sex, severity, age, prior therapy, comorbidity, and other assessment results

including genetic testing. For each study, we also specified the inclusion criteria applied and the initial

8
[Type here]

diagnostic criteria used (i.e., reference test), and we made a judgment as to whether these descriptions

were enough to allow replication. Finally, we also considered the level of confidence in the gold standard

diagnosis (i.e., outcome of the contemporary ASHA (2007) three core-features of CAS reference test).

This is a difficult reference test because of the lack of operational definitions provided in the ASHA

(2007) technical report (see Terband, Maassen, & Maas, 2019). The level of confidence in the diagnosis

was therefore determined through the method of Murray et al. (2015), adapted from (Wambaugh, Duffy,

McNeil, Robin, & Rogers, 2006) using discriminative features of CAS based on ASHA (2007). These

detailed analyses will assist identification of discriminative measures or markers that warrant further

exploration in a future population-based prospective study.

Aims of the present study

The systematic review was completed by the Apraxia of Speech Writing Group, which is part of

the Evidence-Based Clinical Research Committee of the Academy of Neurological Communication

Disorders and Sciences. This systematic review evaluated studies published through December 2018 with

the overall aim of identifying measures that differentiate CAS from at least one other SSD. To meet this

overall aim, our specific aims were to determine the following:

1. Study design quality: the design type, replicability and inter-rater reliability of measures,

circularity (reference vs. index test), and statistical methods used (AAN, 2011).

2. Adequacy of participant descriptions: the level of detail in descriptions of the participants and

level of confidence in the initial diagnoses made by authors (i.e. outcome of reference test).

3. Discriminative diagnostic variables (features/markers): the perceptual, acoustic and/or

kinematic measures that statistically differentiated CAS from the comparison SSD(s).

Method

This systematic review is registered with PROSPERO, registration number: CRD42017056616.

This review was completed using the PRISMA (Moher, Liberati, Tetzlaff, & Altman, 2009) and PRISMA

for Diagnostic Accuracy studies (McInnes et al., 2018) which recommend minimal reporting guidelines

for systematic reviews.

9
[Type here]

Systematic search strategy

We followed the multi-step process of PRISMA (Moher et al., 2009) to find studies that met the

inclusion criteria (see eligibility below). Figure 1 presents the flow chart of the study selection procedure.

-------------------------------------

Insert Figure 1 about here

-------------------------------------

Identification

A comprehensive search of nine databases related to speech-language pathology was completed.

These were Allied and Complementary Medicine (AMED), Cumulative Index to Nursing and Allied

Health Literature (CINAHL), Education Resources Information Center (ERIC), Linguistic Language

Behavior Abstracts (LLBA), Medline, PsycINFO, Scopus, Web of Science and Google Scholar.

Search terms were developed based on the inclusion criteria and in accordance with MeSH

headings so they could be used across databases. The terms were: [“apraxi*” or “dyspraxi*” or

“childhood apraxia of speech” or “motor speech” AND “child*” or “develop*” AND “diagnos*” or

“assess*” or “different*” or “classif*” or “evalua*” or “suspect*” or “marker” or “discrimin* AND

“speech” or “communicat*” or “language” or “articulat*” or “inconsisten*” or “intelligib*” or “prosod*”

or “kinemat*” or “speak*”. Specific MeSH keywords used were: “apraxia, developmental verbal”,

“apraxia, verbal”, “dyspraxia, verbal”, “Developmental Verbal Dyspraxia” – Qualifier – Diagnosis (DI),

classification (CL)]. A total of 4121 references were found.

Screening

References were exported to Endnote X8.2 (2014) for screening. Duplicate references were

removed (n= 2050), leaving 2071 for screening. References were checked and excluded in the following

order: (1) CAS not present, (2) animal subjects only, (3) theoretical / review studies with no primary data,

(4) adult participants only, (5) treatment (not a diagnostic study) and (6) qualitative and/or survey study.

Searches were completed within Endnote of title, abstract and keywords for the terms within the list

above. References were screened by reading the title and then the abstract and for some that were not

10
[Type here]

clear, obtaining and checking the full text as needed. Inter-rater reliability was completed by the first

author and a research assistant for 241 articles (11% of the articles) with 100% agreement. Screening

removed 1878 articles (see Figure 1), leaving 193 studies to be assessed for eligibility.

Eligibility

The full texts were obtained and assessed against the final inclusion criteria before being

reviewed. This review included studies that met all the inclusion criteria, as follows. Participants included

(a) people from birth to 18 years of age; (b) at least one participant diagnosed with CAS (diagnosis could

be made before or during the study) and (c) at least one other participant with a different speech sound

disorder (SSD) for comparison. Synonyms were included for CAS (e.g., Developmental Apraxia of

Speech, Developmental Verbal Dyspraxia, and Verbal Dyspraxia) and a full range of terms relating to

SSDs (e.g., speech/ articulation/ phonological delay or disorder or impairment, inconsistent phonological

disorder, dysarthria (including paediatric or subtypes), and oral apraxia1. Articles included (d) were

written in any language with participants and researchers speaking any language and (e) were either

published articles or doctoral dissertations not published as a journal article. Studies included were

designed to (a) assess CAS compared to other speech sound disorders for differential diagnostic purposes,

or (b) determine markers or measures for speech and/or genotype-phenotype relationships.

From the 193 articles, 53 articles were identified for review. Intra-rater reliability (first author) for

inclusion of 20% of these studies was assessed again at four months after the first decision had been

made. Percent agreement was 98% (n= 49/50). Inter-rater reliability with an independent rater (research

assistant) on 100% of the articles was 97% (n = 187/193; see Supplemental Appendix 2 for the list of

excluded articles and the reason for exclusion). The eight articles with intra- or inter-rater disagreements

were assessed by the entire author team for consensus, with 7/8 excluded. Excluded articles were not

further analyzed.

Data processing & analysis

1
Please note in this study we use ‘phonological disorder’ to refer to phonological impairment/ disorder and
developmental language disorder to refer to language delay for readability.

11
[Type here]

The authors each rated a subset of the 53 included studies, using a procedural manual developed

by the authors, entering ratings into an Excel spreadsheet (Supplemental Appendix 1). The manual was

based on Wambaugh et al (2006) and the variables are presented according to each aim below. Raters

were not blinded to authors of the study as raters were already familiar with the research area. Raters did

not assess any paper they authored or were involved in. Where possible, an author rated studies by the

same research team to assist with determining the evolution of research programs over time and to assist

with methodological details that may have been reported in previous studies.

Five of the 53 studies were rated for reliability by all five authors; two initially after training,

another two mid-way in rating to ensure consistency and then a final one towards the end of the rating

process. . Inter-rater reliability across the five studies for all coding was 84% (SD= 5.1%, range: 79 to

89%). Discrepancies were resolved by consensus and the manual and spreadsheet notes were reviewed to

ensure any issues determined did not affect ratings. If an aspect of a study was difficult to rate, we

discussed this at team meetings and resolved by consensus.

Study Quality (aim 1).

The included studies were assessed for the quality of the diagnostic research design used using

definitions of research studies (cohort studies, case-control studies or other) from (AAN, 2011 pg. 10). To

distinguish further among studies, raters assessed whether there was a clear index (assessment being

tested) compared to a reference (currently used comparison measure) and whether group assignment was

made a priori (diagnosed beforehand) or not (suspected group or had risk factors associated with CAS or

SSD). Raters assessed if each study provided inclusion criteria and diagnostic criteria for CAS and the

other SSDs. Methodological quality was also assessed, in regards to whether the outcome or diagnostic

measures tested were replicable (i.e. could be used again clinically or in future research), reliable (could

be completed by other people with similar results), blinded and were assessed with appropriate statistical

tests/methods.

Statistical analysis, where completed, was assessed by determining whether measures were

parametric or non-parametric, whether testing of assumptions was reported, and whether the analysis was

12
[Type here]

appropriate for the data and question (e.g., use of t-test or ANOVAs to compare between groups). If raters

were unsure in a methodology used, they investigated the method and/or discussed this as a group. From

this analysis, it was possible to compute an overall rating of the diagnostic studies based on American

Academy of Neurology guidelines (AAN, 2011; see Table 1).

Participant descriptions (aim 2).

Raters assessed each study's description of CAS and SSD participant characteristics. Variables

recorded were number of participants, age, sex, severity and etiology of speech disorder, comorbid

diagnoses, country of origin and language spoken, the inclusion criteria of the studies and diagnostic

indicators. The following was also assessed as to whether or not the information was reported for

participants in the studies: any medications taken, neuroimaging data, genetic data, socio-economic status,

cognitive functioning, language functioning, hearing and vision, and if so, described the participant’s

characteristics.

Confidence in CAS diagnosis.

The confidence of diagnosis was assessed using procedures from (Murray, Mc Cabe, Heard, &

Ballard, 2015) based on (Wambaugh et al., 2006) (See Supplemental Appendix 4). For the CAS

participants, confidence was established using the criteria in Murray et al (2015) based on description of

primary features using the three consensus-based CAS features listed in the ASHA technical report

(2007). This was compared to non-discriminative features, which were those shared with other SSDs,

such as poor intelligibility, slow progress, or delayed language (ASHA, 2007; McCabe et al., 1998). For

the SSD participants, the criteria covered the same five levels as CAS with generic descriptors that were

flexible across SSDs. Clear cases of comorbid disorders were also noted. A rating of 1 indicated high

confidence in CAS or SSD diagnosis and a rating of 5 indicated no confidence. Intra-rater reliability (first

author) on CAS diagnosis was 100% and SSD diagnosis was 100% from the four studies used for

calculating reliability. Inter-rater reliability between all raters for the CAS diagnosis was 90% and SSD

diagnosis was 90% for the same four studies. Discrepancies were resolved by consensus and the manual

was reviewed to ensure any issues determined did not affect subsequent ratings.

13
[Type here]

Discriminative diagnostic variables (features/markers) (aim 3).

IMPORTANTE
Raters classified the experimental measures from each study as either perceptual, acoustic and/or

kinematic measures, and documented the specific measures collected. Further assessment determined

whether any measures were discriminative in differentiating CAS from any type of SSD/s, and whether

both discriminative and non-discriminative measures had acceptable reliability. Our benchmarks for

reliability were: percent agreement > 85% (Kratochwill et al., 2010), Kappa statistics >60% (McHugh,

2012) and intra-class correlation coefficients > 0.60 (Hallgren, 2012). Discriminative measures were also

examined for specificity (true positives) and sensitivity (true negatives) values of ≥ 0.90 as per the

standard in clinical medicine (Shriberg & et al., 1997a)

Data analysis

Descriptive statistics were documented for the assessments of interest reported previously. A

statistical meta-analysis of data was not possible with this dataset due to the differing measures,

participants and methodologies across studies. Instead, studies with adequate quality and that

demonstrated variables of diagnostic promise were collated using the criteria in Figure 2. As many

promising papers were close but did not meet the full set of criteria, a second table of papers that (1)

failed to meet one quality criterion (e.g., replication of participants) and/or (2) papers that reported up to

two quality criteria in another source (e.g., reliability of perceptual and acoustic measures were reported

in a previous study by the same research group) were also collated as potential measures for further

research and clinical use. These papers were then described in greater detail within aim 3 to determine

promising discriminative features that differentiated CAS from other SSDs. Quality ratings were

computed by the first and second authors with (n = 53/55) 95% agreement and the two disagreements

were resolved by consensus.

-------------------------------------

Insert Figure 2 about here

-------------------------------------

Results

14
[Type here]

For readability, each criterion assessed is presented in order of most to least rigorous and

references are given in the text for no more than 5 studies. The 53 articles reviewed are listed in

Supplemental Appendix 3. Key findings are presented in summary tables, while the full analyses per

article assessed can be found in Supplemental Appendix 1 (Excel spreadsheet).

Study Quality (aim 1)

Of the 53 studies, 11 (20.8%) used a cohort study design with three (5.7%) using a statistical

multivariate data-driven, prospective method to derive subgroups from a larger, unselected sample of

children with SSD (Peter, 2006; Strand, McCauley, Weigand, Stoeckel, & Baas, 2013; Vick et al., 2014).

Of these three, two compared the resulting subgroups against clinical diagnosis (Peter, 2006; Strand et al.,

2013), whereas the other compared the resulting subgroups against classification criteria reported in the

literature (Vick et al., 2014). Additionally, 41 (77.4%) employed a case-control design and one (1.9%)

involved a case description (Hayden, 1994).

Of the 41 case-control design studies, all but three divided children into subgroups a priori

(before analysis) and then compared these groups on the index tests (i.e. the measure/s of interest being

assessed - e.g., the DEMSS [Strand et al., 2013]). The basis for classification into subgroups varied

considerably with different subgroups and criteria being used. Two studies (Williams, Ingham, &

Rosenthal, 1981; Yoss & Darley, 1974) formed their groups after analysis differentiating ‘functional’

articulation disorder (with no organic disability) versus developmental apraxia of speech (i.e. CAS), and

for one study it was not stated whether groups were formed a priori (Barry, 1995).

In 23 studies (43.4%), there were indications of diagnostic circularity (in which aspects of the

index test were also used in classification by the reference test). In 26 studies (49.1%) there was no

diagnostic circularity, and in the remaining 4 studies (7.5%), it could not be determined whether there was

diagnostic circularity.

Eleven of the 53 studies (20.8%) were prospective; prospective studies were defined as studies in

which data analysis was planned before data collection on index and reference tests (Cohen et al., 2016).

15
[Type here]

The remaining 42 studies were either retrospective (n = 9; 17.0%) or did not clearly state whether the

study was retrospective or prospective (n = 33; 62.3%).

Participants were selected via consecutive sampling in two studies (3.8%; Strand et al., 2013;

Williams et al., 1981) and via convenience sampling in 12 others (22.6%). Most studies (n = 39; 73.6%)

did not explicitly state their sampling method, although descriptions of recruitment methods suggest

convenience sampling in virtually all these studies.

Ten studies (18.9%) clearly indicated that the index test and reference test were administered by

separate, blinded examiners. In two studies (3.8%), the assessors were explicitly not blinded (Aziz,

Shohdi, Osman, & Habib, 2010; Strand et al., 2013). In the remaining cases (n = 41; 77.4%), we were

unable to determine whether the assessors who administered or scored the index test were blinded to the

group classification status based on the reference test; this was in part due to the tendency for articles to

use passive voice construction in reporting procedures, without specifying who conducted the testing

(e.g., “Children were evaluated in a quiet room”).

Only three studies (5.7%) indicated the time period between the index and reference tests (Lewis,

Freebairn, Hansen, lyengar, et al. (2004): approximately 4 years within a longitudinal study; Mei et al.

[2018]: hours; Strand et al. [2013]: weeks). The remaining 50 studies (94.3%) did not state the time

period between index and reference tests.

Overall, of the 53 studies, none met study quality criteria for a Class I rating (AAN, 2011), three

(5.7%) met criteria for a Class II rating (Shriberg et al., 2017a, 2017b; Thoonen, Maassen, Gabreëls, &

Schreuder, 1999) 16 (30.2%) met criteria for a Class III rating, and the remaining 34 (64.2%) were rated

as Class IV studies (see Table 1 for the criteria).

Participant Descriptions (aim 2)

A total of 4,213 participants were included across studies, with an average sample size of 79

participants (median = 41), and a range from 3 (two studies: Betz & Stoel-Gammon, 2005; Smith,

Marquardt, Cannito, & Davis, 1994) to 665 (Lewis et al., 2018). Of these 4,213 participants, 845 (20.1%)

were children with CAS. It should be noted that these numbers do not necessarily reflect the number of

16
[Type here]

unique individuals, as different studies sometimes report data from the same children (e.g., Bradford,

Murdoch, Thompson, & Stokes, 1997; Lewis et al., 2018; Shriberg et al., 2003). Given the lack of details

in some reports, the number of unique children reported in the 53 studies cannot be determined. The

average sample size of children with CAS was 16 (median = 11), with a range from 1 (four studies: Betz

& Stoel-Gammon, 2005; Hayden, 1994; Smith et al., 1994; Terband, Zaalen, & Maassen, 2012) to 63

(Lewis et al., 2018).

The comparison groups varied across studies, and several studies included multiple comparison

groups. All but five studies (48/53, 90.6%) included a comparison group consisting of children with

speech delay/phonological disorder (total of 1,802 children across studies; sample size mean = 38, median

= 13, range = 1 – 251). Other comparison groups included were children with dysarthria (seven studies;

total 99 children; mean sample size = 14, median = 9, range = 1 – 36), children with developmental

language disorder without SSD (four studies; total 54 children, mean sample size = 14, median = 12,

range = 7 – 23), children with developmental language disorder and SSD (three studies; total 226

children, mean = 75, median = 14, range = 14 – 170), children who stutter (one study; sample size = 31),

children with autism spectrum disorder (one study; sample size = 46), and adults with acquired apraxia of

speech (two studies; total 53 participants, range = 22 – 31). Thirty-one studies (58.5%) also included a

comparison group of typically developing children (total 1,041 children; mean sample size = 34, median

= 12, range = 1 – 255), and three also included adults without speech/language disorders (total 46

participants, mean = 15, median = 18, range 10 – 18).

Only 34 studies (64.2%) reported the sex of participants; together, there were 2,822 participants

in these studies. Of the 519 children with CAS reported in these 34 studies, 148 were girls (28.5%) and

371 were boys (71.5%), for a F:M ratio of 1:2.51. In three studies, the sex distribution of the non-CAS

participants was not specified for all children (Lewis, Freebairn, Hansen, lyengar, et al., 2004; Shriberg &

et al., 1997b; Shriberg, Potter, & Strand, 2011); sex was not specified for a total of 63 children across

these studies. Thus, the sex distribution of the non-typical non-CAS comparison group below is based on

a total of 1,567 participants. Among these participants, there were 504 girls (32.2%) and 1063 boys

17
[Type here]

(67.8%), for a F:M ratio of 1:2.11. The language background of the children in the studies was

predominantly English (41/53, 77.4%), with five studies involving Dutch-speaking children (9.4%), and

one study each (1.9%) for Arabic, French, German, and Portuguese; three studies (5.7%) did not report

the language of the children.

Only 24 studies (45.3%) reported disorder severity, and severity in these studies ranged from mild

to severe (for both CAS and non-CAS comparison groups). The ages of the children ranged from 3 to 18

years (in both CAS and non-CAS comparison groups); mean age could not be calculated because multiple

studies reported only age ranges. Etiology for both CAS and non-CAS groups was not reported or was

unknown in most studies, with only ten studies reporting genetic origins for CAS from genetic test

results.

Twenty-seven studies (50.9%) reported that the children with speech disorders had received prior

speech therapy. Thirty-two studies (60.4%) reported on cognitive function, 39 studies (73.6%) reported

on language status, 44 studies (83.0%) reported on hearing status, and 10 studies (18.9%) reported on

vision status of participants. Only four studies (7.5%) reported on socioeconomic status or education, and

no studies (0.0%) reported medication history or neuroimaging findings. Twenty-eight studies (52.8%)

reported comorbid conditions for children with CAS; in most cases, the comorbid condition was language

disorder, but several studies also reported dysarthria, cognitive impairments, atypical oral motor function

(e.g., oral apraxia, fine motor), and hearing loss.

Diagnostic Criteria and Confidence

Seventeen studies (32.1%) failed to explicitly report criteria for diagnosis of CAS; seven of these

studies (13.2% of all studies) referred to other articles for the diagnostic criteria used, in some cases to

other articles in a series (Shriberg et al., 2017a, 2017b). For comparison groups, 16 studies (30.2%) failed

to provide criteria; five of these (9.4% of all studies) referred to other sources for diagnostic criteria used.

Some studies that did report diagnostic criteria were not sufficiently specific or clear to allow replication.

In total, only 26 studies (49.1%) provided sufficient information to enable replication with respect to

sample characteristics; 27 studies (50.9%) were deemed unreplicable based on the diagnostic information

18
[Type here]

provided (one of which would be considered replicable based on reference to other sources (Shriberg et

al., 2011).

Confidence in accuracy of CAS diagnosis spanned the range from 1 to 5 (see Appendix for

definitions), with only seven studies (13.2%) receiving a rating of 1 (Iuzzini-Seigel, Hogan Tiffany, &

Green Jordan, 2017; Iuzzini-Seigel, Hogan, Guarino, & Green, 2015; Maas & Mailend, 2017; Mei et al.,

2018; Murray, Mc Cabe, et al., 2015; Peter, 2006; Zuk, Iuzzini-Seigel, Cabbage, Green, & Hogan, 2018).

No studies received a rating of 2; 18 studies (34.0%) received a rating of 3; 12 (22.6%) received a rating

of 4; and the remaining 16 (30.2%) received a rating of 5. The average rating across studies was 3.57. For

the comparison groups, confidence in diagnosis also spanned the range. For the non-typical comparison

groups, nine studies (17.0%) included groups with a confidence rating of 1; zero (0.0%) with ratings of 2;

seven (13.2%) with ratings of 3; 21 (39.6%) with a rating of 4; and 16 (30.2%) with a rating of 5. The

average rating across studies was 3.66.

Statistical Analysis

Of the 53 studies, 20 (37.7%) did not report reliability data for their indexmeasures. For three

additional studies, reliability of index measures was not reported because measurement did not depend on

human judgment such as transcription or signal segmentation; one involved the Iowa Oral Performance

Instrument (Potter, Nievergelt, & Shriberg, 2013) and the others involved tasks of speech perception

where children with CAS responded to speech tasks (e.g., rhyming discrimination or differentiating d/g

sounds presented on a computer (Nijland, 2009; Zuk et al., 2018). Five further studies (9.4%) referred to

reliability data reported elsewhere. In all, 25 studies (47.2%) reported reliability of index measures or the

basis for their index measures; most studies reported reliability of transcription or acoustic segmentation

rather than the reliability of the specific measures derived from these raw data (e.g., errors, formant

values).

Inferential statistics were used in 38 of the 53 studies (71.7%) and not in the remaining 15

(28.3%). The nature of these statistical tests varied considerably across studies, although ANOVAs, t-

19
[Type here]

tests, and chi-square tests were common. In many cases, it could not be determined from the information

provided in the paper whether statistical assumptions were met, but statistical approaches were considered

generally appropriate for the research design employed in 30/38 studies (78.9%).

Discriminative diagnostic variables (features/markers) (aim 3)

For the 53 studies examined, study variables were categorized as perceptual, acoustic, or

kinematic. Most studies examined perceptual variables (n = 43), including a large number (n = 14) of

experiments that also investigated acoustic and/or kinematic variables; 31 studies reported perceptual

features that were discriminative. Twenty-four studies investigated acoustic variables including 18 that

also examined perceptual and/or kinematic variables; 14 of these studies reported acoustic variables that

were discriminative. Finally, seven studies examined kinematic variables, four of which also investigated

perceptual and/or acoustic variables; four studies reported discriminative kinematic variables. Variables

were further classified based on the type of variable they explored (e.g., auditory perception, prosody) to

demonstrate the breadth and focus of the literature.

Of the 53 studies, 7 met the quality criteria for promising studies based on (1) adequate study

quality (AAN, 2011 Class III or better), (2) adequate diagnostic confidence for participants with CAS

(ratings ≤3), (2) replicable participants, (3), and (4) report of at least 1 reliable discriminative

feature/marker/variable. Table 2 summarizes the seven papers that met these criteria. Table 3 includes an

additional eight papers that were close to meeting all our criteria but (1) failed to meet one quality

criterion (e.g., replicability of participants) and/or (2) papers that referenced up to two quality criteria

from another source (e.g., reliability of perceptual and acoustic measures), rather than explicitly reporting

them within the paper. Both tables report sensitivity and specificity for the discriminating

markers/variables where available. Results below summarize the findings from these 15 papers, organized

by the type of feature or measure found to be discriminative for CAS (e.g. auditory perception;

articulation, speech movements and rate; literacy; maximum performance; nonspeech and prosody) for

comparison. Sensitivity and specificity data where calculated are reported in Table 2 and 3.

-------------------------------------

20
[Type here]

Insert Tables 2 and 3 about here

-------------------------------------

Articulation, Speech Movements, and Rate (including pausing).

Twelve studies (80%) reported discriminative variables that reflected articulation, speech

movements, rate or pausing. Most of these variables investigated articulatory accuracy and speech

inconsistency/consistency. Findings revealed that children with CAS tended to perform poorer than

children with typical development on tasks that assessed percent phonemes or vowels correct (Murray,

McCabe, et al., 2015; Ziethe, Springer, Willmes, & Kröger, 2013) and standardized tests of articulation

and phonology (Peter, 2006). Murray et al. (2015) found that children with CAS could be differentiated

from those with phonological disorder and dysarthria when a combination of measures were used

including an articulation measure – percent phonemes correct. Children with CAS had a lower percentage

of lexical stress matches, a high occurrence of syllable segregation, a relatively higher percentage of

phonemes correct [than children with dysarthria] and low accuracy on the diadochokinetic task ‘peteke’.

Ziethe et al (2013) found that percentage vowels correct in complex, multisyllabic pseudo-words

differentiated children with CAS who performed more poorly than children with phonological disorder.

Similarly, Bradford and Dodd (1996) demonstrated CAS and inconsistent PD performed worse on

learning novel words than PD, SD and TD; however CAS performed more poorly than inconsistent PD,

PD, SD and TD on imitation tasks. Shriberg, Lohmeier, Strand, and Jakielski (2012) showed that children

with CAS demonstrated a higher percentage of within-class manner substitutions and responses

containing one or more additions compared to controls with typically developing speech and speech delay

(with or without comorbid developmental language disorder). Iuzzini-Seigel et al. (2017) found children

with CAS evidenced higher inconsistency on repeated production of monosyllabic words and the phrase

“buy Bobby a Puppy” compared to children with typically developing speech and those with speech

delay. Acoustic analyses revealed that relative to children with speech delay, children with CAS produced

fewer optimal voice onset times for /p/ targets, and reduced vowel space area when auditory feedback was

attenuated with noise masking (Iuzzini-Seigel et al., 2015). Children with CAS and Galactosemia also

21
[Type here]

demonstrated slower speaking and articulation rates and less stable F2 and vowel duration relative to

children with Galactosemia and speech delay (Shriberg et al., 2011). Finally, Shriberg et al. (2017a) found

that children with CAS demonstrated inappropriate pausing (i.e., pause marker) compared with

individuals with speech delay.

Maximum Performance.

Three studies (Murray, McCabe, et al., 2015; Thoonen et al., 1999; Thoonen, Maassen, Wit,

Gabreëls, & Schreuder, 1996) reported maximum performance tasks that discriminated between groups.

Compared to typically developing controls and children with spastic dysarthria, children with CAS

evidenced significantly shorter maximum fricative durations and slower trisyllabic repetition rate

(Thoonen, 1996). In addition, children with CAS evidenced more failed attempts at producing the

sequence ‘pataka’ and required more attempts to produce a correct trisyllable sequence compared to

children with typical development or spastic dysarthria. A follow-up study was conducted to cross-

validate Thoonen and colleagues’ findings and revealed that CAS can be sensitively and specifically

differentiated from TD, speech disorder of unknown origin, and dysarthria on the basis of maximum

repetition rate of ‘pataka’ in combination with maximum fricative duration of /f:/ (Thoonen et al., 1999).

Likewise, Murray and colleagues (2015) found that articulatory accuracy on repetitions of /pǝtǝkǝ/ were

sensitive and specific against expert diagnosis when differentiating CAS from phonological disorder and

dysarthria. This was in combination with syllable segregation, lexical stress matches, and percentage

phonemes correct on a polysyllabic picture-naming task.

Nonspeech.

Three studies (20%) (Bradford & Dodd, 1996; Bradford et al., 1997; Peter, 2006) reported

discriminative nonspeech variables. Bradford and colleagues (1996) used the Bruininks-Oseretsky Test of

Motor Proficiency (Bruininks, 1978) and demonstrated that (1) children with CAS and inconsistent PD

performed more poorly than children with TD, PD or speech delay on the upper limb strength and

dexterity subtest and (2) children with CAS performed poorer than inconsistent PD, PD, SD and TD on

the Visual-Motor Control subtest. Similarly, Peter (2006) used a clustering procedure to reveal that

22
[Type here]

children with CAS fell into two different clusters characterized by (1) low scores on the majority of

nonspeech timing tasks (clapped rhythm and repetitive tapping) and language tasks (as well as across

speech and phonological testing) and (2) a second cluster that had average language scores and

intermediate scores on timing tasks, but low scores on speech and phonological testing. Peter also found

that undetermined handedness was associated with CAS. Finally, Bradford and colleagues (1997) found

that tongue strength and endurance were lower in children with CAS compared to those with inconsistent

PD and deviant PD.

Prosody.

Five studies (33%) reported prosodic variables that discriminated between groups. Relative to

control groups of children with phonological disorder and dysarthria, children with CAS demonstrated

poorer performance on tasks that assessed percent of lexical stress matches and syllable segregation on

the Single-Word Test of Polysyllables (Murray et al., 2015). Within this limited sample, Murray et al.,

(2015) found prosodic accuracy accounted for 80% diagnostic accuracy alone but performed better in a

combination of measures to achieve 100% diagnostic accuracy. Relative to children with speech delay or

those with non-CAS speech disorder, those with CAS evidenced lower stress accuracy (Shriberg et al.,

2017a) or use of inappropriate stress on assessments including the DEMSS, Syllable Repetition Task, and

Prosody-Voice Screening Profile (Shriberg et al., 2012; Shriberg et al., 2011; Shriberg et al., 2017a,

2017b; Strand et al., 2013).

Auditory perception.

One study (7%) found that auditory speech perception ability as measured by a syllable

discrimination task differentiated groups (Zuk et al., 2018). Specifically, children with CAS and normal

language tended to have appropriate speech perception whereas those with CAS and comorbid

developmental language disorder evidenced poor speech perception per this task. Speech perception

therefore helped identify children with comorbid developmental language disorder but poor speech

perception was not considered a core feature of CAS.

Literacy.

23
[Type here]

There were no studies that met or nearly met the criteria which assessed literacy measures.

Discussion

The differential diagnosis of paediatric speech sound disorders (SSDs) is one of the ongoing

challenges in the field of speech-language pathology. Although there is consensus about CAS as a

neurological SSD of motor planning and/or programming, CAS has been associated with a wide variety

of symptoms that overlap with other types of SSD (ASHA, 2007). The primary aim of the present study

was to evaluate the existing literature to determine the discriminative features that might guide and

contribute to the development of a diagnostic protocol to differentiate CAS from other SSDs for future

clinical and research use.

To this end, this systematic review provides a comprehensive summary of the existing research

regarding differential features of CAS. Fifty-three studies were identified that examined speech-language

characteristics that sought to differentiate CAS from at least one other type of SSD. The studies cover a

wide range of speech characteristics that have been investigated to date as well as how they have been

operationalized in concrete index measures. Additionally, studies show large variation in how

methodology and results are reported. As such, studies were difficult to combine and compare, and a

meta-analysis of data was not possible.

Study Quality (aim 1)

None of the studies met the quality criteria for a Class I AAN rating and only three met the

criteria for a Class II AAN rating; consequently, the impact of findings is limited. An important factor

herein is consistency and transparency in reporting. Important methodological aspects were often absent

or not clearly reported. This applies to the full range of the studies’ methodology and includes

information regarding design (e.g., whether the study, including analysis, was planned retrospectively or

prospectively), sampling method (e.g., what method was used: random, consecutive, or convenience

sampling), sample size (e.g., whether the sample size was backed up by a power calculation), inclusion

criteria, test administration (e.g., who administered the assessments; were the assessors for the reference

24
[Type here]

and the index test blinded or not), testing procedure (e.g., what was the time between the administration

of the reference and the index test), data processing and analysis (e.g., who scored the tests; were the

assessors for the reference and the index test blinded or not), and statistics (e.g., whether statistical

assumptions were met or not).

Regarding inclusion criteria, diagnostic criteria, assessment procedures, and data analysis, the

reader is regularly referred to previous work for methodological details without even a brief description.

Although this is not uncommon, motivated by a desire for brevity or by editorial concerns regarding

copyright or plagiarism, extensive reporting is important. Although these issues might not affect the

impact in the research community, which usually has access to the referred sources, this is generally not

the case for clinicians. Furthermore, policy makers and insurance companies are guided by numeric

assessments of scientific evidence. For purposes of health care management and insurance coverage,

clarity and integrity/completeness of reporting is therefore crucial. We therefore strongly recommend that

researchers explicitly adopt existing guidelines such as PRISMA (Moher et al, 2009; McInnes et al,

2018), QUADAS-2 (Whiting et al, 2011), and STARD (Cohen et al., 2015) to ensure study design and

reporting meet these quality and reporting criteria and thus maximize impact.

Participant descriptions and confidence in diagnosis (aim 2)

The size of the CAS participant samples varied widely from 1-63 with a mean of 16, and half of

the studies featured a sample size below 11. Such small samples are problematic, especially combined

with the large age ranges involved. Across studies, the ages of children in the CAS groups ranged from 3

to 18 years and within studies most age ranges that spanned > 5 years. CAS is notorious for its

heterogeneity in symptomatology, and symptoms are subject to change during development and resultant

from treatment (Maassen, Nijland, & Terband, 2010; Maassen, 2015; Strand & McCauley, 2008). The

potential consequences of sampling errors are serious. CAS is slowly but certainly starting to become a

well-studied disorder (at least for the English language) and there is a need for larger sample sizes. This

issue is complicated further by the fact that multiple studies involve the same participants. The total

number of 845 participants with CAS and 1,802 participants with other SSDs included across studies does

25
[Type here]

not represent the same number of unique children. It is paramount that it is explicitly acknowledged if a

study utilizes an existing dataset or if it includes the same children as previous studies. Reporting on these

aspects is improving, but it also appears the practice of studying the same children is becoming more

frequent. Retrospectively exploiting datasets to their maximum and performing multiple studies on the

same participant pool constitute large risks that need to be recognized and not underestimated. The risks

include that we gain most of our understanding about a disorder from a small set of participants that may

not be representative of the larger population and that family-wise statistical errors can overestimate

significant differences and associations.

Like what has been noted above regarding the operationalization of the investigated differential

markers of CAS (index test), we note that there is no consistent diagnostic method (reference test).

Rather, each research group seems to rely on their own protocol with their own inclusion criteria,

diagnostic criteria, and assessment procedures. These differences make it difficult to verify diagnoses and

compare results between studies. It should also be noted that 26 of the 53 papers predated the publication

of the ASHA Technical Report and thus reflect a time of ongoing controversy about CAS and its specific

characteristics. A number of older, influential papers that shaped the three core-features of CAS presented

in the ASHA Technical Report (2007) are not included as promising papers in the current review, in part

due to having poor specification of CAS diagnosis (e.g., Barry, 1995; Lewis, Freebairn, Hansen, lyengar,

et al., 2004; Shriberg & et al., 1997b). Future systematic reviews which focus on studies performed well

after 2007 may identify more studies with diagnosis explicitly utilizing the ASHA characteristics.

Another source of variation in diagnostic protocols is the fact that the ASHA CAS Technical

Report (2007) did not report operational definitions or standardized procedures for measuring the three

diagnostic features. This requires researchers to use their own operational definitions and protocols (e.g.,

Iuzzini-Seigel & Murray, 2017; Terband et al., 2019). An important factor herein is crosslinguistic

differences. Certain features appear to have greater differential diagnostic capacity in some languages and

less so in others. For example, English features a very strong expression of lexical stress. Languages like

Dutch in which lexical stress is expressed less strongly, or French which has a less intricate system of

26
[Type here]

stress assignment compared to English, do not lend themselves to measuring the production accuracy of

word stress in the same way as English. There is a need for more uniform, explicit diagnostic criteria that

are more replicable, cross linguistically validated using specific tests or scores.

However, also with respect to highly language-neutral tasks such as maximum repetition rate

(MRR) or diadochokinesis (DDK) with non-meaningful stimuli (e.g., “pataka”), a variety of methods are

used across languages and research groups (but see Diepeveen, van Haaften, Terband, de Swart, &

Maassen, 2019; Rvachew, Hodge, & Ohberg, 2005 for standardized protocols). What the studies do have

in common though, is that the investigated measures address aspects of an underlying motor impairment,

such as temporal control of the articulators, whether that be lexical stress contrasts in English (Murray et

al, 2015) or for syllabic organization and phonetic context in Dutch (Nijland, Maassen, & Der Meulen,

2003). This highlights that while it may manifest differently across languages, the underlying motor

deficit across people and languages is considered the same.

Discriminative diagnostic variables (features/markers) (aim 3)

Several features were found to be discriminative. A complicating factor in this respect is that

relevant clinical comparison groups varied strongly across studies and comprised six different diagnostic

classifications (that in turn varied in definition between studies), with some comparing to typically

developing children additionally. Furthermore, over half of the studies only reported discriminative

measures on the group level and did not report sensitivity and specificity of these measures in the

identification of individuals with CAS; not due to methodological neglect but simply because these

studies had different aims than validating a marker for differential diagnosis. This included three out of

seven studies that met the quality criteria for promising studies and five out of eight studies that were

close to meeting all our criteria. For example, percentage vowels correct in pseudo-words was found to

differentiate between groups with CAS, PD, and TD for German children, but the index measures were

not analyzed as a differential diagnostic marker (Ziethe et al., 2013). In the following section, we will

discuss the diagnostic measures that have been found to reliably differentiate individuals with CAS per

comparison group, i.e. of which sensitivity and specificity figures are reported. It must be stated,

27
[Type here]

however, that all sensitivity and specificity measures reported across all studies did not include

confidence intervals, meaning the sensitivity and specificity estimates reported could in reality be lower

if random error is considered (Deeks & Altman, 1999; Dollaghan, 2007). Future diagnostic studies should

ensure reporting of confidence intervals with diagnostic accuracy data.

The strongest evidenced differentiation is between CAS and developmental dysarthria, in which

maximum performance tasks play a central role. Thoonen et al. (1999) showed in a replication of

Thoonen et al (1996), CAS in Dutch can be reliably differentiated from spastic dysarthria based on

maximum repetition rate of ‘pataka’ in combination with maximum fricative duration (100%/91%

sensitivity/specificity). For Australian English, Murray et al. (2015) showed reliable differentiation

between CAS and dysarthria based on accuracy on repetition of /pǝtǝkǝ/ in combination with percent

phonemes correct, lexical stress matches and the presence of syllable segregation on the polysyllable test

(100%/100% sensitivity/specificity). This study found that the same set of measures also served to

reliably differentiate between CAS and speech delay/phonological disorder/non-CAS speech disorder. For

American English, the DEMSS motor speech assessment combining inconsistency, segmental accuracy

and prosody measures showed high specificity (97%) but lower sensitivity (65%) (Strand et al., 2013).

Other studies in American English have mainly focused on single diagnostic markers, of which many

measures address prosody. Word level measures show low to reasonable accuracy, with sensitivity and

specificity of stress accuracy and use of inappropriate stress both varying from around 50 to 75 %

(Shriberg et al., 2012). The phrase/sentence level prosody measure--inappropriate pausing--showed

sensitivity and specificity in differentiating CAS from SD, 87% and 99% respectively (Shriberg et al.,

2017a). Other measures that differentiated between CAS and speech delay/phonological disorder/non-

CAS speech disorder in American English-speaking children all showed lower diagnostic accuracy.

Inconsistency on five repeated productions of the phrase “buy Bobby a Puppy” had 70% sensitivity and

80% specificity (Iuzzini-Seigel et al., 2017). Single segmental error measures such as percentage of

within-class manner substitutions and percentage of additions showed sensitivity and specificity levels

varying 53-74% and 52-75% respectively (Shriberg et al., 2012). Among children with Galactosemia,

28
[Type here]

speaking and articulation rate and stability of F2 and vowel duration showed reasonable to high

sensitivity and specificity in differentiating children with CAS from children with Galactosemia without

CAS and children with speech delay (71-88% / 84-100% sensitivity/specificity; Shriberg et al., 2011).

Measures that differentiated CAS and developmental language disorder with reported sensitivity

and specificity were perceptual. Percentage of within-class manner substitutions and percentage of

additions showed low to reasonable accuracy (53-74% and 52-75% sensitivity and specificity

respectively) in differentiating between CAS and language disorder, both with and without concomitant

SSD (Shriberg et al., 2012). With respect to comparison groups of children who stutter, had literacy

difficulties or autism spectrum disorder, none of the studies met or were close to meeting the quality

criteria for promising studies.

In summary, the results of this review indicate that to date there are no individual markers that are

sufficiently sensitive and specific ( > 90%; Shriberg et al., 1997; Thoonen et al., 1999) in differentiating

between CAS and other SSDs. Combinations of measures are more efficacious than single diagnostic

markers, with sensitivity and specificity estimates reaching 100%. It can be concluded that reliable

differential diagnosis requires a combination of measures (e.g., Murray et al., 2015; Thoonen et al., 1999)

to assess across a range of SSDs. The set of diagnostic measures that would be best for use in clinical

practice is further discussed below.

Implications for clinical practice

Children with suspected CAS continue to present to clinics for differential diagnosis and for

evidence-based treatment planning and management. The primary question addressed in this review is

‘what tools do clinicians have that establish whether a child has CAS?’ Clinicians are recommended to

complete an initial, comprehensive test battery, including a hearing test, oral-musculature assessment,

single-word production (including polysyllable words) and connected speech sampling (see Terband et

al., 2019). This assessment can serve to generate a hypothesis as to which speech sound disorder(s) a

child may have.

29
[Type here]

To subsequently test such hypotheses, clinicians need methods and measures that can reliably and

validly determine the presence or absence of CAS. This review evaluated several methods that have been

used in the literature, some of which hold promise in this regard in the interim. Further development and

validation are needed to arrive at definitive methods and measures. Until then, a clinician’s best course of

action would be to identify papers in which the children share characteristics with the client for whom a

differential diagnosis is sought, and to use the measures faithfully to help determine if the child has CAS.

The promising measures identified in this review can provide rigorous methods to this end (see Tables 2

and 3). For example, for a four year old child suspected of having CAS or a phonological disorder, a

clinician could use the DEMSS (i.e., to provide inconsistency, sequencing and prosody information,

Strand & McCauley, 2019; Strand et al., 2013); Iuzzini-Seigel et al.’s (2017) inconsistency measure

supplemented with either the Murray et al (2015) protocol or a combination of measures from the

Robbins and Klee oral musculature assessment (Robbins & Klee, 1987), and polysyllable test (Gozzard,

Baker, & McCabe, 2004) using the formula in the paper or Thoonen et al.’s (1996, 1999) maximum

performance tasks using a tutorial to assist (Diepeveen et al., 2019; Rvachew et al., 2005). If the

differential diagnostic possibilities include CAS or a speech delay, one could use Shriberg et al (2017)’s

pause marker using the protocol from Tilkens et al. (2017) and/or Iuzzini-Seigel et al (2017)’s

inconsistency measure. To differentiate between CAS and dysarthria, there are robust methods from few

studies available. A thorough oral musculature assessment (OMA) investigating muscle tone, strength and

range of movement would be required (Hodge & Hancock, 1994; Murray, McCabe, et al., 2015). If the

clinician suspects spastic dysarthria, Thoonen et al (1999)’s maximum performance tasks will help

provide a dysarthria score (Diepeveen et al., 2019; Rvachew et al., 2005). Murray et al (2015)’s

combination of measures will also help identify dysarthria versus CAS. Although further research is

clearly needed to develop, refine, and validate diagnostic measures, there are tools available to the

informed clinician that can help with differential diagnosis in the interim, especially when used in

combination, to build greater confidence in diagnosis.

Limitations and implications for future research

30
[Type here]

Despite attempting to find papers with participants speaking any language and being written in

any language, the results of this review are very English-dominant. There were five articles that report

children who spoke Dutch and four other articles in which children spoke Arabic, French, German, and

Portuguese; but most of the studies (89%) assessed only English speakers. The results thus predominantly

relate to English speakers.

An adjacent limitation is our use of the English-dominant ASHA (2007) features in determining

confidence in the original CAS diagnosis made. As mentioned above, internationally there appears to be a

consensus that CAS is a pediatric disorder of speech motor planning and/or programming. However, the

way in which the underlying neurological deficit manifests itself symptomatically is likely to differ across

languages, driven by phonetic and phonological characteristics (such as sound inventory, phonotactics,

rules governing phonological alternation, prosody). As a result, the ASHA (2007) features may not

directly apply to studies completed in other languages, which could have affected some studies’ ability to

meet our current quality criteria and be included as promising papers. CAS is starting to become a well-

studied disorder for English-speaking participants and researchers. On the one hand, cross-linguistic

differences limit the generalizability of findings and raise the need to replicate studies and validate

measures in other languages. On the other hand, there is great potential in exploiting cross-linguistic

differences to tease out the underlying mechanisms and specify the underlying deficit, its symptoms and

its preconditions.

Furthermore, none of the findings based on the existing literature have been replicated with a

new, different sample of participants apart from the prospective validation study by Thoonen et al. (1999).

Replication studies featuring larger sample sizes are warranted to determine the reproducibility of earlier

studies’ results. We note that replication studies are not always appreciated or valued by employers and

funding institutions, nor by journals and researcher colleagues. It is often difficult to get a replication

study funded and published, and it is often frowned upon from a researcher’s career perspective.

However, replication is an important part of the scientific endeavor, as fundamental from a

methodological point of view, it is the ultimate and only real test of the validity of a study’s results.

31
[Type here]

Furthermore, it is important to establish the methods and circumstances required for reproducibility

beyond exact replications in order to determine the scope of validity and establish the requirements and

conditions for implementation in clinical practice. We thus call upon employers and funding institutions

to encourage replication studies, and upon journals to be more open to publishing them, including well-

designed and -executed studies with null results.

In this respect, researchers are recommended to register trials and (pre-)publish study protocols.

This custom would minimize possible publication bias and, through unique trial participant registration

numbers, allow identification of participants and data sets across studies. It would also allow for greater

replication, learning and potentially, for collaboration across research groups as publishing of protocols

allows for early feedback and discussion, rather than waiting for the end of trials and providing feedback

through the peer-review process. A data-driven approach featuring theoretically substantiated cluster

analyses allows us to circumvent the issue of reliable initial group assignment and minimize diagnostic

circularity.

Summarizing these points, recommendations for future assessment and diagnostic studies include

careful and prospectively planned studies using AAN and STARD reporting guidelines (Cohen et al.,

2016; Neurology), 2011) and QUADAS-2 quality criteria (Whiting et al., 2011) It is helpful to register

trials on PROSPERO to share protocols and collaborate prior to data collection.

Conclusion

This systematic review of 53 papers sought to discriminate CAS from other SSDs. Results

indicate that no study in the existing literature meet the highest level of study quality (AAN level 1). Most

studies were retrospective, used case-control designs, and lacked reporting of several design aspects. In

terms of diagnostic methods, there was no consistent reference test used across studies. Seven studies

demonstrated promising index tests (examined measures or tools) for future clinical and research use.

Another eight almost met the quality criteria, also demonstrating some promise. Future studies should

replicate existing methods and use larger sample sizes. In addition, research needs to address the

presentation and diagnosis of CAS in languages other than English (e.g., Wong, Lee, & Tong, 2020).

32
[Type here]

Research designs ideally should include cohort studies, prospectively designed according to STARD and

QUADAS-2 guidelines and have data-driven analyses for the greatest rigor. Clinicians are recommended

to complete routine speech assessment and then use promising tools in this study to refine if a child has

CAS compared to or comorbid with another SSD diagnosis.

33
[Type here]

Acknowledgments

We acknowledge support from Lauren Rusedski, the research assistant who supported finding the articles

in the review; the Academy of Neurologic Communication Disorders and Sciences (ANCDS) and the

Apraxia of Speech Writing Committee (Acquired AOS members) in providing the manual and excel

spreadsheet templates we adapted for this CAS review.

References

AAN (American Academy of Neurology). (2011). Clinical Practice Guideline Process Manual. St. Paul,

MN: The American Academy of Neurology.

ASHA (American Speech-Language-Hearing Association). (2007). Childhood Apraxia of Speech:

Technical Report. Retrieved from http://www.asha.org/policy/TR2007-00278/

ASHA (American Speech-Language-Hearing Association). (2016). 2016 Schools Survey report: SLP

caseload characteristics. Retrieved from https://www.asha.org/uploadedFiles/2016-Schools-

Survey-SLP-Caseload-Characteristics.pdf

Aziz, A. A., Shohdi, S., Osman, D. M., & Habib, E. I. (2010). Childhood apraxia of speech and multiple

phonological disorders in Cairo-Egyptian Arabic speaking children: language, speech, and oro-

motor differences. International Journal of Pediatric Otorhinolaryngology, 74(6), 578-585.

https://doi.org//10.1016/j.ijporl.2010.02.003

Barry, R. M. (1995). The Relationship between Dysarthria and Verbal Dyspraxia in Children: A

Comparative Study Using Profiling and Instrumental Analyses. Clinical Linguistics & Phonetics,

9(4), 277-305. https://doi.org/10.3109/02699209508985338

Betz, S. K., & Stoel-Gammon, C. (2005). Measuring articulatory error consistency in children with

developmental apraxia of speech. Clinical Linguistics & Phonetics, 19(1), 53-66. https://doi.org/

10.1080/02699200512331325791

Bowen, C. (2015). Children's Speech Sound Disorders: Second Edition. Sydney, Australia: John Wiley &

Sons, Ltd.Bradford, A., & Dodd, B. (1996). Do All Speech-Disordered Children Have Motor

34
[Type here]

Deficits? Clinical Linguistics & Phonetics, 10(2), 77-101.

https://doi.org/10.3109/02699209608985164

Bradford, A., Murdoch, B., Thompson, E., & Stokes, P. (1997). Lip and Tongue Function in Children

with Developmental Speech Disorders: A Preliminary Investigation. Clinical Linguistics &

Phonetics, 11(5), 363-387. https://doi.org/10.1080/02699209708985201

Broomfield, J., & Dodd, B. (2004). The Nature of Referred Subtypes of Primary Speech Disability. Child

Language Teaching and Therapy, 20(2), 135-151. https://doi.org/10.1191/0265659004ct267oa

Carrigg, B., Parry, L., Baker, E., Shriberg, L., & Ballard, K. (2016). Cognitive, Linguistic, and Motor

Abilities in a Multigenerational Family with Childhood Apraxia of Speech. Archives of Clinical

Neuropsychology, 31. https://doi.org/doi:10.1093/arclin/acw077

Cohen, J. F., Korevaar, D. A., Altman, D. G., Bruns, D. E., Gatsonis, C. A., Hooft, L., . . . Bossuyt, P. M.

M. (2016). STARD 2015 guidelines for reporting diagnostic accuracy studies: explanation and

elaboration. BMJ Open, 6(11), e012799. https://doi.org/10.1136/bmjopen-2016-012799

Deeks, J. J., & Altman, D. G. (1999). Sensitivity and specificity and their confidence intervals cannot

exceed 100%. British Medical Journal, 318(7177), 193.

https://doi.org/10.1136/bmj.318.7177.193b

Diepeveen, S., van Haaften, L., Terband, H., de Swart, B., & Maassen, B. (2019). A Standardized

Protocol for Maximum Repetition Rate Assessment in Children. Folia Phoniatrica et

Logopaedica, 71(5-6), 238-250. https://doi.org/10.1159/000500305

Dodd, B. (2014). Differential Diagnosis of Pediatric Speech Sound Disorder. Current Developmental

Disorders Reports, 1(3), 189-196. https://doi.org/10.1007/s40474-014-0017-3

Dollaghan, C. A. (2007). The Handbook for Evidence-based Practice in Communication Disorders.

Baltimore: Paul H. Brookes Publishing Co.

Eadie, P., Morgan, A., Ukoumunne, O. C., Ttofari Eecen, K., Wake, M., & Reilly, S. (2015). Speech

sound disorder at 4 years: prevalence, comorbidities, and predictors in a community cohort of

35
[Type here]

children. Developmental Medicine and Child Neurology, 57(6), 578-584.

https://doi.org/10.1111/dmcn.12635

Goldman, R., & Fristoe, M. (2015). Goldman-Fristoe Test of Articulation (3rd ed.): Pearson Inc.

Gozzard, H., Baker, E., & McCabe, P. (2004). Single Word Test of Polysyllables: Unpublished work.

Gubiani, M. B., Pagliarin, K. C., & Keske-Soares, M. (2015). Tools for the assessment of childhood

apraxia of speech. CODAS, 27(6), 610-615. https://doi.org//10.1590/2317-1782/20152014152

Hallgren, K. A. (2012). Computing Inter-Rater Reliability for Observational Data: An Overview and

Tutorial. Tutorials in quantitative methods for psychology, 8(1), 23-34.

https://doi.org/10.20982/tqmp.08.1.p023

Hayden, D. (1994). Differential diagnosis of motor speech dysfunction in children. Clinics in

communication disorders, 4(2), 119-141.

Hayden, D., & Square-Storer, P. (1999). VMPAC: verbal motor production assessment for children. San

Antonio, TX: Psychological Corporation.

Hodge, M. M., & Hancock, H. R. (1994). Assessment of children with developmental apraxia of speech: a

procedure. Clinics in communication disorders, 4(2), 102-118.

Iuzzini-Seigel, J. (2019). Motor Performance in Children With Childhood Apraxia of Speech and Speech

Sound Disorders. Journal of Speech, Language, and Hearing Research, 62(9), 3220-3233.

https://doi.org/10.1044/2019_JSLHR-S-18-0380

Iuzzini-Seigel, J., Hogan Tiffany, P., & Green Jordan, R. (2017). Speech Inconsistency in Children With

Childhood Apraxia of Speech, Language Impairment, and Speech Delay: Depends on the Stimuli.

Journal of Speech, Language, and Hearing Research, 60(5), 1194-1210.

https://doi.org/10.1044/2016_JSLHR-S-15-0184

Iuzzini-Seigel, J., Hogan, T. P., Guarino, A. J., & Green, J. R. (2015). Reliance on auditory feedback in

children with childhood apraxia of speech. Journal of Communication Disorders, 54, 32-42.

https://doi.org /10.1016/j.jcomdis.2015.01.002

36
[Type here]

Iuzzini-Seigel, J., & Murray, E. (2017). Speech assessment in children with childhood apraxia of speech.

Perspectives of the ASHA Special Interest Groups, 2(2), 47-60.

https://doi.org/10.1044/persp2.SIG2.47

Kent, R. D. (1996). Hearing and Believing: Some Limits to the Auditory-Perceptual Assessment of

Speech and Voice Disorders. American Journal of Speech-Language Pathology, 5(3), 7-23.

https://doi.org/

Kratochwill, T. R., Hitchcock, J., Horner, R. H., Levin, J. R., Odom, S. L., Rindskopf, D. M., & Shadish,

W. R. (2010). Single-case designs technical documentation. Retrieved from What Works

Clearinghouse website: http://ies.ed.gov/ncee/wwc/pdf/wwc_scd.pdf

Storkel H. L. (2019). Using Developmental Norms for Speech Sounds as a Means of Determining

Treatment Eligibility in Schools. Perspectives of the ASHA Special Interest Groups, 4(1), 67-75.

https://doi.org/10.1044/2018_pers-SIG1-2018-0014

Lewis, B. A., Freebairn, L., Tag, J., Benchek, P., Morris, N. J., Iyengar, S. K., . . . Stein, C. M. (2018).

Heritability and longitudinal outcomes of spelling skills in individuals with histories of early

speech and language disorders. Learning and Individual Differences, 65, 1-11.

https://doi.org/10.1016/j.lindif.2018.05.001

Lewis, B. A., Freebairn, L. A., Hansen, A., Taylor, H., Iyengar, S. K., & Shriberg, L. D. (2004). Family

pedigrees of children with suspected childhood apraxia of speech. Journal of Communication

Disorders, 37(2), 157-175. https://doi.org//10.1016/j.jcomdis.2003.08.003

Lewis, B. A., Freebairn, L. A., Hansen, A. J., lyengar, S. K., & Taylor, H. (2004). School-age follow-up

of children with childhood apraxia of speech. Language, Speech, and Hearing Services in

Schools, 35(2), 122-140. https://doi.org//10.1044/0161-1461%282004/014%29

Liégeois, F. J., & Morgan, A. T. (2012). Neural bases of childhood speech disorders: Lateralization and

plasticity for speech functions during development. Neuroscience & Biobehavioral Reviews,

36(1), 439-458. doi: https://doi.org//10.1016/j.neubiorev.2011.07.011

37
[Type here]

Maas, E., & Mailend, M. L. (2017). Fricative contrast and coarticulation in children with and without

speech sound disorders. American Journal of Speech-Language Pathology, 26(2), 649-663.

https://doi.org/10.1044/2017_AJSLP-16-0110

Maassen, B., Nijland, L., & Terband, H. (2010). Developmental models of Childhood Apraxia of Speech

In B. M. P. V. Lieshout (Ed.), Speech motor control: New developments in basic and applied

research (pp. 243-258). Oxford, UK: Oxford University Press.

Maassen, B., Nijland, L., & Terband, H. (2012). Developmental models of childhood apraxia of speech.

In Speech Motor Control: New Developments in Basic and Applied Research.

Maassen, B. A. M. (2015). Developmental models of childhood apraxia of speech. In Routledge

handbook of communication disorders (pp. 124-133). New York, NY: Routledge/Taylor &

Francis Group; US.

Macrae, T. (2016). Comprehensive Assessment of Speech Sound Production in Preschool Children.

Perspectives of the ASHA Special Interest Groups, 1(1), 39-

56.https://doi.org/10.1044/persp1.SIG1.39

McCabe, P., Rosenthal, J. B., & McLeod, S. (1998). Features of developmental dyspraxia in the general

speech-impaired population? Clinical Linguistics & Phonetics, 12(2), 105-126. https://doi.org/

10.3109/02699209808985216

McCauley, R. J., & Strand, E. A. (2008). A review of standardized tests of nonverbal oral and speech

motor performance in children. American Journal of Speech-Language Pathology, 17(1), 81-91.

https://doi.org/10.1044/1058-0360(2008/007)

McHugh, M. L. (2012). Interrater reliability: the kappa statistic. Biochemica Medica (Zagreb), 22(3), 276-

282.

McInnes, M. D. F., Moher, D., Thombs, B. D., McGrath, T. A., Bossuyt, P. M., Clifford, T., . . . Willis, B.

H. (2018). Preferred Reporting Items for a Systematic Review and Meta-analysis of Diagnostic

Test Accuracy Studies: The PRISMA-DTA Statement. Jama, 319(4), 388-396.

https://doi.org/10.1001/jama.2017.19163

38
[Type here]

McLeod, S., & Baker, E. (2014). Speech-language pathologists' practices regarding assessment, analysis,

target selection, intervention, and service delivery for children with speech sound disorders. Clin

Linguist Phon, 28(7-8), 508-531. https://doi.org/10.3109/02699206.2014.926994

Mei, C., Fedorenko, E., Amor, D. J., Boys, A., Hoeflin, C., Carew, P., . . . Morgan, A. T. (2018). Deep

phenotyping of speech and language skills in individuals with 16p11.2 deletion. European

Journal of Human Genetics, 26(5), 676-686. https://doi.org/10.1038/s41431-018-0102-x

Merlin, T., Weston, A., & Tooher, R. (2009). Extending an evidence hierarchy to include topics other

than treatment: revising the Australian 'levels of evidence'. BMC Medical Research Methodology,

9(1), 34. https://doi.org/10.1186/1471-2288-9-34

Moher, D., Liberati, A., Tetzlaff, J., & Altman, D. G. (2009). Preferred reporting items for systematic

reviews and meta-analyses: The PRISMA statement. BMJ, 339.

https://doi.org/10.1136/bmj.b2535

Morley, M., Court, D., & Miller, H. (1954). Developmental dysarthria. British Medical Journal, 1(4852),

8-10. https://doi.org/10.1136/bmj.1.4852.8

Murray, E., & Iuzzini-Seigel, J. (2017). Efficacious Treatment of Children With Childhood Apraxia of

Speech According to the International Classification of Functioning, Disability and Health.

Perspectives of the ASHA Special Interest Groups 2: Neurogenic Communication Disorders,

2(2). https://doi.org/10.1044/persp2.SIG2.61

Murray, E., Mc Cabe, P., Heard, R., & Ballard, K. J. (2015). Differential diagnosis of children with

suspected childhood apraxia of speech. Journal of Speech, Language, and Hearing Research,

58(1), 43-60. https://doi.org/10.1044/2014_JSLHR-S-12-0358

Murray, E., McCabe, P., & Ballard, K. J. (2015). A randomized controlled trial for children with

childhood apraxia of speech comparing rapid syllable transition treatment and the nuffield

dyspraxia programme–third edition. Journal of Speech, Language, and Hearing Research, 58(3),

669-686. https://doi.org/10.1044/2015_JSLHR-S-13-0179

39
[Type here]

Murray, E., Thomas, D., & McKechnie, J. (2019). Comorbid morphological disorder apparent in some

children aged 4-5 years with childhood apraxia of speech: findings from standardised testing.

Clinical Linguistics & Phonetics, 33(1-2), 42-59.

https://doi.org/10.1080/02699206.2018.1513565

Nijland, L. (2009). Speech perception in children with speech output disorders. Clinical Linguistics &

Phonetics, 23(3), 222-239. https://doi.org/10.1080/02699200802399947

Nijland, L., Maassen, B., & Der Meulen, S. v. (2003). Evidence of motor programming deficits in

children diagnosed with DAS. Journal of Speech, Language, and Hearing Research, 46(2), 437-

450. https://doi.org/10.1044/1092-4388(2003/036)

OCEBM Levels of Evidence Working Group*. The Oxford Levels of Evidence 2. Oxford Centre for

Evidence-Based Medicine. https://www.cebm.net/index.aspx?o=5653

PDF™ Tron Systems (2014). Endnote Software. X8.2.

Peter, B. (2006). Multivariate characteristics and data-based disorder classification in children with

speech disorders of unknown origin. (PhD), University of Washington, Unpublished dissertation.

Retrieved from Linguistics and Language Behavior Abstracts (LLBA) database.

Potter, N. L., Nievergelt, Y., & Shriberg, L. D. (2013). Motor and speech disorders in classic

galactosemia. JIMD Reports, 11, 31-41. https://doi.org//10.1007/8904_2013_219

PROSPERO. (2019). Retrieved from https://www.crd.york.ac.uk/prospero/

Robbins, J., & Klee, T. (1987). Clinical Assessment of Oropharyngeal Motor Development in Young

Children. Journal of Speech and Hearing Disorders, 52, 271-277.

https://doi.org/10.1044/jshd.5203.271

Rupela, V., Velleman, S. L., & Andrianopoulos, M. V. (2016). Motor speech skills in children with Down

syndrome: A descriptive study. International Journal of Speech Language Pathology, 18(5), 483.

https://doi.org/10.3109/17549507.2015.1112836

40
[Type here]

Rusiewicz, H. L., Maize, K., & Ptakowski, T. (2018). Parental experiences and perceptions related to

childhood apraxia of speech: Focus on functional implications. Int J Speech Lang Pathol, 20(5),

569-580. https://doi.org/10.1080/17549507.2017.1359333

Rvachew, S., Hodge, M., & Ohberg, A. (2005). Obtaining and interpreting maximum performance tasks

from children: a tutorial. Journal of Speech-Language Pathology & Audiology, 29(4), 146-157.

Sayahi, F., & Jalaie, S. (2016). Diagnosis of Childhood Apraxia of Speech: A Systematic Review.

Journal of Diagnostics, 3(1), 21-26. https://doi.org/10.18488/journal.98/2016.3.1/98.1.21.26

Shriberg, L. D., Campbell, T. F., Karlsson, H. B., Brown, R. L., McSweeny, J. L., & Nadler, C. J. (2003).

A diagnostic marker for childhood apraxia of speech: the lexical stress ratio. Clinical Linguistics

& Phonetics, 17(7), 549-574. https://doi.org/10.1080/0269920031000138123

Shriberg, L. D., Aram, D.M., Kwiatkowski, J. (1997a). Developmental Apraxia of Speech: I. Descriptive

and Theoretical Perspectives. Journal of Speech, Language, and Hearing Research, 40(2), 273-

285. https://doi.org/10.1044/jslhr.4002.273

Shriberg, L. D., Aram, D.M., Kwiatkowski, J. (1997b). Developmental Apraxia of Speech: II. Toward a

Diagnostic Marker. Journal of Speech, Language, and Hearing Research, 40(2), 286-312.

https://doi.org/10.1080/0269920031000138123

Shriberg, L. D., Lohmeier, H. L., Strand, E. A., & Jakielski, K. J. (2012). Encoding, memory, and

transcoding deficits in Childhood Apraxia of Speech. Clinical Linguistics & Phonetics, 26(5),

445-482. https://doi.org/10.1044/jslhr.4002.286

Shriberg, L. D., Potter, N. L., & Strand, E. A. (2011). Prevalence and Phenotype of Childhood Apraxia of

Speech in Youth With Galactosemia. Journal of Speech, Language & Hearing Research, 54(2),

487-519. https://doi.org/1092-4388(2010/10-0068)

Shriberg, L. D., Strand, E. A., Fourakis, M., Jakielski, K. J., Hall, S. D., Karlsson, H. B., . . . Wilson, D.

L. (2017a). A Diagnostic Marker to Discriminate Childhood Apraxia of Speech From Speech

Delay: II. Validity Studies of the Pause Marker. Journal of Speech, Language & Hearing

Research, 60(4), S1118-s1134. https://doi.org/10.1044/2016_jslhr-s-15-0297

41
[Type here]

Shriberg, L. D., Strand, E. A., Fourakis, M., Jakielski, K. J., Hall, S. D., Karlsson, H. B., . . . Wilson, D.

L. (2017b). A Diagnostic Marker to Discriminate Childhood Apraxia of Speech From Speech

Delay: III. Theoretical Coherence of the Pause Marker with Speech Processing Deficits in

Childhood Apraxia of Speech. Journal of Speech, Language & Hearing Research, 60(4), S1135-

s1152. https://doi.org/10.1044/2016_jslhr-s-15-0298

Smith, B., Marquardt, T. P., Cannito, M. P., & Davis, B. L. (1994). Vowel Variability in Developmental

Apraxia of Speech. In J. A. Till, K. M. Yorkston, & D. R. Beukelman (Eds.), Motor Speech

Disorders: Advances in Assessment and Treatment. Baltimore, MD: Brookes Publishing.

IEPMCS (International Expert Panel on Multilingual Children's Speech) (2012). Multilingual children

with speech sound disorders: Position paper. Retrieved from

http://www.csu.edu.au/research/multilingual-speech/position-paper

Strand, E. A., & McCauley, R. J. (2008). Differential diagnosis of severe speech impairment in young

children. ASHA Leader, 13(10), 10-13. https://doi.org/10.1044/leader.FTR1.13102008.10

Strand, E. A., & McCauley, R. J. (2019). Dynamic Evaluation of Motor Speech Skill (DEMSS) Manual.

Baltimore, MD: Brookes Publishing.

Strand, E. A., McCauley, R. J., Weigand, S. D., Stoeckel, R. E., & Baas, B. S. (2013). A Motor Speech

Assessment for Children With Severe Speech Disorders: Reliability and Validity Evidence.

Journal of Speech, Language & Hearing Research, 56(2), 505-520. https://doi.org/1092-

4388(2012/12-0094)

Terband, H., Maassen, B., Guenther, F. H., & Brumberg, J. (2009). Computational neural modeling of

speech motor control in childhood apraxia of speech (CAS). Journal of Speech, Language &

Hearing Research, 52(6), 1595-1609. https://doi.org/10.1044/1092-4388(2009/07-0283)

Terband, H., Maassen, B., & Maas, E. (2019). A psycholinguistic framework for diagnosis and treatment

planning of developmental speech disorders. Folia Phoniatrica et Logopaedica, 71(5-6), 216-

227. https://doi.org/10.1159/000499426

42
[Type here]

Terband, H., Maassen, B., van Lieshout, P., & Nijland, L. (2011). Stability and composition of functional

synergies for speech movements in children with developmental speech disorders. Journal of

Communication Disorders, 44(1), 59-74. https://doi.org/10.1016/j.jcomdis.2010.07.003

Terband, H., Namasivayam, A., van Brenk, F., Diepeveen, S., Mailend, M.-L., Maas, E., . . . Maassen, B.

(2019). Assessment of Childhood Apraxia of Speech: a review/tutorial of objective measurement

techniques. Journal of Speech, Language and Hearing Research, 62(8S), 2999–3032.

https://doi.org/10.1044/2019_JSLHR-S-CSMC7-19-0214

Terband, H., Zaalen, Y. V., & Maassen, B. (2012). Lateral jaw stability in adults, children, and children

with developmental speech disorders. Journal of Medical Speech-Language Pathology, 20(4),

112-118.

Thoonen, G., Maassen, B., Gabreëls, F., & Schreuder, R. (1999). Validity of maximum performance tasks

to diagnose motor speech disorders in children. Clinical Linguistics and Phonetics, 13(1), 1-23.

https://doi.org/10.1080/026992099299211

Thoonen, G., Maassen, B., Wit, J., Gabreëls, F., & Schreuder, R. (1996). The integrated use of maximum

performance tasks in differential diagnostic evaluations among children with motor speech

disorders. Clinical Linguistics and Phonetics, 10(4), 311-336.

https://doi.org/10.3109/02699209608985178

Tilkens, C. M., Karlsson, H. B., Fourakis, M., Hall, S. D., Mabie, H. L., McSweeny, J. L., . . . Shriberg,

L. D. (2017) A Diagnostic Marker to Discriminate Childhood Apraxia of Speech (CAS) from

Speech Delay (SD): The Pause Marker. Technical Report No. 22.

Vick, J. C., Campbell, T. F., Shriberg, L. D., Green, J. R., Truemper, K., Rusiewicz, H. L., & Moore, C.

A. (2014). Data-Driven Subclassification of Speech Sound Disorders in Preschool Children.

Journal of Speech, Language & Hearing Research, 57(6), 2033-2050. https://doi.org

10.1044/2014_JSLHR-S-12-0193

43
[Type here]

Wambaugh, J. L., Duffy, J. R., McNeil, M. R., Robin, D. A., & Rogers, M. A. (2006). Treatment

guidelines for acquired apraxia of speech: A synthesis and evaluation of the evidence. Journal of

Medical Speech-Language Pathology, 14(2), xv-xxxiii.

Whiting, P. F., Rutjes, A. W., Westwood, M. E., Mallett, S., Deeks, J. J., Reitsma, J. B., . . . Bossuyt, P.

M. (2011). QUADAS-2: a revised tool for the quality assessment of diagnostic accuracy studies.

Annuals of Internal Medicine, 155(8), 529-536. https://doi.org 10.7326/0003-4819-155-8-

201110180-00009

Williams, R., Ingham, R. J., & Rosenthal, J. (1981). A further analysis for developmental apraxia of

speech in children with defective articulation. Journal of Speech & Hearing Research, 24(4),

496-505. https://doi.org http://dx.doi.org/10.1044/jshr.2404.496

Wong, E. C. H., Lee, K. Y. S., & Tong, M. C. F. (2020). The Applicability of the Clinical Features of

English Childhood Apraxia of Speech to Cantonese: A Modified Delphi Survey. American

Journal of Speech-Language Pathology. https://doi.org 10.1044/2019_AJSLP-19-00118

Yoss, K. A., & Darley, F. L. (1974). Developmental apraxia of speech in children with defective

articulation. Journal of Speech & Hearing Research, 17(3), 399-416. https://doi.org

/10.1044/jshr.1703.399

Ziethe, A., Springer, L., Willmes, K., & Kröger, B. J. (2013). Study of core features of children with

childhood apraxia of speech aged between 4 and 7 years. Sprache Stimme Gehor, 37(4), 210-214.

https://doi.org 10.1055/s-0032-1323786

Zuk, J., Iuzzini-Seigel, J., Cabbage, K., Green, J. R., & Hogan, T. P. (2018). Poor speech perception is not

a core deficit of childhood apraxia of speech: Preliminary findings. Journal of Speech, Language,

and Hearing Research, 61(3), 583-592. https://doi.org 10.1044/2017_JSLHR-S-16-0106

44
[Type here]

Figures and Figure Legends

45
[Type here]

Figure 2

Quality criteria for analysis

46
 [Type here]

Tables

Table 1.

American Academy of Neurology Classes for Rating Diagnostic Articles (AAN [American Academy of
Neurology], 2011).

Class Description
Class I A cohort [cross-sectional] study with prospective data collection of a broad spectrum of
persons with the suspected condition, using an acceptable reference standard for case
definition. The diagnostic test is objective or performed and interpreted without
knowledge of the patient’s clinical status. Study results allow calculation of measures of
diagnostic accuracy.
Class II A case control study of a broad spectrum of persons with the condition established by an
acceptable reference standard compared to a broad spectrum of controls or a cohort study
where a broad spectrum of persons with the suspected condition where the data was
collected retrospectively. The diagnostic test is objective or performed and interpreted
without knowledge of disease status. Study results allow calculation of measures of
diagnostic accuracy.
Class III A case control study or cohort study where either persons with the condition or controls
are of a narrow spectrum. The condition is established by an acceptable reference
standard. The reference standard and diagnostic test are objective or performed and
interpreted by different observers. Study results allow calculation of measures of
diagnostic accuracy.
Class IV Studies not meeting Class I, II or III criteria including consensus, expert opinion or a
case report.

47

View publication stats