DEVELOPMENTAL MEDICINE & CHILD NEUROLOGY ORIGINAL ARTICLE

The Trunk Control Measurement Scale: reliability and

discriminative validity in children and young people with
neuromotor disorders
PETRA MARSICO 1,2 | ELENA MITTEREGGER3,4 | JULIA BALZER1,2 | HUBERTUS J A VAN HEDEL 1,2
1 Paediatric Rehab Research Group, Rehabilitation Centre, University Children’s Hospital Zurich, Affoltern am Albis; 2 Children’s Research Centre, University Children’s
Hospital Z€urich, Z€urich; 3 ZHAW, Institute for Physiotherapy, Z€urich University of Applied Studies, Z€urich; 4 Children Therapy Centres of the Foundation Regional
Group Zurich (Stiftung RgZ), Z€urich, Switzerland.
Correspondence to Petra Marsico at Paediatric Rehab Research Group, Rehabilitation Centre for Children and Adolescents, University Children’s Hospital Z€urich, M€uhlebergstrasse 104,
CH-8910 Affoltern am Albis, Switzerland. E-mail: [email protected]

This article is commented on by Saavedra on page 671 of this issue.

PUBLICATION DATA AIM This study investigated the intra- and interrater reliability of the Trunk Control
Accepted for publication 8th February 2017. Measurement Scale (TCMS) German version, with its subscores, in children with neuromotor
Published online 4th April 2017. disorders. Further, the discriminative validity of the TCMS was assessed by comparing the
TCMS scores with the Functional Independence Measure for children.
ABBREVIATIONS METHOD Bland–Altman analyses and intraclass correlation coefficients were applied to
AUC Area under the curve investigate reliability. The discriminative ability of the TCMS was evaluated with receiver
GMFCS Gross Motor Function operating characteristics.
Classification System RESULTS Ninety children (mean age 11y 5mo; range 5y–18y 11mo) participated for the
ICC Intraclass correlation coefficient reliability, and 50 for the discriminative validity study. The reliability proved to be excellent
SDD Smallest detectable difference (intrarater: bias=0.57 points, 95% confidence interval [CI] 3.71 to 4.85; interrater: bias= 0.31
TCMS Trunk Control Measurement points, 95% CI 5.77 to 5.10). A change in the TCMS total score of six points (10%) can be
Scale considered a true change. The TCMS subscores appeared to be clinically relevant because
WeeFIM Functional Independence children with less than around 80% of the static balance score, less than 55% of the dynamic
Measure for children reaching score, or less than around 35% of the selective movement control score needed
support for daily life activities.
INTERPRETATION The TCMS is a reliable and clinically relevant assessment for children aged
5 years and older with different neurological impairments.

The central role of trunk control on function and activity dynamic sitting balance; the latter is divided into selective
in children and young people with neuromotor disorders is movement control and dynamic reaching.7 This assessment
currently being discussed in the literature.1,2 Often, postu- tool shows good relative reliability in children aged 8 to
ral control is impaired in persons with a neurological con- 15 years with spastic CP.7 Information on absolute measure-
dition, and this can affect trunk stability, mobility, and ment errors is needed to interpret whether observed changes
selectivity of the upper limbs.2–5 In clinical practice, we can be considered ‘true’ or not. While the same study inves-
address these impairments with different approaches, such tigated also the absolute measurement error of the total
as balance training during physical therapy or hippother- score (the smallest detectable difference [SDD] was 4.66
apy. To assess the efficacy of such interventions, we need points for intrarater reliability and 5.47 points for interrater
valid, reliable, and responsive outcome measures to assess reliability), such information is missing for subscores of the
(changes in) trunk control. TCMS. Furthermore, these results were obtained in a rela-
In a recent review, the psychometric properties of sitting tively small group of children (n=26) and were restricted to
balance measures for children with cerebral palsy (CP) were children with spastic CP. In clinical practice, impaired trunk
discussed.6 Seven outcome assessments were included: the control can also be observed in younger children and those
Pediatric Reach Test; the Level of Sitting Scale; the Sitting with other neurological diagnoses (e.g. dyskinetic or ataxic
Assessment for Children with Neuromotor Dysfunction; the CP and other neuromotor disorders, such as acquired brain
Segmental Assessment of Trunk Control; the Seated Postu- injuries, myelomeningocele, or different syndromes), so it is
ral Control Measurement; the Trunk Impairment Scale; and important to investigate whether these findings also apply to
the Trunk Control Measurement Scale (TCMS).6 With the a broader group.
TCMS, an objective outcome measure is available to score Besides the reliability of the TCMS, we were also inter-
trunk ability in sitting. The TCMS consists of static and ested to learn how much trunk control a child needs to be

706 DOI: 10.1111/dmcn.13425 © 2017 Mac Keith Press

 14698749, 2017, 7, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/dmcn.13425 by Universidad Nacional Mayor De San Marcos, Wiley Online Library on [26/04/2023]. 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
independent for mobility and self-care in daily life, as the What this paper adds
child’s independence in daily living is important for the chil- • The Trunk Control Measurement Scale (TCMS) is highly reliable in children
dren and their families. Therefore, we also investigated the with various neuromotor disorders.
discriminative validity by comparing the TCMS and its sub- • The TCMS is reliable in children aged 5 years and older.
scores with the mobility and self-care domains of the Func- • Measurement errors of TCMS total and subscores allow for the
interpretation of longitudinal changes.
tional Independence Measure for children (WeeFIM).8,9 • The TCMS subscores are relevant concerning daily life, mobility, and
self-care.
METHOD
Participants The WeeFIM consists of three categories: self-care
Children were recruited from the inpatient and outpatient (eight items), mobility (five items), and cognition (five
setting of the Rehabilitation Centre for Children and Ado- items). Each item is rated on a 7-point ordinal scale. While
lescents of the University Children’s Hospital Z€ urich in a score of 1 indicates that total assistance is required, a
Affoltern am Albis and the Children Therapy Centres of score of 7 indicates complete independence. We were
the Foundation Regional group Z€ urich (Stiftung RgZ). interested in the self-care and mobility categories, because
Inclusion criteria were: (1) neurological diagnosis such as we expected that trunk control should be relevant for these
CP (Gross Motor Function Classification System categories. To assess the self-care domain, three of eight
[GMFCS] level I–IV); (2) acquired brain injury; (3) spinal items were selected: eating, dressing upper body, and
cord injury; and (4) age 5 to 19 years. The children needed dressing lower body. Grooming, bathing, toileting, blad-
to have the ability to follow easy-to-understand instruc- der, and bowel function were excluded, because these mea-
tions. Exclusion criteria were surgery or botulinum toxin sures do not solely depend on voluntary motor function
injection within the last 3 months, and pain or medical (e.g. modifications in the environment or autonomic blad-
restriction for weight bearing. Parents and adolescents der or bowel control). For the mobility section, all items
aged 15 years and above signed an informed consent form. were included: transfer to a chair or wheelchair, transfer to
Children below the age of 15 years agreed to participate in the toilet, transfer to the bath or shower, mobility such as
this study. The study was approved by the ethics commit- walking or by means of a wheelchair, and walking stairs.
tee of the Canton of Z€ urich, complied with the Declara-
tion of Helsinki, and followed the guidelines of good Assessment procedure
clinical practice. The aim was to collect data on at least 50 Two physiotherapists (EM and PM) with more than
children. According to the COSMIN group, a sample size 10 years of experience in treating children and adolescents
greater than 50 is considered good.10 with neurological disorders tested all children in a quiet
room. The child sat on a height-adjustable bench with the
Measures feet unsupported. Orthoses, shoes, and socks were taken
The TCMS has been translated and validated in the Kor- off. The testing lasted 20 to 30 minutes. The testing was
ean and German language.11,12 We performed the TCMS recorded by video in the frontal plane except for items 2,
according to the German description of Mitteregger 6, 7, and 12 (these items were recorded in the sagittal
et al.12 The TCMS total score and the three subscores plane for better analysis). For the intrarater reliability, the
were evaluated. The maximum value for the total TCMS TCMS tests were scored by the same rater, first directly
is 58 points (20 for the category ‘static sitting balance’, 28 after the original test situation from the video and the sec-
for ‘selective movement control’, and 10 for ‘dynamic ond time from the video at least 2 months later. For the
reaching’). A higher TCMS scores indicates a better per- interrater reliability, both therapists (rater A and B) used
formance in trunk control. the video recordings to score the TCMS more than
The TCMS score correlates well with the GMFCS and 2 months after the assessment was conducted. These rat-
the Gross Motor Function Measure.4,7,12 While the ings were compared.
GMFCS and the Gross Motor Function Measure provide
information about gross motor function and capacity Data analysis
respectively, they do not assess independence in daily life We assumed that if the TCMS could differentiate well (i.e.
activities.13,14 To assess daily life independence, we had a a good discriminative validity) between children being
trained research nurse who routinely assessed the WeeFIM dependent versus independent in self-care and mobility,
values of inpatients. The WeeFIM is an 18-item tool used this could indicate the clinical relevance of the TCMS.
to observe daily life performance and independence in chil- Therefore, we dichotomized the WeeFIM self-care (three
dren. The assessment is suitable for application in all chil- items) and mobility (five items) subcategory scores. To be
dren and adolescents whose functioning level lies below defined as independent the child had to reach a value ≥5 in
that of a typically developed peer of 7 years old (i.e. a typi- each included item – that is, children who were able to
cally developed 7-year-old child should achieve the maxi- perform the activity without personal assistance (from
mum score of the WeeFIM). In addition, it can be applied supervision until complete independence, i.e. WeeFIM
to children with developmental disabilities aged 6 months item scores of 5–7) were grouped as ‘independent’. In
to 21 years. contrast, children who could not perform the activity

Reliability of the Trunk Control Measurement Scale Petra Marsico et al. 707
 14698749, 2017, 7, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/dmcn.13425 by Universidad Nacional Mayor De San Marcos, Wiley Online Library on [26/04/2023]. 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
(i.e. WeeFIM item score of 0) or needed assistance to per- the highest sensitivity plus highest specificity minus one).20
form the activity (total assistance until minimal assistance, The area under the curve (AUC) was taken as an indicator
i.e. WeeFIM item score of 1–4) were grouped as ‘depen- of the accuracy and interpreted by the following bench-
dent’. We included the scores made by rater A for these marks: outstanding (AUC ≥0.90); excellent (AUC 0.80–
comparisons. 0.89); acceptable (AUC 0.70–0.79). In general, a was set at
0.05.21
Statistical analysis
We performed statistical calculations with SPSS (IBM RESULTS
SPSS Statistics 19; IBM, Armonk, NY, USA). All variables Participants
were visually inspected for normal distribution (histogram), Ninety children (44 females, 46 males) with a mean age of
and skewness and kurtosis were analysed. 11 years and 5 months (range 5y–18y 11mo) were included
For the relative intra- and interrater reliability analysis of in the reliability study. Twenty-one children were younger
the TCMS total and subscores, intraclass correlation coeffi- than 8 years (mean age 6y 3mo [SD 10mo]). The distribu-
cients (ICC) and 95% confidence intervals (CIs) with the tion of diagnoses and GMFCS levels of both groups (relia-
two-way random-effect model were calculated.15 The ICC bility and clinical importance) are shown in Table I.
was calculated as follows: ICC=rs2/(rs2+rt2+re2), where WeeFIM data were available from the 50 inpatients.
rs2 reflects the between-subjects’ effect, rt2 the amount of
trial effect or systematic error (between the two ratings), Reliability
and re2 the amount of residual variance (i.e. random error In two children, item 15 (reach across the midline –
variance).16 For interpretation, the following benchmarks subscore dynamic reaching) was not video-recorded. For
were used: very high reliability (ICC >0.90); high reliability these children, the reliability analyses of the dynamic
(ICC 0.70–0.89); moderate validity (ICC 0.50–0.69); low reaching and total TCMS score were based on items 13
reliability (ICC 0.26–0.49); poor reliability (ICC <0.25).17 and 14. All other datasets were complete. Rater A per-
To quantify the absolute reliability, the standard error of formed the assessment in 54 children, and rater B in 36
measurement (SEM) and the SDD were calculated, with children.
the following formulae: SEM=√(rt2+re2) and The relative intra- and interrater reliability were very
18
SDD=1.969√29SEM. We also expressed the SDD as a high, with ICC >0.90 for the total and all subscores
percentage of the maximum score. Additionally, for visual (Table II). The absolute measurement errors of the total
interpretation, we provided Bland–Altman plots.19 With and subscores were calculated. The SDD of the total score
the Blind–Altman plots we show the bias (i.e. the mean dif- of the TCMS remained lower than 10% of the maximal pos-
ference between the two ratings) and the limits of agree- sible score (Table II). The Bland–Altman plots represented
ment (biastwo times the standard deviation [SD]). The the bias and the limits of agreement (95% CI) both for the
upper limit of agreement and the lower limit of agreement intrarater (Fig. 1a) and interrater (Fig. 1b) reliability. The
mark the 95% CI. As most TCMS studies did not evaluate 95% CIs were well in agreement with the SDD values.
psychometric properties of the TCMS in children younger For the intrarater reliability, children younger than
than 8 years, we performed separate analyses for this group. 8 years of age had a bias of 0.67 (95% CI 4.37 to 6.09),
To estimate whether TCMS total and subscores could while children aged 8 years and older showed a bias of
distinguish between children who were independent in 0.55 (95% CI 3.53 to 4.63). For the interrater reliability,
mobility or self-care, receiver operating characteristics children younger than 8 years had a bias of 0.83 (95%
were performed. For each measure, corresponding cut-off CI 6.54 to 4.88), while children aged 8 years and older
levels were calculated, based on the Youden Index (this is had a bias of 0.16, (95% CI 5.48 to 5.16). A separate

Table I: Number of participants apportioned by diagnosis and disability level according to the Gross Motor Function Classification System (GMFCS)

GMFCS Spastic cerebral Ataxic and dyskinetic Acquired brain Others Total
Analysis level palsy (n) cerebral palsy (n) injuries (n) (n)a (n)

Reliability I 20 6 13 2 41
II 13 6 3 1 23
III 12 3 0 2 17
IV 4 3 1 1 9
Total 49 18 17 6 90
Clinical I 5 1 10 1 17
relevance II 6 3 4 1 14
III 8 1 0 2 11
IV 4 2 1 1 8
Total 23 7 15 5 50
a
Includes four children with myelomeningocele/hydrocephalus, one child with Guillain–Barre  syndrome, and one child with hereditary
paraplegia (this child was an outpatient and was only included in the reliability analysis).

708 Developmental Medicine & Child Neurology 2017, 59: 706–712

 14698749, 2017, 7, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/dmcn.13425 by Universidad Nacional Mayor De San Marcos, Wiley Online Library on [26/04/2023]. 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
Table II: Intraclass correlation coefficients (ICCs) for reliability with 95% confidence intervals (CIs), standard error of measurement (SEM) and the
smallest detectable difference (SDD)

Reliability TCMS ICC (95% CI) SEM SDD SDD/total (%)

Intrarater Static sitting balance 0.96 (0.94–0.97)a 0.61 1.70 9

Selective movement control 0.96 (0.94–0.98)a 1.43 3.97 14
Dynamic reaching 0.95 (0.93–0.97)a 0.65 1.81 18
Total score 0.98 (0.97–0.99)a 1.58 4.39 8
Interrater Static sitting balance 0.96 (0.94–0.97)a 1.06 2.94 15
Selective movement control 0.94 (0.91–0.96)a 1.73 4.78 17
Dynamic reaching 0.96 (0.94–0.97)a 0.66 1.84 18
Total score 0.98 (0.97–0.99)a 1.90 5.27 9
a
p<0.001. TCMS, Trunk Control Measurement Scale.

(a) Intrarater reliability: TCMS total score (n=90) (b) Interrater reliability: TCMS total score (n=90)
6 6
Difference (1–2 scoring)

5 5 Upper LOA:

Difference (rater A–B)

Upper LOA:
4 4 5.10
4.85 3
3 2
2 1
1 Bias: 0 Bias:
0 0.57 –1 –0.31
–1 –2
–2 –3
Lower LOA: –4
–3 –5
–4 –3.71 Lower LOA:
–6
–5 –7 –5.77
0 5 10 15 20 25 30 35 40 45 50 55 58 0 5 10 15 20 25 30 35 40 45 50 55 58
Mean of first and second scoring Mean of rater A and B
<8y ≥8y

Figure 1: Intra- and interrater reliability of the Trunk Control Measurement Scale (TCMS) total score. (a) Intrarater reliability and (b) interrater reliability
expressed by Bland–Altman plots, with the bias and the limits of agreement. LOA, limit of agreement. [Colour figure can be viewed at wileyonlinelibrary.com]

Bland–Altman plot of the children with ataxic and dyski- sitting, dynamic reaching, and selective movement control
netic CP, acquired brain injury, and other neurological dis- respectively. Similar values were obtained to discriminate
orders (n=41) is included in Figure S1 (online supporting between children with dependent versus independent
information). All results were in line with those from the mobility: 82.5%, 55.0%, and 41.1% respectively. Eight
whole group. children were, according to the WeeFIM self-care domain,
dependent, despite the fact that their total TCMS score
Discriminative validity exceeded the cut-off level of 30.5 points. In the mobility
The mean age of the 50 children collected by the WeeFIM part, six children (two with acquired brain injury, one with
was 11 years and 5 months (range 5y –18y 11mo). These myelomeningocele, two with spastic CP, and one with
children showed a mean TCMS total score of 32 (SD 16), dyskinetic CP), who were rated as dependent, even if their
with a minimum score of 1 and a maximum score of 57. TCMS total score exceeded the cut-off value of 30.5
Results of the subscores were as follows: mean static sitting points. On the contrary, three children (self-care) and two
balance 14 (SD 5), range 1 to 20; mean selective movement children (mobility) were rated as independent, despite the
control 12 (SD 8), range 0 to 27; mean dynamic reaching 6 fact that their TCMS total score was below the cut-off
(SD 3), range 0 to 10 points. The mean WeeFIM score value. These were children diagnosed with spastic CP (two
amounted to 95 (SD 35) and ranged from 27 to 126. All children) and one child with myelomeningocele out of the
receiver operating characteristics showed a significant AUC group of children younger than 8 years of age (in self-care
(Fig. 2). The TCMS total score could discriminate with and mobility).
best combined sensitivity and specificity between dependent
and independent children at a cut-off value of 30.5 (i.e. DISCUSSION
52.6% from the maximal possible TCMS score), both for The aim of this study was to determine the relative and
self-care and mobility WeeFIM subscores (see also Fig. 2). absolute reliability and the discriminative validity of the
The relative cut-off values that could discriminate best German TCMS version in children aged 5 to 19 years
between children with dependent versus independent self- with neuromotor disorders. The ICCs indicated a very
care amounted to 77.5%, 55.0%, and 30.4% for static high intra- and interrater reliability. The results were in

Reliability of the Trunk Control Measurement Scale Petra Marsico et al. 709
 14698749, 2017, 7, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/dmcn.13425 by Universidad Nacional Mayor De San Marcos, Wiley Online Library on [26/04/2023]. 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
(a) WeeFIM self-care and TCMS scores

Sensitivity: 0.86, specificity: 0.72 Sensitivity: 0.81, specificity: 0.69 Sensitivity: 0.95, specificity: 0.69 Sensitivity: 0.81, specificity: 0.83
AUC: 0.83 (p<0.001) AUC: 0.81 (p<0.001) AUC: 0.81 (p<0.001) AUC: 0.89 (p<0.001)
60 20 30 10
55 18 27 9

Selective movement control

50 16 8
24

Static sitting balance

45

Dynamic reaching
14 15.5 21 7
40
Total TCMS

35 12 18 6
5.5
30 30.5 10 15 5
25 8 12 4
20 6 9 3
15 8.5
4 6 2
10
5 2 3 1
0 0 0 0
Dependent Independent Dependent Independent Dependent Independent Dependent Independent
WeeFIM self-care WeeFIM self-care WeeFIM self-care WeeFIM self-care

(b) WeeFIM mobility and TCMS scores

Sensitivity: 0.91, specificity: 0.79 Sensitivity: 0.86, specificity: 0.86 Sensitivity: 0.82, specificity: 0.79 Sensitivity: 0.77, specificity: 0.82
AUC: 0.88 (p<0.001) AUC: 0.88 (p<0.001) AUC: 0.87 (p<0.001) AUC: 0.86 (p<0.001)
60 20 30 10
55 18 27 9

Selective movement control

50 16
Static sitting balance

16.5 24 8

Dynamic reaching
45
14 21 7
Total TCMS

40
35 12 18 6
30 30.5 10 15 5 5.5
25 8 12 4
20 6 9 11.5 3
15
4 6 2
10
5 2 3 1
0 0 0 0
Dependent Independent Dependent Independent Dependent Independent Dependent Independent
WeeFIM mobility WeeFIM mobility WeeFIM mobility WeeFIM mobility
<8y ≥8y

Figure 2: Discriminative validity. Cut-off values, sensitivity, specificity, and area under the curve (AUC) values of the Trunk Control Measurement Scale
(TCMS) scores (total score, static sitting balance, selective movement control, and dynamic reaching scores) to distinguish between children who are,
according to the Functional Independence Measure for children (WeeFIM), dependent versus independent for (a) the self-care domain and (b) mobility.
[Colour figure can be viewed at wileyonlinelibrary.com]

line with those of previous studies.7,11 More importantly, we scored the TCMS using video recordings, we did not
the small SEM and SDD values indicate a good absolute account for the natural variability regarding performance,
agreement. As we also presented the SEM and SDD values motivation, mood, and compliance, as this would be con-
of the three subscores, we can better interpret the results ducted in a test–retest design. While these factors could
of a recently published study.22 In the study by Pham reduce the reliability of the TCMS, especially in younger
et al.,22 the change scores of the TCMS after were investi- children, our results are still in line with those of Heyrman
gated 3 weeks of intensive physical training. All changes et al.,7 who used a repeated measurement procedure with a
lied under the SDD of the subscores, and therefore they mean time of 10 days in between the assessments. Accord-
could be a result of measurement error and not a result of ing to our results, the TCMS score should increase by at
a real improvement in trunk control. As the measurement least six points or 10% of its total score to account for the
errors were smaller when tested by the same rater, we measurement error and be considered a true change. In
recommend that if the TCMS is used to picture clinical future studies the responsiveness and minimal clinically
progress, the same rater should assess the child. important difference should be determined. This additional
The TCMS was reliable in children with highly information would help therapists to better interpret the
impaired trunk control, as well as in children with slight TCMS test results.
impairments. Even in children younger than 8 years, the The TCMS was able to differentiate well between chil-
test showed reliable results (see Fig. 1, distribution of dren who are independent in self-care and mobility (mea-
mean scores). The TCMS could be applied reliably to chil- sured with the WeeFIM) and those who are not, indicating
dren with neurological diagnosis other than spastic CP in a good discriminative validity. The excellent AUCs were
our study (Fig. S1). The children with dyskinetic CP had significant, indicating that the discrimination of the TCMS
more difficulty reaching higher points in the selective was above chance. This was the case for the TCMS total
trunk control part, as they show many involuntary move- score and for all three subscores. Interestingly, the relative
ments and the selectivity was markedly impaired. Because cut-off values of the static sitting balance, dynamic

710 Developmental Medicine & Child Neurology 2017, 59: 706–712

 14698749, 2017, 7, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/dmcn.13425 by Universidad Nacional Mayor De San Marcos, Wiley Online Library on [26/04/2023]. 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
reaching, and selective movement control subscores The cut-off values showed a high sensitivity and specificity,
decreased consecutively. On the one hand, this finding which indicates that certain TCMS subscores can differen-
could be influenced by the scaling of the subscores, which tiate well between children with different levels of self-
appears roughest for the static sitting balance subscore and dependence, but we did not integrate other factors that
finest for the selective movement control subscore. On the could contribute to daily activities such as cognition, global
other hand, the highest ceiling effect occurred in the static motor function, or muscle strength.
sitting balance subscore, while no child achieved the maxi-
mal selective movement control subscore. What we experi- CONCLUSION
enced in practice was that the items of selective movement The reliability of the TCMS was high in this group of
control were more difficult to perform than items of the children aged 5 to 19 years with neuromotor impairments.
other parts. Because the WeeFIM does not consider qual- When using the TCMS total score as an outcome mea-
ity of movement, we assume that even if children have sure, the change should exceed six points to be higher than
moderately impaired selective movement control of the the measurement error. With the cut-off values of the
trunk, these children could still be independent in self-care TCMS we can discriminate between the children who are
and mobility, as measured with the WeeFIM. We suggest independent or dependent in self-care and mobility. The
that clinicians should be careful in using these cut-off val- relative cut-off values were different between the TCMS
ues because they rely on various methodological decisions, subscores, which might reflect the differences in scaling of
such as the dichotomization of the WeeFIM in ‘dependent’ these subscores.
and ‘independent’. Changing the definitions or using
another measure for dichotomization would likely result in A CK N O W L E D G E M E N T S
different TCMS cut-off values, and this should be consid- The authors wish to thank the children and their parents for their
ered when interpretation our results. participation in this study; also highly appreciated is the support
Typically developing children can reach functional inde- of the physiotherapists and their help in the organization of the
pendence with the WeeFIM items we included in this measurements, Nicole Iten for collecting the WeeFIM data, and
study at the age of about 5 years; however, the majority of Rob Labruyere for the support in writing the manuscript. This
children younger than 8 years of age in our study were work was sponsored, in part, by a grant from the M€axi-Founda-
rated as dependent. This could be due to the sensorimotor tion. Julia Balzer was paid by grants from PhysioSwiss, Switzer-
impairments, which might have contributed to a reduction land; Physiotherapy Science Foundation, Switzerland; and Swiss
or delay in self-dependence in these children. National Science Foundation (Project 32003B_156646), Switzer-
This study has some methodological considerations. For land.
the reliability analysis, we re-analysed a previously assessed
TCMS using the video recordings. Therefore, our reliabil- SUPPORTING INFORMATION
ity analyses do not include variability caused by a second The following additional material may be found online:
performance of the child or adolescent. Concerning the Figure S1: Intrarater and interrater reliability of the Trunk
discriminative validity analysis, this study does not investi- Control Measurement Scale (TCMS) total score for the group:
gate causality between trunk control and self-dependence. ataxic and dyskinetic CP, ABI, and others (n=41).

REFERENCES
1. Curtis DJ, Butler P, Saavedra S, et al. The central role 6. Ba~
nas BB, Gorgon EJ. Clinimetric properties of sitting 10. Terwee CB, Mokkink LB, Knol DL, Ostelo RW, Bou-
of trunk control in the gross motor function of children balance measures for children with cerebral palsy: a sys- ter LM, de Vet HC. Rating the methodological quality
with cerebral palsy: a retrospective cross-sectional study. tematic review. Phys Occup Ther Pediatr 2014; 34: 313– in systematic reviews of studies on measurement proper-
Dev Med Child Neurol 2014; 57: 351–7. 34. ties: a scoring system for the COSMIN checklist. Qual
2. Saether R, Helbostad JL, Adde L, Braendvik S, 7. Heyrman L, Molenaers G, Desloovere K, et al. A clini- Life Res 2012; 21: 651–7.
Lydersen S, Vik T. The relationship between trunk con- cal tool to measure trunk control in children with cere- 11. Jeon JY, Shin WS. Reliability and validity of the Korean
trol in sitting and during gait in children and adoles- bral palsy: the Trunk Control Measurement Scale. Res version of the Trunk Control Measurement Scale
cents with cerebral palsy. Dev Med Child Neurol 2015; Dev Disabil 2011; 32: 2624–35. (TCMS-K) for children with cerebral palsy. Res Dev Dis-
57: 344–50. 8. Msall ME, DiGaudio K, Rogers BT, et al. The abil 2014; 35: 581–90.
3. Brogren Carlberg E, L€
owing K. Does goal setting in Functional Independence Measure for Children 12. Mitteregger E, Marsico P, Balzer J, van Hedel HJA.
activity-focused interventions for children with cerebral (WeeFIM). Conceptual basis and pilot use in children with Translation and construct validity of the Trunk Control
palsy influence treatment outcome? Dev Med Child Neu- developmental disabilities. Clin Pediatr 1994; 33: 421–30. Measurement Scale in children with brain lesions. Res
rol 2013; 55: 47–54. 9. Ziviani J, Ottenbacher KJ, Shephard K, Foremen S, Dev Disabil 2015; 45–46: 343–52.
4. Heyrman L, Desloovere K, Molenaers G, et al. Clinical Astbury W, Ireland P. Concurrent validity of the 13. Russell DJ, Avery LM, Rosenbaum PL, Raina PS, Walter
characteristics of impaired trunk control in children with Functional Independence Measure for Children (Wee- SD, Palisano RJ. Improved scaling of the gross motor func-
spastic cerebral palsy. Res Dev Disabil 2013; 34: 327–34. FIMTM) and the Pediatric Evaluation of Disabilities tion measure for children with cerebral palsy: evidence of
5. van der Heide JC, Hadders-Algra M. Postural muscle Inventory in children with development disabilities and reliability and validity. Phys Ther 2000; 80: 873–5.
dyscoordination in children with cerebral palsy. Neural acquired brain injuries. Phys Occup Ther Pediatr 2001; 14. Palisano R, Rosenbaum P, Walter S, Russel D, Wood
Plast 2005; 12: 197–204. 21: 91–101. E, Galuppi B. Development and reliability of a

Reliability of the Trunk Control Measurement Scale Petra Marsico et al. 711
 14698749, 2017, 7, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/dmcn.13425 by Universidad Nacional Mayor De San Marcos, Wiley Online Library on [26/04/2023]. 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
system to classify gross motor function in children 17. Katz JN, Larson MG, Phillips CB, Fossel AH, Liang MH. 20. Youden WJ. Index for rating diagnostic tests. Cancer
with cerebral palsy. Dev Med Child Neurol 1997; 39: Comparative measurement sensitivity of short and longer 1950; 3: 32–5.
214–33. health status instruments. Med Care 1992; 30: 917–25. 21. Hosmer D, Lemeshow S, Sturdivant R. Applied Logistic
15. Shrout P, Fleiss J. Intraclass correlations: uses in 18. De Vet HCW, Terwee CB, Knol DL, Bouter LM. Regression. 3rd edn. Hoboken, NJ: Wiley, 2013.
assessing rater reliability. Psyschological Bull 1979; 86: When to use agreement versus reliability measures. J 22. Pham HP, Eidem A, Hansen G, et al. Validity and
420–8. Clin Epidemiol 2006; 59: 1033–9. responsiveness of the Trunk Impairment Scale and
16. Rousson V, Gasser T, Seifert B. Assessing intrarater, 19. Bland JM, Altman DG. Comparing methods of mea- Trunk Control Measurement Scale in young individuals
interrater and test-retest reliability of continuous mea- surement: why plotting difference against standard with cerebral palsy. Phys Occup Ther Pediatr 2016; 36:
surements. Stat Med 2002; 21: 3431–46. method is misleading. Lancet 1995; 346: 1085–7. 440–52.

712 Developmental Medicine & Child Neurology 2017, 59: 706–712