Professional Documents
Culture Documents
See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
DEVELOPMENTAL MEDICINE & CHILD NEUROLOGY ORIGINAL ARTICLE
1 Department of Occupational Therapy, Kaohsiung Medical University, Kaohsiung. 2 Department of Physical and Rehabilitation Medicine, Kaohsiung Medical University Chung-Ho
Memorial Hospital, Kaohsiung. 3 Department of Physical and Rehabilitation Medicine, Kaohsiung Municipal Hsiaokang Hospital, Kaohsiung, Taiwan.
Correspondence to Dr Chwen-Yng Su, Department of Occupational Therapy, Kaohsiung Medical University, 100, Shi-Chuan 1st Rd., Kaohsiung City 807, Taiwan. E-mail: cysu@cc.kmu.edu.tw
PUBLICATION DATA AIM To examine the internal consistency, test–retest reliability, and responsiveness of the
Accepted for publication 19th August 2011. Movement Assessment Battery for Children–Second Edition (MABC-2) Test for children with
developmental coordination disorder (DCD).
ABBREVIATIONS METHOD One hundred and forty-four Taiwanese children with DCD aged 6 to 12 years (87 males,
DCD Developmental coordination disorder 57 females) were tested on three separate occasions: two baseline measurements with a 20-day
MABC-2 Movement Assessment Battery for interval before the intervention, and a follow-up measurement after 6 months of rehabilitation.
Children–Second Edition The therapists rated the performance of children in school-related physical tasks at baseline and
MDC Minimal detectable change after intervention.
MID Minimal important difference
RESULTS Internal consistency for the MABC-2 Test was a = 0.90. Test–retest reliability for the total
PTPS Physical tasks performance scale
ROC Receiver operating characteristic score was excellent, with an intraclass correlation coefficient of 0.97. A small to medium magni-
SRM Standardized response mean tude of treatment effect was captured by the MABC-2 Test. The minimal detectable change (MDC)
was 0.28 points whereas the minimal important difference (MID) values were from 2.36 to 2.50. All
subscales except balance showed acceptable validity in differentiating groups of children whose
physical performance had improved or remained stable.
INTERPRETATION The MABC-2 Test is a reliable and valid measure to assess motor competence
in children with DCD. The MID and MDC scores provide the reference point for clinical decision-
making in managing the individual child.
Developmental coordination disorder (DCD) is one of the yield consistent results across repeated measurements and
most commonly reported disorders in school-age children, detect subtle changes in motor function.
with prevalence rates of up to 6%.1 A most recent UK popula- The Movement Assessment Battery for Children (MABC)7
tion-based study found that 1.8% of 7-year-olds have DCD is one of the most widely used measures for identifying impair-
according to strict Diagnostic and Statistical Manual of Mental ments in motor performance of children and has been trans-
Disorders, 4th edition (DSM-IV) criteria.2 These children lated into several major European and Asian languages, such
form a heterogeneous group, differing in the extent of their as Italian, Chinese, and Japanese.8–10 The second edition
motor difficulties and the extent to which they display other (MABC-2)11 is an updated version of the original MABC and
conditions such as attention deficits, speech ⁄ articulation diffi- differs from the first in several respects, including age exten-
culties, and non-verbal learning disabilities.3,4 Several studies sion, reduction of age bands from four to three, item revision
have shown that children with DCD are at greater risk of devel- and addition of new items, installation of an innovative score
oping secondary impairments (e.g. decreased strength and lack interpretation method, inclusion of a more representative
of fitness) owing to their limited participation in physical activ- standardization sample, and rearrangement of subtests. The
ity.5,6 Furthermore, in the absence of intervention, many of the MABC-2 Test has shown good to excellent reliability in
these abnormalities persist into adulthood and might continue healthy children,11,12 but there are as yet no data in this area
to interfere with performance in various aspects of community for children with DCD. For validity, several investigators have
functioning. As a result, early identification and therapeutic studied the effectiveness of different treatment strategies for
intervention is particularly important to enhance motor func- children with DCD using the MABC.13,14 Nevertheless, in
tion and promote success in school and daily living in children none of these studies was an attempt made to determine the
with DCD. To monitor the effectiveness of an intervention, it ability of the MABC to detect small but important clinical
is crucial to use reliable and sensitive measures that are able to changes in motor performance over time (responsiveness).
160 DOI: 10.1111/j.1469-8749.2011.04177.x ª The Authors. Developmental Medicine & Child Neurology ª 2012 Mac Keith Press
14698749, 2012, 2, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1469-8749.2011.04177.x by Spanish Cochrane National Provision (Ministerio de Sanidad), Wiley Online Library on [19/11/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
These are changes that have been induced by treatments of What this paper adds
known efficacy, are associated with changes in another out- • It assesses the reliability of the Movement Assessment Battery for Children–
come measure of similar constructs, or that clients and clini- Second Edition (MABC-2) Test in developmental coordination disorder (DCD).
cians perceive as important.15 Knowledge of an instrument’s • It evaluates the responsiveness of the MABC-2 Test in DCD.
responsiveness aids the interpretation of score changes after • It performs cross-cultural validation of the MABC-2 Test to determine its clini-
cal use in children with DCD in Taiwan.
treatment and selection of an appropriate outcome measure.
To address such shortcomings in previous reliability and sixth grade.17 It is a judgment-based criterion-referenced
validity studies, the purpose of the present work was to evalu- assessment that comprises three sections: participation, task
ate the MABC-2 Test’s internal consistency, test–retest reli- supports, and activity performance. There are two major cate-
ability, and responsiveness to change in Taiwanese children gories in the activity performance section: physical tasks and
with DCD. cognitive ⁄ behavioural tasks. The physical tasks performance
scale (PTPS) was used as an external criterion for clinically rel-
METHOD evant change in this study. The PTPS consists of 161 items
Participants (activities) divided into 12 domains and is rated on a scale of 1
Children with DCD were recruited from 13 hospitals, seven to 4, where 1=does not perform, 2=partial performance,
clinics, and 14 elementary schools in southern Taiwan. Inclu- 3=inconsistent performance, and 4=consistent performance.
sion criteria for participation in the study were (1) age between Reliability and discriminant validity of the Chinese version of
6 and 12 years, (2) a diagnosis of DCD according to DSM-IV the School Function Assessment were similar to those of the
criteria,16 (3) absence of serious emotional or behavioural dis- original English version.18,19
turbances, and (4) participation in physical or occupational
therapy programmes at the time of research. Children who had Procedure
previous history of neurological disorders such as traumatic This study was approved by the Institutional Review Board at
brain injury, muscular dystrophy, and epilepsy were excluded. the Kaohsiung Medical University Chung-Ho Memorial Hos-
A total of 216 children met the inclusion and exclusion criteria. pital. After obtaining informed consent from all parents, the
Of these, 72 (33.3%) children dropped out of the study for var- principal investigator (Y-PW) of the study invited physical
ious reasons. In the final sample (n=144), 40% (n=57) were and occupational therapists (n=15) who treated these children
female, and the mean age was 7 years 7 months (SD 2y 1mo, before the study to participate in the intervention stage. All
range 6y–12y 9mo). The number of children in each age group children received a conventional paediatric rehabilitation pro-
was as follows: ages 6 years to 6 years 11 months (n=22); gramme, at least 1 day a week, for 6 months.
7 years to 7 years 11 months (n=22); 8 years to 8 years Children were assessed with the MABC-2 Test on three
11 months (n=21); 9 years to 9 years 11 months (n=21); different occasions. Two baseline measurements (T1 and T2)
10 years to 10 years 11 months (n=20); 11 years to 11 years were performed with a 20-day interval in between. No inter-
11 months (n=19); and 12 years to 12 years 11 months (n=19). ventions were undertaken between these two time points. A
third follow-up measurement (T3) was conducted immedi-
Instruments ately after completing the 6-month rehabilitation programme.
MABC-2 Test All three assessments were performed by the principal investi-
The MABC-2 Test is designed to identify and describe gator, a certified occupational therapist with 15 years’ clinical
impairments in motor performance of children and adoles- experience in paediatric rehabilitation, using standardized
cents from 3 to 16 years of age divided into three age bands procedures for administration specified in the test manual.
(3–6y, 7–10y, 11–16y). The test is composed of two parts: the Assessments were made at the same time of day for each visit.
Performance Test and the Checklist. Only the Performance To decrease possible experimenter bias, the examiner did not
Test was used in the current study; it involves children com- re-acquaint herself with the child’s scores from the first assess-
pleting a series of eight fine and gross motor tasks grouped ment when conducting the retests. Testing lasted approxi-
into three subscales: Manual Dexterity, Aiming and Catching, mately 30 minutes and was performed on an individual basis
and Balance. The MABC-2 Test yields several scores: raw in quiet locations identified at each child’s respective school,
score and standard score on the individual test items, subscale or paediatric occupational therapy unit at hospitals. In addi-
standard score with equivalent centiles, and a total standard tion, seven therapists who did not participate in the interven-
score and centile equivalent. The total impairment score is tion stage were recruited to complete the PTPS
translated to age-related centile norms with the fifth centile as independently at T1 and T3. The therapists were provided
the cut-off point to indicate children with movement difficul- with written and oral instructions for filling out the PTPS and
ties. Standard scores have a mean of 10 and an SD of 3. had high levels of agreement (intraclass correlation coefficient
[ICC]=0.95–0.99).
School Function Assessment–Chinese version
The School Function Assessment–Chinese version is designed Statistical analysis
to measure a student’s performance of functional tasks that All data were analysed using SPSS for Windows (SPSS, Chi-
support his or her participation in both academic and social cago, IL, USA). Standard scores of the MABC-2 Test were
school-related activities for students in grades kindergarten to used in the following analyses.
Reliability and Responsiveness of the MABC-2 Test Yee-Pay Wuang et al. 161
14698749, 2012, 2, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1469-8749.2011.04177.x by Spanish Cochrane National Provision (Ministerio de Sanidad), Wiley Online Library on [19/11/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Distribution the correspondence between the chosen anchor and the
The score distribution of the MABC-2 Test was examined for change scores from the MABC-2 Test. Each of the above-
floor and ceiling effects. Floor or ceiling effects exceeding mentioned methods was further specified as follows.
20% of the participants were considered substantial.20
Effect size
Internal consistency The effect size is a standardized measure of change obtained
Internal consistency indicates the extent to which items are all by dividing the mean change between initial and the third fol-
measuring the same construct. Cronbach’s alpha (a) coefficient low-up measurements by the SD of the initial measurement.
was calculated for internal consistency of the MABC-2 Test As a guide to interpreting these values, Cohen25 labelled an
using data on all children receiving initial assessment. Values effect size ‘small’ if the effect size is between 0.2 and 0.5, ‘mod-
of Cronbach’s a above 0.7 are generally regarded as accept- erate’ if the effect size is between 0.5 and 0.8, or ‘large’ if the
able, over 0.8 good, and over 0.9 excellent.21 effect size is at least 0.8.
Subscales
Table II: Test–retest reliability of the Movement Assessment Battery for Manual 1.4 (1.7) )1 to 5 0.63 0.68 0.72 0.76 1.50
Dexterity
Children–Second Edition (MABC-2) Test (n = 144) Aiming and 1.5 (2.8) )4 to 7 0.49 0.54 2.14 1.78 2.50
Catching
MABC-2 ICC (95% CI) SEM Balance 1.1 (2.8) )5 to 9 0.37 0.41 1.26 0.52 0.50
Total score 1.3 (2.1) )1 to 7 0.42 0.63 1.21 2.36 2.50
Items
a
Manual Dexterity 1 0.96 (0.94–0.97) 0.41 Difference scores were calculated by subtracting the first baseline
Manual Dexterity 2 0.98 (0.97–0.99) 0.30 score from the third retest score (i.e. T3–T1). A positive score means
Manual Dexterity 3 0.92 (0.89–0.95) 0.51 that performance increased at retest, a negative score means that
Aiming and Catching 1 0.88 (0.83–0.92) 0.74 performance decreased at retest. The numbers under ‘range’ represent
Aiming and Catching 2 0.96 (0.95–0.98) 0.61 the lowest and highest difference score.
b
Balance 1 0.99 (0.98–0.99) 0.35 Standardized response mean (SRM) is calculated as the average
Balance 2 0.97 (0.95–0.98) 0.44 change in scores between initial and the third follow-up
Balance 3 0.96 (0.95–0.97) 0.62 measurements divided by the SD of that change score.
c
Subscales Minimal detectable change (MDC) was calculated as 1.65 · 2 · SEM,
Manual Dexterity 0.97 (0.95–0.98) 0.31 at a 10% level of significance (90% confidence interval).
d
Aiming and Catching 0.91 (0.82–0.95) 0.92 Minimal important difference (MID) was defined as the difference in
Balance 0.97 (0.95–0.98) 0.54 mean change of the MABC-2 Test scores between children classified
Total score 0.97 (0.96–0.98) 0.52 as ‘improved’ and those classified as ‘no change’.
e
The Receiver operating characteristic (ROC) curves were used to
ICC, intraclass correlation coefficient; SEM, standard error of examine various cut-off values for the MABC-2 Test change scores for
measurement. their sensitivity and specificity in detecting clinical improvement.
Reliability and Responsiveness of the MABC-2 Test Yee-Pay Wuang et al. 163
14698749, 2012, 2, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1469-8749.2011.04177.x by Spanish Cochrane National Provision (Ministerio de Sanidad), Wiley Online Library on [19/11/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
PTPS scores. None of the children had a change in PTPS effect of all subscales. One possible reason might be that bal-
score of less than 0. The MID for the MABC-2 Test total ance skills measured by this subscale were somewhat simple
score was identified to be 2.36 points; that is, a change of 2 for children with DCD because over half of them (58%)
points or more is likely to represent a therapist-perceived scored between the mean and 1SD above the mean at baseline.
important change in the MABC-2 Test (Table III). A multi- This leaves little room for improvement. Effect size scores
variate analysis of variance applied to the mean change scores were found to be slightly lower than SRM scores, suggesting
for the three subscales revealed significant differences between that the SD of the initial assessment scores (used in calculating
the two groups (Wilks’ k = 0.84, F4,96 = 6.123, p < 0.01). Sub- the effect size) were more heterogeneous than the SD of
sequent analyses of variance on the individual composite change scores (used in calculating the SRM).
scores were performed. Significant univariate differences were Given that average effects across a group might not be
found in the Manual Dexterity (F = 211.235, p < 0.001), Aim- meaningful to the individual child, reference values are
ing and Catching (F = 322.005, p < 0.001), and Balance required to determine if changes in the child’s scores before
(F = 184.382, p = 0.003) subscales, with the ‘improved’ group and after the intervention reach statistical significance. The
scoring higher than the ‘no change’ group. MDC is such a reference, which can guide clinicians in mak-
The MID calculated from the ROC curve using the cut-off ing the decision to modify the child’s therapy programme
point nearest the upper left-hand corner of the graph was 2.5 accordingly. In other words, this information allows the clini-
points for the MABC-2 Test total score (sensitivity 71%, spec- cians to differentiate between true change and change due to
ificity 46%), 1.5 points for the Manual Dexterity subscale (sen- measurement error in score. The MDC at the 90% confidence
sitivity 63%, specificity 49%), 2.5 points for the Aiming and interval ranged from 0.72 to 2.14 points for the total and sub-
Catching subscale (sensitivity 76%, sensitivity 49%), 0.5 scales of the MABC-2 Test, and from 0.70 to 1.72 points for
points for the Balance subscale (sensitivity 60%, specificity the items. In contrast, the MID serves as another threshold
44%). The values of the area under the curve for the changes that determines the importance of changes in outcome scores
in total score, Manual Dexterity, Aiming and Catching, and from the clinician’s viewpoint. Our results indicated that the
Balance subscales were 0.80, 0.64, 0.68, and 0.50 respectively. values of the MID calculated from the difference in mean
change scores between the improved and stable groups were
DISCUSSION in the neighbouring range to those from the ROC curves,
This study examined, for the first time, the reliability and the ranging from 0.5 to 2.5 points. These MID values are benefi-
responsiveness of the MABC-2 Test in 6 to 12-year-old cial in terms of aiding the clinical decision-making process
Taiwanese children with DCD and showed that this test is a because the anchor (performance on school-related motor
reliable and valid outcome measure, one that can be used in tasks) was measured reliably and was relevant to the motor
clinical practice to monitor treatment progress. The MABC-2 skills tapped by the MABC-2 Test. However, care must be
Test demonstrated good to excellent internal consistency taken in using the MID score of the Balance subscale gener-
(Cronbach’s a range 0.81–0.90). The total instrument, as well ated from the ROC statistic when interpreting a child’s perfor-
as its separate items and subscales, exhibited excellent test– mance, as this subscale had poor discriminating ability.
retest reliability. These ICC values were higher than those The major limitation of the present study is that generaliza-
reported for healthy adolescents.29 Apart from these, because tion of the findings is limited to Taiwanese children with
the reliability coefficients alone do not provide sufficient infor- DCD aged 6 to 12 years. Further studies including children
mation about the precision of test scores of individual child, aged 3 to 5 years and 13 to 16 years are warranted to explore
this study provides additional psychometric data about the var- fully the use of the MABC-2 for younger children and adoles-
iability of the errors of measurement based on the SEM. The cents with DCD. In addition, our findings should be
SEM for the total MABC-2 Test was 0.52, for the items ran- replicated in Western countries as studies have found cross-
ged from 0.30 to 0.74, and for the subscales ranged from 0.31 cultural differences in some M-ABC items when comparing
to 0.92. These values were smaller than those reported in performance between preschool children from Asian countries
healthy children (range 1.20–1.56),11 lending further support and those from the USA.12
to the stability of the MABC-2 Test scores over repeated test- In conclusion, the MABC-2 Test can produce reliable
ing in children with DCD. results in repeated assessments over a 20-day interval and is
The values of effect size and SRM for the MABC-2 Test responsive to change in children with DCD enrolled in reha-
corresponded to small to moderate effect sizes except for the bilitation programmes. The MID and MDC scores serve as
first item (Balance 1) of the Balance subscale, which yielded guidelines for clinicians in determining intervention effective-
negligible effect. Indeed, the Balance subscale had the lowest ness from either clinical judgment or statistical significance.
REFERENCES
1. Wann J. Current approaches to intervention in children with der using the DSM-IV at 7 years of age: a UK coordination disorder using different motor tests. Res Dev
developmental coordination disorder. Dev Med Child Neurol population-based study. Pediatrics 2009; 123: e693– Disabil 2009; 30: 1367–77.
2007; 49: 405. 700. 4. Tsai CL, Pan CY, Cherng RJ, Hsu YW, Chiu HH. Mecha-
2. Lingam R, Hunt L, Golding J, Jongmans M, Emond 3. Chen YW, Tseng MH, Hu FC, Cermak SA. Psychosocial nisms of deficit in visuospatial attention deficit in children
A. Prevalence of developmental coordination disor- adjustment and attention in children with developmental with developmental coordination disorder: a neurophysiolog-
Reliability and Responsiveness of the MABC-2 Test Yee-Pay Wuang et al. 165