Screening accuracy of the parent-completed Ages and Stages Questionnaires – second edition as a broadband screener for motor problems in preschoolers with autism spectrum disorders

Abstract

Children with autism spectrum disorders are at risk for motor problems. However, this area is often overlooked in the developmental evaluation in autism diagnostic clinics. An alternative can be to identify children who should receive intensive motor assessment by using a parent-based screener. The aim of this study was to examine whether the Ages and Stages Questionnaires – second edition may be used to identify gross and fine motor problems in children. High-functioning children with autism spectrum disorder (n = 43, 22–54 m) participated in this study. Sensitivity, specificity, predictive values and areas under the receiver operating characteristic curve were calculated by comparing the Ages and Stages Questionnaires – second edition scores to the developmental evaluation of the Peabody Developmental Motor Scale – second edition. The results revealed that both the Ages and Stages Questionnaires – second edition gross and fine motor domain may be used to identify children without motor problems. In contrast, sensitivity analyses revealed the likelihood of under screening motor problems in this population. The Ages and Stages Questionnaires – second edition met only the criteria of a fair to good accuracy to identify poor gross motor (sensitivity = 100%) and below-average fine motor development (sensitivity = 71%) in this sample. Hence, the capacity of the Ages and Stages Questionnaires – second edition to identify motor problems in preschoolers with autism spectrum disorder appears to be limited. It is recommended to include a formal standardized motor test in the diagnostic procedure for all children with autism spectrum disorder.

Keywords

Ages and Stages Questionnaires autism spectrum disorders broadband screening motor development

Introduction

Autism spectrum disorders (ASDs) are a set of heterogeneous neurodevelopmental conditions, characterized by early-onset difficulties in social communication and unusually restricted, repetitive behaviour and interests (American Psychiatric Association (APA), 2013). The worldwide population prevalence is about 1%, making ASD one of the most frequent childhood neurodevelopmental disorders (Lai et al., 2014). Comorbidity is common in ASD; for instance, up to 79% of individuals with ASD, including those with intellectual disabilities, have concurrent motor problems (Lai et al., 2014). In a comprehensive meta-analysis, Fournier et al. (2010) pooled the findings of 51 studies involving between-group comparisons for ASD and typically developing controls. They calculated the group differences with respect to motor problems, such as motor coordination, arm movement, gait and postural stability deficits to the appropriate control group, resulting in a significant standardized mean difference effect equal to 1.20 (p < 0.0001). This large effect indicates substantial motor coordination deficits in ASD. The authors suggest that motor problems may be considered as a cardinal feature of ASD and have significant functional implications. It leads to decreased abilities to perform activities of daily living, including getting dressed, handwriting and participating in educational, sport and recreational activities (Fournier et al., 2010). Accordingly, motor problems can lead to difficulties with autonomy, limit social opportunities and impact several quality-of-life outcomes in this population.

Gross and fine motor skills of young children with ASD become progressively more delayed with age, even when controlling for nonverbal problem-solving skills. Unfortunately, motor skills of young children with ASD, and especially of those without intellectual disability, are usually not a priority for early assessment and intervention teams who may focus primarily on communication and behavioural concerns (Lloyd et al., 2013). Hence, motor problems in young children with ASD are not being identified as early as possible. As a result, these children must wait to get the help they need. In order to prevent further decline in motor skills, gross and fine motor skills should be learned and practiced (Lloyd et al., 2013).

An alternative for a formal motor assessment can be to rely on parents’ concerns using a screening questionnaire. Screening is a brief assessment procedure designed to identify children who should receive more intensive assessment. Any concerns regarding motor development raised during motor screening should be promptly addressed by standardized motor tests (Council on Children with Disabilities (CCD) et al., 2006). Parent-based broadband developmental screeners are frequently used due to the fact that they are relatively inexpensive, accurate and may be completed in the home setting. Among parent-based broadband screens, the Ages and Stages Questionnaires – second edition (ASQ-2) are widely used (Squires et al., 1999). The questionnaires of ASQ-2 are designed to screen young children for developmental delays and to identify those children who are in need of further thorough evaluation and those who appear to be developing typically. The questionnaires cover five different domains: communication, gross motor, fine motor, problem-solving and personal social skills. Referral for further assessment is advised when the score on any domain falls below the cut-off point, which is set at 2 standard deviations (SDs) below the mean of the reference group (Kerstjens et al., 2009; Squires et al., 1997). The original ASQ has been proven to be reliable and cost-effective. The normative sample of 2008 children consisted of 19% low-risk and 81% high-risk children, including infants with medical risk factors discharged from the neonatal intensive care unit and children with environmental risk factors due to poverty. Overall sensitivity (75%) and specificity (86%) were acceptable in this sample (Squires et al., 1997). However, screening accuracy that is established on individuals from one population cannot automatically be attributed to another population (Altman and Bland, 1994b).

The aim of this study is to examine whether the broadband parent-completed age-specific ASQ-2 questionnaires may be used to screen motor problems in young children without intellectual disability who are diagnosed with ASD in autism diagnostic clinics. The assessment of the children took place in two stages. In the first stage, parental concerns were assessed using the gross and fine motor domain section of the ASQ-2. In the second stage, gross and fine motor development was measured with the use of the locomotion and visual-motor integration subtests of the Peabody Developmental Motor Scale – second edition (PDMS-2) (Folio and Fewell, 2000). The PDMS-2 is used in clinical and research settings for young children at risk for ASD and with ASD (e.g. Libertus and Landa, 2013; Provost et al., 2007; Vanvuchelen et al., 2011). Examiners who have given and interpreted the PDMS-2 in this study were certificated physical therapists experienced in working with and sensitive to the needs of children with an ASD. The PDMS-2 locomotion and visual-motor integration items are quite similar to the gross and fine motor tasks of the ASQ-2; therefore, the screening accuracy of the ASQ-2 was measured by comparing the ASQ-gross and ASQ-fine motor scores to different cut-off scores of the developmental evaluation of the PDMS-2.

Methods

Participants

In all, 43 preschoolers with ASDs (10 female and 33 male) participated in this study. We took a sub-sample of children with a performance IQ above 70 from a larger sample of children with ASD (see Vanvuchelen et al., 2011). Children’s mean chronological age was 39.8 months (SD = 8.3 months, age range = 22–54 months) and mean nonverbal mental age was 39.6 months (SD = 11.0 months, mental age range = 19–71 months).

Children were diagnosed according to a multidisciplinary clinical consensus classification for ASD in University Autism Clinics (Flanders, Belgium). In 39 participants, the clinical diagnosis was confirmed by the Autism Diagnostic Observation Schedule – Generic (ADOS-G) classification (Lord et al., 2003), whereby 14 participants received module 1 (mean total score = 9.4, SD = 1.8) and 25 participants module 2 (mean total score = 12.8, SD = 3.9). In four participants, the clinical diagnosis was not confirmed by the ADOS-G classification, whereby two participants received module 1 (total score of 5 and 6) and two participants module 2 (both with a total score of 7).

All participants were free from any medical condition and had no visual or hearing impairment. All families gave written informed consent for the participation of their child. The study was approved by the ethics committees of the University Hospitals Louvain, Antwerp, Brussels and Ghent (Flanders, Belgium) before the collection of data.

Measures

Nonverbal mental age

Standardized tests appropriate for the child’s age were used to assess the children’s nonverbal mental age: 12 participants were measured with the Dutch modification of the nonverbal version of the mental scale of the Bayley Scales of Infant Development – second edition – Nederlandse versie (BSID-II-NL; Van der Meulen et al., 2000) and 31 participants with the revised version of the Snijders-Oomen Nonverbal Intelligence Test for Children (SON-R 2.5-7; Tellegen et al., 1998).

Parental concern

The parents filled in the ASQ-2 (Squires et al., 1999). The ASQ-2 consists of 19 different questionnaires covering the age range of 4–60 months. The 22-month (n = 1), 24-month (n = 2), 27-month (n = 2), 30-month (n = 2), 33-month (n = 2), 36-month (n = 16), 42-month (n = 9) and 48-month (n = 9) questionnaires from the Dutch translation of the ASQ-2 were used. Only the data of the gross and fine motor domains were analysed for the purpose of this study. Each domain was assessed by six questions on developmental milestones. The gross motor section of the ASQ-2 for children between 36 and 41 months of age, for example, contains tasks about kicking, jumping, climbing stairs, standing on one foot and overhand throwing. The fine motor section of the ASQ-2, on the other hand, contains tasks about drawing figures, string items, cutting paper and holding a pen. They are chosen so as to represent a developmental quotient of 75%–100% (Kerstjens et al., 2009; Squires et al., 1997). Parents could answer the questions with ‘yes’, ‘sometimes’ or ‘not yet’, with a respective score of 10, 5 or 0 points, with a minimum sum score of 0 and a maximum sum score of 60 per domain. The questionnaires are scored by comparing the sum scores for each domain with the empirically derived cut-off points for that domain. The cut-off points were derived by subtracting 2 SDs from the mean for each domain and vary by developmental domain and by questionnaire interval. If a child’s score for any domain is at or below the cut-off point, the child is recommended for a referral for further developmental evaluation.

Gross and fine motor problems

A measure of motor development commonly used for preschool-aged children in University Autism Clinics (Flanders, Belgium) is the PDMS-2 (Folio and Fewell, 2000). The PDMS-2 is a criterion and norm referenced test, designed to measure motor abilities typically seen in the early developmental stages of life. The test consists of two domains (gross and fine motor) and six subtests categories (reflex, balance, locomotion, object manipulation, grasping and visual-motor integration). In this study, the gross and fine motor levels were measured with the use of the locomotion and visual-motor integration subtest. The 89-item locomotion subtest of the PDMS-2 measures a child’s ability to move from one place to another. The actions measured include walking, running, hopping, jumping and climbing stairs. The 72-item visual-motor integration subtest measures a child’s ability to use his or her visual perceptual skills to perform complex eye–hand coordination tasks, such as building with blocks, lacing, cutting and copying designs (Folio and Fewell, 2000). The PDMS-2 is criterion referenced; each test item is designed to evaluate the performance of the skill. Each test item has specific criteria that are measured to be either proficient (score 2), showing signs of mastery (score 1), or are not performed (score 0). According to the PDMS-2 manual, raw scores were converted to subtest standard scores (SSs), which allows examiners to make comparisons across subtests and between instruments. SS ranged between 1 and 20 (mean of 10, SD of 3). An SS of 6 or 7 indicates a below-average motor development. An SS 4 or 5 indicates a poor motor development. An SS below 4 indicates a very poor motor development (Folio and Fewell, 2000).

The PDMS-2 locomotion and visual-motor integration subscales evidence a high degree of reliability. This reliability is consistently high across three types of reliability: error associated with content sampling (Cronbach’s coefficient alphas of 0.96 and 0.95), error due to time sampling (test–retest reliability coefficients of 0.94 and 0.92) and error due to examiner variability in scoring (interscoring correlation coefficients of 0.99 and 0.98). The magnitude of the coefficients strongly suggests that the PDMS-2 possesses little test error and that users can have confidence in its results (Folio and Fewell, 2000).

Data analyses

The screening accuracy of the ASQ-2 parental report (binary classification of suspected vs not suspected of having motor problems) was assessed in relation to the PDMS-2 scores (binary classification of having vs not having motor problems). Four comparison methods were conducted for the evaluation of the screening accuracy of the ASQ-2 with respect to the SSs of the PDMS-2: method 1 PDMS-2 SS ⩽ 7 (below or equal to percentile rank 16), method 2 PDMS-2 SS ⩽ 6 (below or equal to percentile rank 9), method 3 PDMS-2 SS ⩽ 5 (below or equal to percentile rank 5) and method 4 PDMS-2 SS ⩽ 4 (below or equal to percentile rank 2).

For each method, the prevalence rate of gross and fine motor problems was calculated. Next, sensitivity and specificity were measured. Sensitivity and specificity are the percentages of true positives and true negatives that were correctly identified by the screener (Altman and Bland, 1994a). Next, the predictive values of the ASQ-2 were calculated. The positive predictive value (PPV) is the probability that children have a motor delay when the ASQ-2 is at or below the cut-off score. The negative predictive value (NPV) is the probability that children have no motor delay when the ASQ-2 is above the cut-off score. The predictive values of a screener in clinical practice, however, depend critically on the prevalence of the abnormality in the patients being tested, which may differ from the prevalence in a published study assessing the usefulness of the test (Altman and Bland, 1994b). Next, a receiver operating characteristic (ROC) curve was obtained by calculating the sensitivity and specificity of every observed data value and plotting sensitivity against 1-specificity (Altman and Bland, 1994c). An area under the curve (AUC) of 0.7 or higher reflects a fair to good accuracy in the screening capacity of an instrument (Swets, 1998). Finally, the percentages of children who are correctly (accuracy AC) and incorrectly (misclassification MC) identified either with or without motor problems were calculated for each method.

Results

Prevalence of motor problems

Tables 1 and 2 represent an overview of the prevalence rate of gross and fine motor problems depending on the cut-off scores of the PDMS-2, respectively. The mean SS for the PDMS-2 locomotion subtest was 7.7 (SD = 2.0) ranging between 4 and 12. Locomotion abilities of 23 of 43 children were in the average range (SS = 8–12). In other words, the prevalence rate of gross motor problems (SS ⩽ 7) was 46% (20 children). More specifically, 10 children had a below-average gross motor development (SS = 6 or 7) and 10 children had a poor gross motor development (SS = 4 or 5).

Table 1.

Overview of screening accuracy measures for ASQ-2 gross motor domain with respect to the PDMS locomotion subtest.

	Prevalence	SE (95% CI)	SP (95% CI)	PPV (95% CI)	NPV (95% CI)	AUC (95% CI)	p	AC (%)	MC (%)
PDMS-2 locomotion SS ⩽ 7	46% (n = 20)	30 (11–54)	86 (66–97)	66 (29–92)	58 (40–75)	0.58 (0.41–0.75)	0.34	60	40
PDMS-2 locomotion SS ⩽ 6	32% (n = 14)	35 (12–64)	86 (68–96)	55 (21–86)	73 (55–87)	0.61 (0.42–0.79)	0.24	70	30
PDMS-2 locomotion SS ⩽ 5	23% (n = 10)	40 (12–73)	84 (68–94)	44 (13–78)	82 (65–93)	0.62 (0.41–0.83)	0.23	74	26
PDMS-2 locomotion SS ⩽ 4	2% (n = 1)	100 (2.5–100)	80 (65–91)	11 (0–48)	100 (89–100)	0.90 (0.76–1.0)	0.17	81	19

ASQ-2: Ages and Stages Questionnaires – second edition; PDMS-2: Peabody Developmental Motor Scale – second edition; SE: sensitivity; CI: confidence interval; SP: specificity; PPV: positive predictive value; NPV: negative predictive value; AUC: area under the curve; AC: accuracy; MC: misclassification; SS: standard score.

Table 2.

Overview of screening accuracy measures for ASQ-2 fine motor domain with respect to the PDMS visual-motor integration subtest.

	Prevalence	SE (95% CI)	SP (95% CI)	PPV (95% CI)	NPV (95% CI)	AUC (95% CI)	p	AC (%)	MC (%)
PDMS-2 visual-motor integration SS ⩽ 7	25% (n = 11)	54 (23–83)	90 (74–98)	66 (29–92)	85 (68–95)	0.72 (0.53–0.92)	0.02*	81	19
PDMS-2 visual-motor integration SS ⩽ 6	16% (n = 7)	71 (29–96)	88 (73–96)	55 (21–86)	94 (80–99)	0.80 (0.59–1.0)	0.01*	86	14
PDMS-2 visual-motor integration SS ⩽ 5	6% (n = 3)	66 (9–99)	82 (67–92)	22 (2–60)	97 (84–99)	0.74 (0.42–1.0)	0.16	81	19

p<.05

The mean SS for the PDMS-2 visual-motor integration subtest was 9.2 (SD = 2.3) ranging between 5 and 14. Visual-motor integration abilities of 32 children (75%) were in the average range (SS = 8–12) or above average range (SS = 13 or 14). In other words, the prevalence rate of fine motor problems (SS ⩽ 7) was 25% (11 children). More specifically, eight children had a below-average fine motor development (SS = 6 or 7) and three children had a poor fine motor development (SS = 5).

Screening accuracy of the ASQ-2 gross motor domain with respect to the PDMS-2 locomotion subtest

Table 1 visualizes the proportion of children who were correctly and incorrectly identified by the ASQ-2 gross motor domain with respect to the four methods (different cut-off scores of the PDMS-2 locomotion SS). As shown in Table 1, the sensitivity of the ASQ-2 to identify children at risk for gross motor problems varied between 30% and 100%. The sensitivity was only acceptable with respect to poor gross motor development (method 4: SS ⩽ 4 indicating 2 SD below PDMS-2 mean). In contrast, specificity of the ASQ-2 gross motor domain varied between 80% and 86%. Hence, the accuracy of the ASQ-2 to correctly identify children who are free from gross motor problems was acceptable. As expected, PPV decreased and NPV increased with decreasing prevalence rates of motor problems. PPV values varied between 11% and 66%. Hence, for those children with delayed locomotion skills, the ASQ-2 gross motor domain scores were almost meaningless. NPV values varied between 58% and 100%. The ASQ-2 gross motor domain was only an accurate measure to exclude poor gross motor development (PDMS-SS ⩽ 4 and 5). The AUC values varied between 0.58 and 0.90. The ASQ-2 met only the criteria of a fair to good accuracy to identify poor gross motor development (method 4: SS ⩽ 4 indicating 2 SD below PDMS-2 mean). The majority (60%–81%) of the children were correctly identified with gross motor problems. In other words, between 19% and 40% of the children were misclassified.

Screening accuracy of the ASQ-2 fine motor domain with respect to the PDMS-2 visual-motor integration subtest

Table 2 visualizes the proportion of children who were correctly and incorrectly identified by the ASQ-2 fine motor domain with respect to the different methods and cut-off scores of the PDMS-2 visual-motor integration SS. As shown in Table 2, the sensitivity of the ASQ-2 to identify children at risk for fine motor problems varied between 54% and 71%. Only for children with below-average fine motor development (method 3: SS ⩽ 6), the sensitivity (71%) was acceptable. Specificity of the ASQ-2 fine motor domain varied between 82% and 90%. The accuracy of the ASQ-2 to correctly identify children without fine motor problems was acceptable. PPV values varied from 22% to 66%. Hence, for those children with delayed visual-motor integration skills, the ASQ-2 fine motor domain scores were almost meaningless. In contrast, NPV values varied from 85% to 97%. The ASQ-2 fine motor domain was an accurate measure to exclude fine motor problems. The AUC values varied between 0.72 and 0.80 suggesting that the ASQ-2 met the criteria of a fair to good accuracy to identify fine motor problems regardless of its severity. The majority (between 81% and 86%) of the children were correctly identified with fine motor problems. Only a minority (between 14% and 19%) of the children were misclassified.

Discussion

To the best of our knowledge, this is the first study to examine the screening accuracy of the ASQ-2) for motor problems in a sample of high-functioning preschoolers with ASDs within the context of autism diagnostic clinics. Based on the results of the PDMS-2, 46% of the children have gross motor problems, and 25% fine motor problems defined as at least 1 SD below the mean (i.e. SS below or equal to 7, an equivalent of percentile rank 16). Using PDMS-2 to identify motor problems, we found that the screening accuracy of the ASQ-2 for gross and fine motor problems was rather low. The ASQ-2 gross motor domain was only able to identify poor gross motor development (minus 2 SDs below the age-specific mean), and not at all below-average gross motor development. In contrast, only in children with below-average fine motor development (−1 SDs below the age-specific mean), the sensitivity of the ASQ-2 fine motor domain was acceptable. On the other hand, specificity analyses revealed the likelihood of correct identification of children without gross and fine motor problems. The specificity of the ASQ-2 gross as well as fine motor domain was acceptable. Thus, the ASQ-2 is useful to identify those children with ASD without motor problems. Based on this parental information, a considerable number of children were correctly identified as not necessarily being picked up for a thorough motor assessment. However, while both these values are important, greater consideration should be given to the sensitivity of the screening tool. Since the ASQ-2 questionnaires are meant to identify at-risk children for further evaluation, and not to provide a definitive diagnosis of motor delay, it seems more prudent to maximize sensitivity so as to miss the fewest number of possible cases. The poor screening accuracy of the ASQ-2 was related to poor sensitivity. The ASQ-2 questionnaires seem to incline to one side, in particular to the incorrect identification of children with ASD as normal with respect to motor development. If a screener incorrectly identifies a child as normal, it results in under referrals for a thorough motor assessment. In this case, based on the parental information, a considerable number of children would not have been picked up for a formal standardized motor assessment, and by consequence will not get appropriate help.

The finding of this study contrasts with results of the original study which revealed that overall sensitivity of the ASQ was 75% (Squires et al., 1997). There are several possible explanations for so many false-negative cases accordingly to the ASQ-2 scores. First, in Squires et al. (1997), sensitivity was measured by comparing the classifications of the child’s performance based on the parent-completed ASQ with the classification of the child’s performance on another standardized test, the BSID-II. Importantly, in their study a sample of very young children (ranging between 4 and 36 months of age) with severe disabilities were involved. The majority (81%) of the 2008 children in their normative sample were children at risk for developmental problems. The obtained sensitivity values in this study are also in contrast to other research validating the ASQ-2 against the BSID-II in an Australian sample of ex-premature born children (Skellern et al., 2001). It is important to note, however, that we did use the PDMS-2 instead of the BSID-II. The choice for the PDMS-2 was motivated by the age range of our participants (22–54 m) which rather fitted the age norms of the PDMS-2 (0–60 m) and not of the BSID-II (0–42 m). Only a moderate to high association and a low concurrent validity exists between the SSs of these tests (Provost et al., 2004). Therefore, a comparison between both designs is not entirely applicable.

On the other hand, the results of a recent French study in ex-premature born children revealed that the capacity of the ASQ-2 to identify children with mild delay appears also to be limited (Halbwachs et al., 2013). In addition, the results of Gollenberg et al. (2010) indicate a high sensitivity (SE = 100%) and moderate specificity (SP = 87%) of the ASQ-2 for severely delayed status in contrast to a very low sensitivity (39%) for mild delays. The latter is quite similar to the results of the current study. The preschoolers in our sample were rather mildly delayed with respect to their motor development.

Another explanation of the low sensitivity in our study could be that we used the US version of the ASQ-2. The questionnaires were translated in Dutch and back-translated into English by two independent translators. However, we did not examine the psychometric properties of the questionnaires after translation in our own cultural setting of Flanders, the Dutch-speaking part of Belgium. The same was true for the PDMS-2. However, the results of a representative cohort of 120 preschoolers (mean chronological age = 65 months, SD = 3.1 months) revealed that Flemish children performed as well as American children on all subtests of the PDMS-2, with exception of visual-motor integration. On this fine motor subtest, Flemish children performed above average (mean SS = 13.4, SD = 1.8). This is probably due to the fact that school entrance in our country starts from the age of 30 months (Vanvuchelen et al., 2003). The latter may also be an explanation of the low prevalence rate of fine motor problems detected by both the PDMS-2 and ASQ-2 in this study.

A final and maybe the most important clarification of the low sensitivity level in this study could be the context of an autism diagnostic clinic. Without any doubt, relying on parents’ concerns is useful because these are accurate and efficient indicators of problems. Parents have far greater opportunity to observe their children than professionals do. Parents are enormously interested in their child’s development and so welcome the opportunity to answer questions about how their child is learning and behaving (Glascoe and Marks, 2011). Although it is plausible that parents of children referred to an autism diagnostic clinic worry about their children’s behaviour and social–communicative development and not (yet) about their gross and fine motor development, although these problems are proven to negatively impact many activities of daily living (Fournier et al., 2010). It can be that parents have misrepresented their children’s motor abilities due to their major concern regarding social and communicative development and repetitive and stereotype behaviour. For that reason, we believe that this broadband questionnaire is not suitable to be used in the context of autism diagnostic clinics to select children with mild motor problems for formal motor assessment.

As with all research, the results of this study must be considered within the limitations of its design and sample. Although the PDMS-2 (e.g. Libertus and Landa, 2013; Provost et al., 2007; Vanvuchelen et al., 2011) and the ASQ-2 (e.g. Eom et al., 2014; Hardy et al., 2015) are used in clinical and research settings for young children with ASD, a noteworthy limitation of this study is that neither the PDMS-2 nor the ASQ-2 was validated in children with ASD. The normative data of the PDMS-2 and ASQ-2 were based on US children, and applicability of these norms to detect mild motor impairments in Flemish children may be questionable. As we already mentioned, we did not examine the psychometric properties of the translated ASQ-2. It should also be noted that we simply used the PDMS-2 locomotion and visual-motor integration subtests to evaluate the accuracy of the ASQ-2 motor domains because the other PDMS-2 subscales were not administered. Furthermore, many children with ASD have difficulties to imitate an examiner. It is possible that children did not want to show a skill on demand during the PDMS-2 administration which may have led to an under-identification of their skills and over-identification of motor problems. On the other hand, examiners in diagnostic clinics are very experienced in assessing young children with ASD and the children were accompanied by their parents, who may have modelled the requested tasks. Finally, the sample size was rather small and only included children with an IQ above 70. This may have decreased the statistical power in our study and may complicate the generalization of the results. Therefore, the results should be interpreted with caution until the study is replicated in a larger sample, potentially including children with other developmental delays, and in other recruitment settings.

Conclusion

Low sensitivity of the ASQ-2 gross and fine motor subscales was found in a sample of preschoolers diagnosed with ASD in University Autism Clinics. Hence, the broadband parental-based ASQ-2 gross and fine motor subscales are not useful to identify children with motor problems in this context. These parental questionnaires may only be used to identify those children without motor problems. In the context of autism diagnostic centres, a formal assessment of motor skills in all children is recommended.

Footnotes

Acknowledgements

The authors would like to thank all children and parents who participated in this study. The authors also thank the student Lander Huybrechts (Faculty of Medicine and Life Sciences, Department Rehabilitation Sciences and Physiotherapy, Hasselt University, Belgium) for his assistance with the data analyses. The authors are grateful to Tin Vochten and Karine Rafferty-Swerts for skilfully editing the manuscript.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was supported by a foundation Marguerite-Marie Delacroix grant (Tienen, Belgium) to M. Vanvuchelen.

References

Altman

Bland

(1994a) Diagnostic tests 1: sensitivity and specificity. British Medical Journal 308(6943): 1552.

Altman

Bland

(1994b) Diagnostic tests 2: predictive values. British Medical Journal 309(6947): 102.

Altman

Bland

(1994c) Diagnostic tests 3: receiver operating characteristic plots. British Medical Journal 309(6948): 188.

American Psychiatric Association (APA) (2013) Diagnostic and Statistical Manual of Mental Disorders. 5th ed. (DSM-5). Washington, DC: APA.

Council on Children with Disabilities (CCD), Section on Developmental Behavioral Pediatrics, Bright Futures Steering Committee,. (2006) Identifying infants and young children with developmental disorders in the medical home: an algorithm for developmental surveillance and screening. Pediatrics 118: 405–420.

Eom

Fisher

Dezort

. (2014) Routine developmental, autism, behavioral, and psychological screening in epilepsy care settings. Developmental Medicine and Child Neurology 56(11): 1100–1105.

Folio

Fewell

(2000) Peabody Developmental Motor Scales. 2nd ed. (PDMS-2). Austin, TX: Pro Ed.

Fournier

Hass

Naik

. (2010) Motor coordination in autism spectrum disorders: a synthesis and meta-analysis. Journal of Autism and Developmental Disorders 40: 1227–1240.

Glascoe

Marks

(2011) Detecting children with developmental behavioral problems: the value of collaborating with parents. Psychological Test and Assessment Modeling 2: 258–279.

10.

Gollenberg

Lynch

Jackson

. (2010) Concurrent validity of the parent-completed Ages and Stages Questionnaires, 2nd ed. with the Bayley Scales of Infant Development II in a low-risk sample. Child: Care Health and Development 36: 485–490.

11.

Halbwachs

Muller

Nguyen The Tich

. (2013) Usefulness of parent-completed ASQ for neurodevelopmental screening of preterm children at five years of age. PLoS ONE 8: e71925.

12.

Hardy

Haisley

Manning

. (2015) Can screening with the Ages and Stages Questionnaire detect autism? Journal of Developmental and Behavioral Pediatrics 36(7): 536–543.

13.

Kerstjens

Bos

ten Vergert

. (2009) Support for the global feasibility of the Ages and Stages Questionnaire as developmental screener. Early Human Development 85: 443–447.

14.

Lai

Lombardo

Baron-Cohen

(2014) Autism. Lancet 383: 896–910.

15.

Libertus

Landa

(2013) The Early Motor Questionnaire (EMQ): a parental report measure of early motor development. Infant Behavior and Development 36(4): 833–842.

16.

Lloyd

MacDonald

Lord

(2013) Motor skills of toddlers with autism spectrum disorders. Autism 17: 133–146.

17.

Lord

Rutter

DiLavore

. (2003) Autism Diagnostic Observation Schedule. Los Angeles, CA: Western Psychological Services.

18.

Provost

Heimerl

Lopez

(2007) Levels of gross and fine motor development in young children with autism spectrum disorder. Physical & Occupational Therapy in Pediatrics 27(3): 21–36.

19.

Provost

Heimerl

McClain

. (2004) Concurrent validity of the Bayley Scales of Infant Development II Motor Scale and the Peabody Developmental Motor Scales-2 in children with developmental delays. Pediatric Physical Therapy 16(3): 149–156.

20.

Skellern

Rogers

O’Callaghan

(2001) A parent-completed developmental questionnaire: follow up of ex-premature infants. Journal of Paediatrics and Child Health 37: 125–129.

21.

Squires

Bricker

Potter

(1997) Revision of a parent-completed development screening tool: Ages and Stages Questionnaires. Journal of Pediatric Psychology 22: 313–328.

22.

Squires

Bricker

Potter

(1999) Ages and Stages Questionnaires User’s Guide. 2nd ed. Baltimore, MD: Paul Brookes Publishing.

23.

Swets

(1998) Measuring the accuracy of diagnostic systems. Science 240(4857): 1285–1293.

24.

Tellegen

Winkel

Wijnberg-Williams

. (1998) Snijders-Oomen Niet-verbale Intelligentietest: SON-R 2½-7: Handleiding en Verantwoording. Lisse: Swets Test Publishers.

25.

Van der Meulen

Ruiter

SAJ

Lutje Spelberg

. (2000) Bayley Scales of Infant Development, Second Edition, Nederlandse versie (BSID-II-NL). Lisse: Swets Test Publishers.

26.

Vanvuchelen

Mulders

Smeyers

(2003) Onderzoek naar de bruikbaarheid van de recente Amerikaanse Peabody Developmental Motor Scales-2 voor vijfjarige, Vlaamse kinderen. Signaal 45: 24–41.

27.

Vanvuchelen

Roeyers

De Weerdt

(2011) Imitation assessment and its utility to the diagnosis of autism: evidence from consecutive clinical preschool referrals for suspected autism. Journal of Autism and Developmental Disorders 41: 484–496.