1. No category

PDF hosted at the Radboud Repository of the Radboud University
Nijmegen
This full text is a publisher's version.
For additional information about this publication click this link.
http://hdl.handle.net/2066/92729
Please be advised that this information was generated on 2014-10-15 and may be subject to
change.
Longitudinal
motor performance in
very preterm infants
Anjo Janssen
Financial support for the printing of this thesis has been generously provided
by: The Royal Dutch Society for Physical Therapy (KNGF) and The Dutch Society
for Pediatric Physical Therapy (NVFK).
Design and lay-out: In Zicht Grafisch Ontwerp, www.promotie-inzicht.nl
Printed: Ipskamp Drukkers, Enschede, www.ipskampdrukkers.nl
ISBN 978-90-9026607-7
© A.J.W.M. Janssen, Molenhoek 2012
All rights reserved. No part of this thesis may be reproduced or transmitted in any form by any means,
without prior permission of the copyright owner. The copyright of the articles that have been published
has been transferred to the respective journals.
Longitudinal motor performance
in very preterm infants
Een wetenschappelijke proeve op het gebied van
de Medische Wetenschappen
Proefschrift
ter verkrijging van de graad van doctor
aan de Radboud Universiteit Nijmegen
op gezag van de rector magnificus prof. mr. S.C.J.J. Kortmann,
volgens besluit van het college van decanen
in het openbaar te verdedigen
op woensdag 29 februari 2012
om 15.30 uur precies
door
Anna Josepha Wilhelmina Maria Janssen
geboren op 31 maart 1960
te Nijmegen
Promotoren
Prof. dr. MWG Nijhuis-van der Sanden
Prof. dr. LAA Kollée
Prof. dr. RAB Oostendorp
Manuscriptcommissie
Prof. dr. R de Groot, voorzitter
Prof. dr. ALM Verbeek
Prof. dr. MJ Jongmans (UMC Utrecht)
“It is not the strongest of the species that survives, nor the most intelligent, but
the one most responsive to change.”
Charles Darwin (English Naturalist. 1809-1882)
Contents
List of abbreviations
8
Chapter 1
General introduction
11
Part I
Chapter 2Influence of behavior and risk factors on motor performance in preterm infants at age 2 to 3 years
Dev Med Child Neurol. 2008; 50: 926-931
31
Chapter 3Unstable longitudinal motor performance in preterm infants from 6 to 24 months on the Bayley Scales of
Infant Development – second edition
Res Dev Disabil. 2011; 32: 1902-1906
49
Part II
Chapter 4A model to predict motor performance in preterm infants at 5 years
Early Hum Dev. 2009; 85: 599-604
69
Chapter 5
Comparing the predictive validity of two infant motor tests for motor performance at 5 years in preterm infants
Submitted for publication
89
Part III
Chapter 6
Development of a movement quality measurement tool for children
Accepted Phys Ther.
Part IV
113
Chapter 7Summary
157
Chapter 8General discussion
167
Samenvatting in het Nederlands (Summery in Dutch)177
Dankwoord (Acknowledgements)187
Curriculum Vitae190
List of publications191
List of abbreviations
List of abbreviations
A&C
AGA
AIMS
BAL
BSID-II
BSID-II-BRS
BSID-II-MS
BSITD-III
CA
CAMPB
CI
CLD
CP
DOR
ELBW
EPT
GA
GM
GMFCS
GMPM
ICF
ICF-CY
aiming and catching
appropriate for gestational age
Alberta infant motor scale
balance
Bayley scales of infant development, second edition
Bayley scales of infant development, second edition, behavior rating scale
Bayley scales of infant development, second edition, motor scale
Bayley scales of infant and toddler development, third edition
corrected age
combined assessment of motor performance and behavior
confidence interval
chronic lung disease
cerebral palsy
diagnostic odds ratio
extremely low birth weight
extremely preterm
gestational age
general movements
gross motor function classification system
gross motor performance measure
International Classification of Functioning, Disability and Health
International Classification of Functioning, Disability and Health for
Children and Youth
IMP
infant motor profile
IQR
Interquartile range
IVH
intraventricular hemorrhage
MABC
movement assessment battery for children
MABC-2-NL movement assessment battery for children, second edition, Dutch version
MD
manual dexterity
MMT
Maastricht’s motor test
MND
minor neurological dysfunction
mo
months
MQ
movement quality
n
number
NBI
neuromotor behavioral inventory
NEC
necrotizing enterocolitis
NGT
nominal group technique
NICU
neonatal intensive care unit
NPV
negative predictive value
ns
not significant
OMQ
observable movement quality
OR
odds ratio
8
List of abbreviations
PDI
PGMQ
PPT
PPV
PVL
QUEST
RDS
ROP
rp
SD
SES
SGA
SOMP
TIME
TTS
VLBW
VPT
wk
ZNA
psychomotor developmental index
preschooler gross motor quality scale
pediatric physical therapist
positive predictive value
periventricular leukomalacia
quality of upper extremity skills test
respiratory distress syndrome
retinopathy of prematurity
spearman correlation coefficient
standard deviation
socioeconomic status
small for gestational age
structured observation of motor performance
toddler and infant motor evaluation
total test score
very low birth weight
very preterm
weeks
Zürich neuromotor assessment
9
Chapter 1
General introduction
This thesis describes studies concerning longitudinal motor performance
in very preterm (VPT) infants in the context of the follow-up.
General introduction
Preterm birth
Preterm birth is defined as childbirth before 37 completed weeks or 259 days of
gestation as measured from the first day of the last menstrual period.1 The World
Health Organization estimated that in 2005, around the world 9.6% of all births
were preterm. This would result in a total of about 12.9 million preterm births
worldwide. The lowest rate was found in Europe, where 6.2% of the births were
preterm. 2 In 2008, in the Netherlands 7.7% of all births, 13.649 infants, were born
before 37 weeks.3
The lower the gestational age (GA) of the infant, the less the organs are adapted
for the challenges outside the womb. As a consequence, after preterm birth,
hypothermia because of poor temperature regulation, (severe) respiratory
problems, compromised blood circulation and oxygenation of the brain and
other vital organs, infections and feeding problems can occur. Childbirth before
32 weeks of gestation often requires intensive care treatment to survive.
Specialized care in neonatal intensive care units (NICUs), becoming available in
the developed world from the early seventies, resulted in the survival of an
increasing number of critically ill preterm and even extremely preterm infants.4
In the early follow-up studies, cohorts of preterm infants often were defined by
birth weight, since gestational age was unknown in many cases. Low birth
weight infants are defined as less than 2500 grams birth weight and they are
subdivided into:
· Very Low Birth Weight (VLBW): birth weight <1500 g
· Extremely Low Birth Weight (ELBW): birth weight <1000 g
At present, preterm infants usually are defined by gestational age. Preterm
infants are defined as born before 37 weeks of gestation and they are subdivided
into:
· very preterm infants (VPT): born before 32 weeks of gestation
· extremely preterm (EPT): born before 28 weeks of gestation
Combining gestational age and birth weight allows subdividing the group of
preterm infants into small for gestational age (SGA) and appropriate for
gestational age (AGA) infants. SGA is defined as birth weight at least two
standard deviations below the mean for the gestational age.5
13
1
Chapter 1
In follow-up studies both categories VLBW or VPT infants are included.6;7
However, inclusion based on birth weight increases the percentage of SGA
infants. These infants are more at risk for developmental impairments and
inclusion based on VLBW in study cohorts might bias study outcomes.8;9
In the studies of this thesis only infants born before 32 weeks of gestation
were included. In 2008, in the Netherlands 2637 infants (1.5%) were born
VPT and 1674 infants were admitted to the neonatal intensive care units
(NICUs). 2 Approximately 150 were treated at the NICU from the Radboud
University Nijmegen Medical Centre, Nijmegen, The Netherlands.
Follow-up
In the 1980s and 1990s there was concern about the potentially increasing
number of impaired NICU survivors. Neurodevelopmental assessments were
used as predictors of long-term outcome.10-12 However, assessments at term age
turned out to be a weak predictor of long term outcome and follow-up programs
for NICU survivors were established. In 1993, the Dutch Ministry of Health
recommended the NICUs to establish follow-up programs for the developmental
outcome of NICU survivors.13 In 2000, this was confirmed again by the ministry.14
The National Neonatal Follow-up network established a protocol recommending
standardized multidisciplinary assessments in preterm infants at the corrected
age (CA; chronological age reduced by the number of days born before 40
weeks of gestation) of 6, 12 or 18 and 24 months and at a chronological age of
5 and 8 years.15
Until 1995, follow-up of NICU survivors in Nijmegen was performed by the
neonatologist only. The pediatric physical therapist (PPT) was present at the
outpatient clinic and was consulted during follow-up visits if considered
indicated by the neonatologist. Also, there were no strict inclusion criteria for
follow-up and the PPT assessed the infants suspected of motor impairments.
Therefore, systematic scientific evaluation of motor performance outcome was
impossible. An extensive standardized multidisciplinary neonatal follow-up
program was introduced in the Radboud University Medical Centre Nijmegen in
1995. It aimed to evaluate neonatal care by analysis of the long term outcome
14
General introduction
and to intervene in case of developmental delays in individual infants. The multi­
disciplinary follow-up team consisted of a neonatologist, a child psychologist, a
speech therapist and a PPT. The infants were assessed in a standardized
protocol, during consecutive visits to the outpatient clinic. Included were all
infants born before 32 weeks gestation and/or with a birth weight below 1500
grams, treated in the NICU at Nijmegen. Initially, the infants were invited for
follow-up at 24 months CA and the chronological age of 5 years. Since 2003,
infants were also invited for follow-up at 6 and 12 months CA. Follow-up
evaluation at 8 years of age was never introduced in Nijmegen because of
financial restrictions.
Initial follow-up studies reported mainly on survival rates. However, with
improved survival, the focus of the reports shifted to morbidity and to the quality
of the child’s development. Follow-up studies on specific sub-populations such
as those being SGA, or having chronic lung disease (CLD) or intraventricular
hemorrhage (IVH) 16 were published more frequently than follow-up studies that
included all preterm infants. Moreover, specific neonatal interventions such as
oxygen administration, surfactant treatment and indomethacin treatment were
evaluated for their long-term consequences.4 From these studies it became
clear that there may be a discrepancy between short-term neonatal outcome
and long-term post-discharge outcome.4 An example is the liberal use of oxygen
in the early years, resulting in improvement of the respiratory status of infants
but increased the number of infants with visual impairments due to retinopathy.1719 20
Similarly, the postnatal use of steroids improved the respiratory status and
facilitated weaning from the ventilator for infants with severe CLD. However, later
was shown that this treatment was associated with increased rates of cerebral
palsy (CP). 21;22
Later studies in VPT or VLBW infants focused on childhood motor, cognitive and
behavioral impairments compared to term infants. 23;24 Despite improving
survival, the rate of children with daily life impairments has remained relatively
constant, and up to 50% of them show motor, cognitive or behavioral
impairments. 23;24 Some impairments even increase over the years. 25;26 Five to
15% of the survivors are diagnosed with cerebral palsy, 27;28 but there is a
decreasing incidence and severity of cerebral palsy.9;29 Children without cerebral
palsy but with developmental problems show impairments which are complex,
often subtle, and usually affect various developmental aspects. Minor motor
15
1
Chapter 1
impairments, classified as Developmental Coordination Disorder (DSM-IV),
have also been found more often in VLBW children.30-34 These motor impairments
persist into adolescence and can affect school performance and self esteem. 35-37
Also in adulthood, VLBW children continue to exhibit higher rates of neurosensory
impairments, with lower academic scores and a lower high school graduation
rate compared to adults born with a normal birth weight.38;39 In the Netherlands,
at 19 years of age VPT infants showed a 3 fold higher rate of no school attendance
or unemployment compared to the general Dutch population.39 A recent study
on adults in the USA showed that the majority of VLBW survivors born during the
early years of neonatal intensive care report fairly normal social lives and quality
of life. If differences are noted they are usually due to neurosensory disabilities.7
The adult outcomes described belong mainly to preterm infants in the 1970s
and the early 1980s, a time when neonatal mortality was high and few extremely
preterm infants survived. Medical progress since 1990 has resulted in an
increase in survival of high risk children, even when born extremely preterm,
who previously would have died.40 Therefore long-term longitudinal multi­
disciplinary follow-up of infants is still needed.
The number of published follow-up studies on motor performance outcome
is less compared to studies on school and cognitive outcome. Moreover,
Nijmegen participated in a multicenter follow-up study to investigate if
pediatricians detect motor impairments at 5 years of age.3 This study pointed
out that even after standardized and thorough neurological assessment,
pediatricians may overlook functional motor impairments. Long term followup studies that do not include detailed standardized tests for multiple
domains may underestimate developmental impairments.7 The role of the
PPT in the follow-up in Nijmegen is to assess the child’s motor performance
and to indicate if PPT intervention is necessary.
Human movement
Human movement behavior is extremely complex and fundamental to
development.43 Each individual generates movements to meet the demands of
the task being performed within a specific environment. Shumway-Cook and
Woollacott tried to better understand the nature of human movement and
16
General introduction
defined the concept of movement as follows: ‘Movement emerges from the
interaction of the individual, the task, and the environment. Movement is both
task specific and constrained by the environment’. ‘The individual’s capacity to
meet interacting tasks and environmental demands determines that person’s
functional capability’.44 ‘Therefore the evaluation of a child’s movement is a multidimensional process involving various aspects’. Two important aspects of
movement that can be assessed are the increasing capacity to perform new
skills and the quality of the movements. The increasing capacity is termed motor
performance, and the control and coordination of movements is termed
movement quality.
Motor Performance in preterm infants
Motor performance is delayed in preterm infants. For example, VLBW infants sit
unsupported and walk later than term infants.45;46 This is confirmed in the new
reference values for the Alberta Infants Motor Scale (AIMS) in ‘healthy’ Dutch
preterm infants. The preterm infants scored significantly lower than the Canadian
reference values for term infants, even after full correction for preterm birth.47
In a meta-analysis including 41 studies and 9653 children it was found that VPT
and VLBW infants without congenital anomalies show a significant motor
impairment throughout childhood.6 Although motor outcomes show a catch-up
in the first years there is a tendency to a greater deficit with increasing age
during elementary school and early adolescence. Moreover, there is a some
evidence that early gross motor performance and cognitive function at later
school age are related.48 Unknown is, whether the children whose motor
performance catches-up at young ages deteriorate again in childhood.
In the studies of this thesis motor performance was assessed with the Bayley
Scales of Infant development, 2nd edition Motor Scale (BSID-II-MS) at 6, 12
and 24 months CA.1 At 6 and 12 months CA the Alberta Infant Motor Scale
(AIMS) was used as well.4 The Movement Assessment Battery for Children
(M-ABC) and from 2009 the latest 2nd edition, Dutch version (M-ABC-II-NL)
were recorded at the chronological age of 5 years.5;6
17
1
Chapter 1
Movement quality in preterm infants
Movement quality in preterm infants differs from that in term infants.53-55 In
preterm infants more extension in the limbs is seen after birth, while flexion tone
increases to term age, flexion stays less than in term infants. 56 Preterm infants
tend to have less extensor tone in the neck in the sitting posture.54 Kinematic
analysis of kicking movements of the legs showed that preterm infants under 30
weeks gestation without overt sonographic brain abnormalities have a higher
kicking frequency but exhibited a shorter flexion phase. They also showed lower
variation in the interlimb coordination pattern at 2 and 4 months CA,57 which was
associated with a delayed walking attainment. 58 During analysis of stepping on
a treadmill preterm infants had more extended leg postures during the stand
phase.59 Furthermore, kicking coordination in preterm infants with and without
white matter disorders tend to differ from each other.60
Kinematic analysis of reaching movements in low risk preterm infants at 4 and 6
months CA shows a temporary acceleration of the development in reaching,
which is related to a better functional outcome at 12 months CA. High-risk
preterm infants grow into dysfunctional reaching behavior at 6 months CA.61
There also is a tendency that in 5 years old children born very preterm, fast
aiming tasks with a stylus on a digitizer tablet require a longer movement time.
The time needed in 5 years old children born very preterm is even longer than
the time of 3 year old term peers. When aiming at a small target the 3 year old
children born very preterm tended to revert to more segmented velocity profiles,
indicating less fluent movements than term 3 year old children.62
In conclusion, movement quality in preterm infants differs from term infants and
evaluation of movement quality in these children is important. PPTs usually
describe movement quality in free text. Therefore, the different descriptions
hamper comparison between therapists.
18
General introduction
In the studies of this thesis we used a questionnaire of the child’s behavior
during the test (so-called ‘test-taking behavior’). The BSID-II-BRS consist for
6 to 24 months olds of three subscales: orientation/engagement, emotional
regulation and movement quality. This subscale movement quality consists of
only eight questions. Although several measurement tools including movement
quality are available,1;10-15 all have limitations. To allow comparable qualitative
assessment between PPTs, longitudinal evaluation and prediction of motor
performance, it is essential that a concise measurement tool becomes available.
Risk factors for delayed motor performance in
preterm infants
Motor performance in preterm infants is associated with known medical risk
factors such as birth weight and head circumference at birth 69, gestational
age70, periventricular leukomalacia (PVL)29, intraventricular hemorrhage (IVH)29,
respiratory distress syndrome (RDS)71, Chronic Lung Disease (CLD)72-75 and
necrotizing enterocolitis (NEC).76 However, they account for only a part of the
variance associated with long-term outcomes.77 Non-medical factors such as
social class, parental education, parenting style, parental mental health, family
structure, family functioning and the home environment are also associated
with developmental outcome of children born preterm.78;79
In the studies in this thesis registered risk factors for a delayed motor
performance were included. To gain insight in the complexity of motor
performance in children born preterm the International Classification
of Functioning, Disability and Health for Children and Youth (ICF-CY) was
used to compose a model of influencing factors. In Figure 1 the cursive bold
factors are the factors we included in the studies described in this thesis.
19
1
Chapter 1
Figure 1 The International Classification of Functioning, Disability and
Health for Children and Youth (ICF-CY) model composed for the
study of this thesis on motor performance in preterm infants.
Health condition
Preterm birth before 32 weeks
Body Functions &
structures
Activities
Participation
Attention functions
Bowel functions
Cognitive functions
Condition
Motor functions
Respiratory functions
Visual functions
Movement quality
Head circumference
Height
Weight
Motor skills
Motor Performance
Interaction parents/child
Kindergarten/school
Play, sports, hobbies
‘Test-taking’ behavior
Environmental factors
Personal factors
Coping parents
Expectancy parents
Learning environment
Maternal education
NICU treatment
Stress parents
SES parents
Endotracheal intubation
Caesarian section
Multiple births
Birth weight
Character
Coping child/stress child
Ethnic background
Gender
Gestational age
CLD
NEC
Neonatal convulsions
PVL
ROP
IVH I, II, III, IV
Apgar score
Sepsis
Meningitis/encephalitis
IRDS
Pneumonia
Hyperbilirubinemie
Abbreviations: CLD =chronic lung disease, IRDS= idiopathic respiratory distress syndrome,
IVH=intraventricular hemorrhage, NEC =necrotizing enterocolitis, NICU=neonatal intensive care
unit, PVL=periventricular leukomalacia, ROP=retinopathy of prematurity, SES =Socioeconomic
status
20
General introduction
Summarizing, preterm infants are at risk for a delayed motor performance and
different movement quality compared to full-term infants. Longitudinal cohort
studies which give insight in long-term motor performance and risk factors for
delayed motor performance are rarely published. Therefore, studies are needed
to evaluate the long-term predictive value of neonatal risk factors and post
discharge assessments.4 Studies investigating the prediction of motor performance
to improve the efficiency of the follow-up in preterm infants are not available. So,
if we could improve prediction this might benefit efficiency of the follow-up
program. Moreover, it is beneficial for both child and parents to effectively
intervene and reduce unnecessary interventions and follow-up evaluations.
Concluding, insight in longitudinal motor performance, risk factors, prediction
of motor performance and movement quality in preterm infants is important in
order to be able to start pediatric physical therapy when necessary, to prevent
both over- and under-referral and to evaluate the effect of PPT interventions in
future studies.
Aims and outline of the thesis
This thesis aims to determine motor performance in very preterm infants in the
follow-up and to investigate the influence of the child’s behavior during the test
(‘test-taking’ behavior) and risk factors on longitudinal motor performance. In
addition, to study the predictive value of motor performance assessment and to
develop a prediction model for longitudinal follow-up in children born preterm.
Finally, to develop a concise movement quality measurement tool for children
that can be used at different ages.
The two cohorts described in this thesis consisted of surviving infants born
between January 1996 and December 2001 and between March 2003 and
December 2006. We included infants with a GA of less than 32 weeks, who had
been admitted to the neonatal intensive care unit of the Radboud University
Nijmegen Medical Centre and risk factors related to motor outcome.
The following research questions are being studied in this thesis:
· What is the motor performance outcome and influence of behavior and risk
factors on longitudinal motor performance in preterm infants in follow-up?
21
1
Chapter 1
· Can we improve the efficiency of the follow-up for longitudinal motor performance
assessment by predicting normal and delayed motor performance taking risk-­
factors into account?
· Which tests or combinations of tests, at which assessment moment, predicts
motor performance best?
· Can we develop a concise measurement tool for evaluation of observable
movement quality?
Outline of the thesis
The thesis is divided in four parts, the first concerning motor performance
outcome, influence of risk factors and stability of assessments, the second
about prediction of motor performance for outcome at 5 years of age and in the
third focusing on assessment of observable movement quality. The fourth part
consist of the summary, general discussion, and curriculum vitae. Chapter 1
gives an overview and definitions of terminology used in this thesis. Chapter 2
presents the influence of risk factors on motor performance at 2 years of age in
the first cohort. This is important to understand which risk factors need to be
taken into account to start pediatric physical therapy treatment, even when
motor performance at that time is unsuspected for motor performance
impairments. In chapter 3 we explore longitudinal motor performance and
influence of risk factors in a multilevel analysis on motor trajectories from 6 to 24
months, in the second cohort. This to investigate the highly variable motor
performance in the first two years of life, and to understand more of the variation
in motor performance in preterm infants. Chapter 4 presents a model to predict
motor performance from 2 to 5 years of age, to improve the efficiency of follow-up
for motor performance in the first cohort. With the increasing health care costs
the follow-up should be as efficient as possible without reducing the quality of
care. Chapter 5 studies the prediction of motor performance from 6 months to
5 years of age in the second cohort. All measurements performed at 6, 12 and
24 months and 5 years were included to study which measurement at what
moment best predicts motor performance. In chapter 6 we developed a concise
movement quality measurement tool for children from the PPTs perspective.
This is a first step to explicit the way movement quality is judged by PPTs.
Chapter 7 summarizes the main findings of this thesis and presents suggestions
for future research and chapter 8 provides a general discussion.
22
General introduction
References
(1) World Health Organization. International classification of diseases and related health problems,
10th revision. Geneva: 1992.
(2) Beck S, Wojdyla D, Say L, Pilar Betran A, Merialdi M, Harris Requejo J et al. The worldwide
incidence of preterm birth: a WHO systematic review of maternal mortality and morbidity. Bull
World Health Organ 2009; 88: 1-19.
(3) The Netherlands Perinatal registry (PRN). Perinatal care in the Netherlands 2008. Utrecht: 2011.
(4) Vohr BR. How should we report early childhood outcomes of very low birth weight infants? Semin
Fetal Neonatal Med 2007; 12: 355-362.
(5) Lee PA, Chernausek SD, Hokken-Koelega AC, Czernichow P. International Small for Gestational
Age Advisory Board consensus development conference statement: management of short
children born small for gestational age, April 24-October 1, 2001. Pediatrics 2003; 111: 1253-1261.
(6) de Kieviet JF, Piek JP, arnoudse-Moens CS, Oosterlaan J. Motor development in very preterm and very
low-birth-weight children from birth to adolescence: a meta-analysis. JAMA 2009; 302: 2235-2242.
(7) Hack M. Adult outcomes of preterm children. J Dev Behav Pediatr 2009; 30: 460-470.
(8) Arnold CC, Kramer MS, Hobbs CA, McLean FH, Usher RH. Very low birth weight: a problematic
cohort for epidemiologic studies of very small or immature neonates. Am J Epidemiol 1991; 134:
604-613.
(9) Claas MJ, Bruinse HW, Koopman C, van H, I, Peelen LM, de Vries LS. Two-year neurodevelopmental outcome of preterm born children ≤ 750 g at birth. Arch Dis Child Fetal Neonatal Ed 2011; 96:
F169-F177.
(10) Amiel-Tison C. Neuromotor status. In: Taeusch HW, Yegman M.W., editors. Follow-up management
of the high-risk infant. Boston, MA: Little Brown & Company; 1987: 115-126.
(11) Dubowitz LM, Dubowitz V, Palmer PG, Miller G, Fawer CL, Levene MI. Correlation of neurologic
assessment in the preterm newborn infant with outcome at 1 year. J Pediatr 1984; 105: 452-456.
(12) Brazelton TB, Nugent J.T. The Neonatal Behavioral Assessmenst Scale (NBAS). Cambridge, MA:
MacKeith Press; 1995.
(13) Health Council ofthe Netherlands. Planningsdecision neonatal intensive care at birth. The Hague
1993.
(14) Health Council ofthe Netherlands. Intensive care at birth. Health Council of the Netherlands
publication no. 2000/08. The Hague: 2000 .
(15) Baar AL van, den Ouden AL, Kollée LAA. Ontwikkeling van kinderen met perinatale risicofactoren:
theoretische achtergrond, literatuurgegevens en implementatie in de praktijk. Tijdschrift voor Kindergeneeskunde 2000; 68: 210-216.
(16) Papile LA, Burstein J, Burstein R, Koffler H. Incidence and evolution of subependymal and intraventricular hemorrhage: a study of infants with birth weights less than 1,500 gm. J Pediatr 1978; 92:
529-534.
(17) Campbell K. Intensive oxygen therapy as a possible cause of retrolental fibroplasia; a clinical
approach. Med J Aust 1951; 2: 48-50.
(18) Flynn JT, Bancalari E, Snyder ES, Goldberg RN, Feuer W, Cassady J et al. A cohort study of
transcutaneous oxygen tension and the incidence and severity of retinopathy of prematurity. N
Engl J Med 1992; 326: 1050-1054.
(19) Tin W, Gupta S. Optimum oxygen therapy in preterm babies. Arch Dis Child Fetal Neonatal Ed
2007; 92: F143-F147.
(20) Tokuhiro Y, Yoshida T, Nakabayashi Y, Nakauchi S, Nakagawa Y, Kihara M et al. Reduced oxygen
protocol decreases the incidence of threshold retinopathy of prematurity in infants of <33 weeks
gestation. Pediatr Int 2009; 51: 804-806.
(21) Vohr BR, Coll CG, Lobato D, Yunis KA, O’Dea C, Oh W. Neurodevelopmental and medical status of
low-birthweight survivors of bronchopulmonary dysplasia at 10 to 12 years of age. Dev Med Child
Neurol 1991; 33: 690-697.
23
1
Chapter 1
(22) Yeh TF, Lin YJ, Lin HC, Huang CC, Hsieh WS, Lin CH et al. Outcomes at school age after postnatal
dexamethasone therapy for lung disease of prematurity. N Engl J Med 2004; 350: 1304-1313.
(23) Doyle LW. Neonatal intensive care at borderline viability--is it worth it? Early Hum Dev 2004; 80:
103-113.
(24) Bhutta AT, Cleves MA, Casey PH, Cradock MM, Anand KJ. Cognitive and behavioral outcomes of
school-aged children who were born preterm: a meta-analysis. JAMA 2002; 288: 728-737.
(25) Roberts G, Anderson PJ, Doyle LW. Neurosensory disabilities at school age in geographic cohorts
of extremely low birth weight children born between the 1970s and the 1990s. J Pediatr 2009; 154:
829-834.
(26) Roberts G, Anderson PJ, De LC, Doyle LW. Changes in neurodevelopmental outcome at age eight
in geographic cohorts of children born at 22-27 weeks’ gestational age during the 1990s. Arch Dis
Child Fetal Neonatal Ed 2010; 95: F90-F94.
(27) Tin W, Wariyar U, Hey E. Changing prognosis for babies of less than 28 weeks’ gestation in the
north of England between 1983 and 1994. Northern Neonatal Network. BMJ 1997; 314: 107-111.
(28) Vohr BR, Wright LL, Poole WK, McDonald SA. Neurodevelopmental outcomes of extremely low
birth weight infants <32 weeks’ gestation between 1993 and 1998. Pediatrics 2005; 116: 635-643.
(29) van Haastert I, Groenendaal F, Uiterwaal CS, Termote JU, van dH-J, Eijsermans MJ et al.
Decreasing incidence and severity of cerebral palsy in prematurely born children. J Pediatr 2011;
159: 86-91.
(30) Polatajko HJ, Macnab JJ, Anstett B, Malloy-Miller T, Murphy K, Noh S. A clinical trial of the processoriented treatment approach for children with developmental co-ordination disorder. Dev Med
Child Neurol 1995; 37: 310-319.
(31) Holsti L, Grunau RV, Whitfield MF. Developmental coordination disorder in extremely low birth
weight children at nine years. J Dev Behav Pediatr 2002; 23: 9-15.
(32) Marlow N, Roberts BL, Cooke RW. Laterality and prematurity. Arch Dis Child 1989; 64: 1713-1716.
(33) Goyen TA, Lui K. Developmental coordination disorder in “apparently normal” schoolchildren born
extremely preterm. Arch Dis Child 2009; 94: 298-302.
(34) Lingam R, Hunt L, Golding J, Jongmans M, Emond A. Prevalence of developmental coordination
disorder using the DSM-IV at 7 years of age: a UK population-based study. Pediatrics 2009; 123:
e693-e700.
(35) Powls A, Botting N, Cooke RW, Marlow N. Motor impairment in children 12 to 13 years old with a
birthweight of less than 1250 g. Arch Dis Child Fetal Neonatal Ed 1995; 73: F62-F66.
(36) Marlow N, Hennessy EM, Bracewell MA, Wolke D. Motor and executive function at 6 years of age
after extremely preterm birth. Pediatrics 2007; 120: 793-804.
(37) Wocadlo C, Rieger I. Motor impairment and low achievement in very preterm children at eight
years of age. Early Hum Dev 2008; 84: 769-776.
(38) Hack M, Flannery DJ, Schluchter M, Cartar L, Borawski E, Klein N. Outcomes in young adulthood
for very-low-birth-weight infants. N Engl J Med 2002; 346: 149-157.
(39) Hille ET, Weisglas-Kuperus N, van Goudoever JB, Jacobusse GW, Ens-Dokkum MH, de Groot L et
al. Functional outcomes and participation in young adulthood for very preterm and very low birth
weight infants: the Dutch Project on Preterm and Small for Gestational Age Infants at 19 years of
age. Pediatrics 2007; 120: e587-e595.
(40) de Kleine MJ, den Ouden AL, Kollee LA, Ilsen A, van Wassenaer AG, Brand R et al. Lower mortality
but higher neonatal morbidity over a decade in very preterm infants. Paediatr Perinat Epidemiol
2007; 21: 15-25.
(41) de Kleine MJ, Nijhuis-van der Sanden MW, Lya Den OA. Is paediatric assessment of motor development
of very preterm and low-birthweight children appropriate? Acta Paediatr 2006; 95: 1202-1208.
(42) de Kleine MJ, den Ouden AL, Kollee LA, Nijhuis-van der Sanden MW, Sondaar M, van
Kessel-Feddema BJ et al. Development and evaluation of a follow up assessment of preterm
infants at 5 years of age. Arch Dis Child 2003; 88: 870-875.
(43) Edelman GM. Neural Darwinism: selection and reentrant signaling in higher brain function. Neuron
1993; 10: 115-125.
24
General introduction
(44) Shumway-Cook A, Woollacott MH. Motor control: issues and theories. Motor Control; Translating
research into clinical practice, third edition. Philadelphia, Pennsylvania: Lippincott Williams &
Wilkins; 2007: 3-20.
(45) Pin TW, Darrer T, Eldridge B, Galea MP. Motor development from 4 to 8 months corrected age in
infants born at or less than 29 weeks’ gestation. Dev Med Child Neurol 2009; 51: 739-745.
(46) Jeng SF, Yau KIT, Liao HF, Chen LC, Chen PS. Prognostic factors for walking attainment in very
low-birthweight preterm infants. Early Human Development 2000; 59: 159-173.
(47) van Haastert I, de Vries LS, Helders PJ, Jongmans MJ. Early gross motor development of preterm
infants according to the Alberta Infant Motor Scale. J Pediatr 2006; 149: 617-622.
(48) Piek JP, Dawson L, Smith LM, Gasson N. The role of early fine and gross motor development on
later motor and cognitive ability. Hum Mov Sci 2008; 27: 668-681.
(49) Bayley N. Bayley Scales of Infant Development, second edition. San Antonio: The Psychological
Corporation; 1993.
(50) Piper MC, Pinnell LE, Darrah J, Maguire T, Byrne PJ. Construction and validation of the Alberta
Infant Motor Scale (AIMS). Can J Public Health 1992; 83 Suppl 2: S46-S50.
(51) Henderson SE, Sugden DA. Movement Assessment Battery for Children. London: Psychological
Corporation, Harcourt Brace Jonanovich; 1992.
(52) Henderson S.E., Sugden D.A., Bernett A.L. Movement Assessment Battery for Children-2 second
edition (Movement ABC-2). London: Pearson Education, Inc.; 2007.
(53) de GL, Hopkins B, Touwen B. Motor asymmetries in preterm infants at 18 weeks corrected age and
outcomes at 1 year. Early Hum Dev 1997; 48: 35-46.
(54) Mercuri E, Guzzetta A, Laroche S, Ricci D, van Haastert I, Simpson A et al. Neurologic examination
of preterm infants at term age: comparison with term infants. J Pediatr 2003; 142: 647-655.
(55) Piek JP. Is a quantitative approach useful in the comparison of spontaneous movements in fullterm
and preterm infants? Hum Mov Sci 2001; 20: 717-736.
(56) Amiel-Tison C, Grenier A. Normal Development during the first year of life: identification of
anomalies and use of the grid. Neurological assessment during the first year of life. New York
Oxford: Oxford university Press; 1986: 46-95.
(57) Jeng S, Chen L, Yau KT. Kinematic analysis of kicking movements in preterm infants with very low
birth weight and full-term infants. Phys Ther 2002; 82: 148-159.
(58) Jeng SF, Chen LC, Tsou KI, Chen WJ, Luo HJ. Relationship between spontaneous kicking and age
of walking attainment in preterm infants with very low birth weight and full-term infants. Phys Ther
2004; 84: 159-172.
(59) Davis DW, Thelen E, Keck J. Treadmill stepping in infants born prematurely. Early Hum Dev 1994;
39: 211-223.
(60) Fetters L, Chen YP, Jonsdottir J, Tronick EZ. Kicking coordination captures differences between
full-term and premature infants with white matter disorder. Hum Mov Sci 2004; 22: 729-748.
(61) Fallang B, Saugstad OD, Grogaard J, Hadders-Algra M. Kinematic quality of reaching movements
in preterm infants. Pediatric Research 2003; 53: 836-842.
(62) Sagnol C, Debillon T, Debu B. Assessment of motor control using kinematics analysis in preschool
children born very preterm. Dev Psychobiol 2007; 49: 421-432.
(63) Hemgren E, Persson K. A model for combined assessment of motor performance and behaviour
in 3-year-old children. Ups J Med Sci 1999; 104: 49-85.
(64) Sun SH, Zhu YC, Shih CL, Lin CH, Wu SK. Development and initial validation of the Preschooler
Gross Motor Quality Scale. Res Dev Disabil 2010; 31: 1187-1196.
(65) Boyce W, Gowland C, Rosenbaum P, Hardy S, Lane C, Plews N et al. Gross motor performance
measure manual. Kingston, ON: Queen’s University, School of Rehabilitation Therapy; 1998.
(66) De Matteo C, Law M, Russel D, Pollock N, Rosenbaum P, Walter S. QUEST: Quality of upper
extremity skills test. Hamilton ON: Mc Masters University, Neurodevelopmental Clinical Research
Unit; 1992.
(67) Prechtl HF, Einspieler C, Cioni G, Bos AF, Ferrari F, Sontheimer D. An early marker for neurological
deficits after perinatal brain lesions. Lancet 1997; 349: 1361-1363.
25
1
Chapter 1
(68) Kraus de Camargo O, Storck M, Bode H. Video-based documentation and rating system of the
motor behaviour of handicapped children treated with physiotherapy, a new outcome measure.
Pediatr Rehabil 1998; 2: 21-26.
(69) Marin Gabriel MA, Pallas Alonso CR, De La Cruz BJ, Caserio CS, Lopez MM, Moral PM et al. Age
of sitting unsupported and independent walking in very low birth weight preterm infants with
normal motor development at 2 years. Acta Paediatr 2009; 98: 1815-1821.
(70) Raz S, Debastos AK, Newman JB, Batton D. Extreme prematurity and neuropsychological
outcome in the preschool years. J Int Neuropsychol Soc 2010; 16: 169-179.
(71) Engle WA. Surfactant-replacement therapy for respiratory distress in the preterm and term
neonate. Pediatrics 2008; 121: 419-432.
(72) Katz-Salamon M, Gerner EM, Jonsson B, Lagercrantz H. Early motor and mental development in
very preterm infants with chronic lung disease. Arch Dis Child Fetal Neonatal Ed 2000; 83: F1-F6.
(73) Yeo CL, Chan C. Motor development of very low birthweight infants with chronic lung disease, a
comparative study. Ann Acad Med Singapore 2005; 34: 411-416.
(74) Bohm B, Katz-Salamon M. Cognitive development at 5.5 years of children with chronic lung
disease of prematurity. Arch Dis Child Fetal Neonatal Ed 2003; 88: F101-F105.
(75) Laughon M, O’Shea MT, Allred EN, Bose C, Kuban K, Van Marter LJ et al. Chronic lung disease and
developmental delay at 2 years of age in children born before 28 weeks’ gestation. Pediatrics
2009; 124: 637-648.
(76) Rees CM, Pierro A, Eaton S. Neurodevelopmental outcomes of neonates with medically and
surgically treated necrotizing enterocolitis. Arch Dis Child Fetal Neonatal Ed 2007; 92: F193-F198.
(77) Vohr BR, Wright LL, Dusick AM, Mele L, Verter J, Steichen JJ et al. Neurodevelopmental and
functional outcomes of extremely low birth weight infants in the National Institute of Child Health
and Human Development Neonatal Research Network, 1993-1994. Pediatrics 2000; 105:
1216-1226.
(78) Hogan DP, Park JM. Family factors and social support in the developmental outcomes of very
low-birth weight children. Clin Perinatol 2000; 27: 433-459.
(79) Msall ME, Park JJ. The spectrum of behavioral outcomes after extreme prematurity: regulatory,
attention, social, and adaptive dimensions. Semin Perinatol 2008; 32: 42-50.
26
General introduction
27
1
Part I
Chapter 2
Influence of behavior and
risk factors on motor performance in
preterm infants at age 2 to 3 years
A J W M Janssen1, MSc PT, M W G Nijhuis-van der Sanden2 PhD PT,
R P Akkermans 3 MSc, R A B Oostendorp2 PhD PT, L A A Kollée 4 PhD MD
Radboud University Nijmegen Medical Centre, Nijmegen, The Netherlands,
Department of Pediatric Physical Therapy
2
Radboud University Nijmegen Medical Centre, Nijmegen, The Netherlands,
Centre of Quality of Care Research, Research Centre of Allied Health Sciences
3 Radboud University Nijmegen Medical Centre, Nijmegen, The Netherlands,
Centre of Quality of Care Research, Information Processing and Statistical Support
4 Radboud University Nijmegen Medical Centre, Nijmegen, The Netherlands,
Department of Pediatrics
1
Dev Med Child Neurol. 2008; 50: 926-931
Chapter 2
Abstract
The aim of this cross-sectional study was to determine the influence of test-taking
behavior and risk factors for delayed motor performance in 437 (244 males, 193
females; ≤32 weeks of gestation) at the corrected age of 2 to 3 years (mean
29mo [SD 3.3]). Other mean sample demographics were: gestational age 29+5
weeks (1+5), range 25+0 -32+0; birth weight 1213.7g (331.7), range 468-2350; and
days in the neonatal intensive care unit 21.1 (21.3), range 1-165. Children (n=23)
with a severe disability were excluded. We assessed motor performance and
behavior with the Motor Scale and the Behavior Rating Scale (BRS) of the Bayley
Scales of Infant Development, 2nd edition (BSID-II). Risk factors were tested
against delayed motor performance as the dependent variable in binary logistic
regression analysis. Median score on the Motor Scale in terms of the BSID-II
Psychomotor Developmental Index (PDI) was 86. ‘Delayed’ motor performance
was observed in 46.5% of the children tested, and behavior was ‘not-optimal’ in
31.4%. The Motor Scale and BRS scores were significantly correlated (rs=0.62,
p<0.01). Risk factors for delayed motor performance were: neonatal convulsions
(odds ratio [OR] 4.5; 95% confidence interval [CI] 1.6–12.9), low maternal
educational level (OR 3.3; 95% CI 1.7–6.5), male sex (OR 2.8; 95% CI 1.8–4.3),
and chronic lung disease (OR 2.1; CI 1.1–4.1). We conclude that preterm infants
are at high risk of delayed motor performance and non-optimal test-taking
behavior.
32
Influence of behavior and risk factors on motor performance in preterm infants
Introduction
Multidisciplinary follow-up programs for preterm infants that include motor
performance assessment are recommended to determine long-term morbidity
and the need for intervention.1-4 The Dutch Neonatal Follow-up Study Group
recommends the inclusion of preterm infants with a gestational age of less than
32 weeks in such program. To date, three studies that have included this group
of infants have evaluated motor performance at the age of 2 to 3 years;5-8 however,
one used a screening test.5-7
Motor performance ‘emerges from the interactions between the individual, the
task and the environment’.9(p2) Similarly, the International Classification of
Functioning, Disability and Health (ICF) focuses on body functions and
structures, activities, and participation in interaction with personal and
environmental factors.10 This implies that not only should perinatal risk factors
for delayed motor performance be taken into account, but also the interaction
between the test setting and environmental factors. Moreover, test-taking
behavior (behavior displayed during testing) reflects how the child copes with
the differential task demands. The Bayley Scales of Infant Development, 2nd
edition (BSID-II), which became available in 1993, is the most widely used
assessment tool of motor performance in preterm follow-up studies.11 The
BSID-II also evaluates the test-taking behavior of the child, thus facilitating
interpretation of the test results.4
The aim of the present study was threefold: (1) to determine motor performance
in a 6-year cohort (1996–2001) of preterm infants without overt disabilities at a
corrected age of 2 to 3 years; (2) to determine whether and to what extent
test-taking behavior affects the motor score, and (3) to identify potential risk
factors for delayed motor performance.
Method
Motor assessments evaluated in this prospective, cross-sectional cohort study
constitute components of a standardized multidisciplinary follow-up assessment
program for preterm infants at the Radboud University Nijmegen Medical
Centre, in the Netherlands. We performed general physical pediatric and
neurological examinations to exclude children with visual impairments and
33
2
Chapter 2
cerebral palsy (CP). Trained pediatric physical therapists, blinded to the
children’s full medical histories, assessed the children.
The assessment protocol used is the standard in the Netherlands and is based
on a national consensus on neonatal follow-up. Results were documented
anonymously and approval by the medical ethical committee was not required.
Participants
The study cohort comprised all surviving infants born between January 1996
and December 2001 at a gestational age of ≤32 weeks and treated in the
Neonatal Intensive Care Unit. Children with known chromosomal disorders,
neuromuscular diseases, or CP and children unable to complete the test were
excluded. All children were tested once at the corrected age of 2 to 3 years.
Accompanying parents or caregivers remained present throughout the test
procedure.
Assessment of motor performance
Motor performance was assessed with the BSID-II Motor Scale and behavior
was recorded with the BSID-II Behavior Rating Scale (BRS).11 The content,
construct, and concurrent validity of the Motor Scale and BRS have been shown
to be comprehensive and appropriate.4;11;12 The predictive validity for motor
performance increases if children are tested at 2 years of age or older.11 As the
Dutch norm references for the BSID-II did not became available until 2003, and
the norms were not validated until 2005,13 we chose to use the US norm
references of the BSID-II.
The Motor Scale assesses both gross and fine motor skills, and items are arranged
in age groups to assess specific ages. It is a norm-referenced test giving scores
for the Psychomotor Developmental Index (PDI; mean of 100 [SD 15]; range
50–129). The scores are transformed into the following classification: scores
below 70 (–2SD) indicates ‘significantly delayed’ motor performance, scores
between 70 and 84 (–1SD) ‘mildly delayed’ motor performance, scores between
85 and 114 reflect performance that is ‘within normal limits’ and scores of 115
and above (>1SD) denote ‘accelerated’ motor performance. Because PDI scores
below 50 (–3SD) are not discriminative, we also calculated an age-equivalent
developmental motor-age in months using the table in the BSID-II manual.
34
Influence of behavior and risk factors on motor performance in preterm infants
The BRS assesses qualitative aspects of the child’s behavior during the test
(so-called ‘test-taking behavior’) and comprises 30 questions each scored on a
5-point scale. The first two questions focus on parental assessment of the
child’s typical behavior during the test and the adequacy of the test, and are not
included in the total score. The pediatric physical therapist completes the BRS
and calculates the total score after the Motor Scale assessment. For 2 to 3
year-olds, three BRS subscales are scored: (1) orientation/engagement, (2)
emotional regulation, and (3) motor quality. In the motor quality subscale the
appropriateness of the child’s movement control and coordination in the tested
fine and gross motor tasks are scored. Furthermore the presence or absence of
hypotonicity, hypertonicity, tremulousness, slow and delayed, and frenetic
movements are scored. Both the subscale scores and the total score are
transformed into centiles. Scores below the 10th centile reflect ‘non-optimal’
behavior, scores between the 11th and 25th centiles reflect ‘questionable’
behavior, and scores above the 26th centile reflect behavior ‘within normal
limits’.
Risk factors
The Dutch Neonatal Follow-up Study Group has developed a database to
document the risk factors for delayed outcome uniformly. Of the 46 registered
risk factors we chose 20 mentioned in the literature and on clinical experience,
that are expected to influence delayed motor performance, for inclusion in the
analyses: sepsis; meningitis/encephalitis/ventriculitis; intra-cranial hemorrhage;
post-hemorrhagic ventricle dilatation; ischemic cerebral abnormality; neonatal
convulsions; retinopathy of prematurity; idiopathic respiratory distress syndrome;
chronic lung disease (mechanical ventilation or oxygen after 36 weeks’ gestation);
respiratory problems (pneumonia/pneumothorax/atelectasis); hyperbilirubinemia;
necrotizing enterocolitis; male sex; gestational age; birth weight; multiple births;
Caesarean section; Apgar score; congenital malformation; endotracheal intubation
after delivery. A neonatologist recorded for each child the potential risk factors
at discharge from the hospital. Maternal educational level was added in three
categories: low (no/elementary school/vocational training); average (lower and
upper general secondary education/secondary vocational education); and high
(pre-university education/bachelor degree and higher).
35
2
Chapter 2
Statistical methods
PDI scores <50 were set at 49. The PDI and developmental motor-age scores
were tested for normality. We used the Mann–Whitney U test to reveal sex-related
differences in motor performance, the Kruskall–Wallis analysis of variance
(ANOVA) to evaluate differences between age groups, and the c2 test to compare
the Motor Scale and BRS classifications of our study sample with the BSID-II
reference group. Spearman’s rank correlation coefficient was computed to
determine the correlation between the PDI and BRS scores.
We first entered the selected factors in cross-tabulations and calculated c2,
second we performed univariate logistic regression, and third we checked for
colinearity. Finally, 14 factors with a significance level of <0.5 were entered in
the binary logistic stepwise regression procedure with the Motor Scale
classification as the dependent variable. The four categories were dichotomized
by combining accelerated and normal as ‘normal’ motor performance and
mildly and significantly delayed as ‘delayed’ motor performance. We applied a
stepwise selection procedure (backward and forward elimination (likelihood
ratio), p in <0.05 and p out >0.05). The final model included only factors with p
values of <0.05. Results are presented as odds ratios (OR) with 95% confidence
intervals (CI). Two-sided p values of <0.05 were considered to be statistically
significant. All analyses were performed using SPSS software (version 12.0).
Results
Cohort
The flow chart in Figure 1 provides an overview of the inclusion procedure. The
final 6-year cohort comprised 787 children. Of these, 123 were not invited to
participate because: (1) the test facilities were temporarily unavailable (n=13),
(2) the age inclusion criteria had been changed (n=99, all from the 1999 and
2001 cohorts), or (3) the children were known to have a disability and were
already under multidisciplinary treatment (n=11). In total, 664 children were
eligible for follow-up. Of these, 81 children could not be traced or had moved
abroad, the parents of 94 children declined to participate, and there were no
BSID-II data available for 20 children due to follow-up elsewhere.
Of the 466 infants that were actually tested, 29 had to be excluded from the
analysis (Fig. 1), of whom 12 were excluded on the clinical and neurological
36
Influence of behavior and risk factors on motor performance in preterm infants
Figure 1 Flow chart describing composition and inclusion process per
cohort-year and for final 1996–2001 cohort
2
1996
1997
1998
1999
2000
2001
Combined 1996-2001 cohorts
n
114
4
n
142
1
n
154
2
n
137
1
n
124
2
n
128
2
Discharges from the hospital
Deaths after discharge
3
2
4
10
1
59
3
1
40
0
Not invited
Disability child
105
139
141
74
121
86
Eligible
13
14
20
9
9
16
Untraceable/living abroad
15
31
25
3
20
94
3
23
70
466
4
12
4
9
3
+ 1
29
67
437
%
n
799
12
112
11
100
664
12
81
3
3
4
1
-
1
3
8
1
3
5
5
5
1
1
6
2
2
5
5
6
4
9
3
8
Reasons for refusal to participate
:
Too problematic/taxing
Unknown
Parents report no problems
Recently checked by general practitioner
Follow-up Elsewhere
+
14
-
4
6
2
9
2
Did not show up/ill/cancelled
81
103
98
55
85
44
Tested
Reasons for exclusion from analysis
:
Disability
Tested in wrong ageband
Uncooperative during testing
Language problems
Ill at the scheduled time of testing
79
99
92
47
80
40
Analyzed
75
71
65
63
66
47
% of children per eligible cohort year
12
15
14
7
12
6
% of children as proportion of total eligible cohort
examination mainly for visual impairments and CP which made them unable to
perform the test. Thus, 4.5% of the infants (11+12 =23) in the final 6-year cohort
were considered to have a disability. The data of 437 children (244 males, 193
females) were analyzed. At assessment, the children’s mean corrected age was
29 months (SD 3.3). Other mean (SD) sample demographics were: gestational
age 29+5 weeks (1+5), range 25+0 –32+0, birth weight 1213.7 grams (331.7), range
468–2350, and days in the neonatal intensive care unit 21.1 (21.3), range 1–165.
37
Chapter 2
Motor scale
PDI scores and the developmental motor-age scores were not normally
distributed. Three peaks could be distinguished in the PDI data, around a score
of 50, 84, and 98 respectively. We, therefore, decided that the use of medians
was appropriate, but means are also given to allow comparison with the results
of other studies. Table I lists the PDI scores and Table II the Motor Scale
classification. The median PDI score was 86 and the mean developmental motor
age was 25.7 months (SD 5.6). Overall, 46.5% of the children scored within the
category of delayed motor performance (i.e. 27% mildly delayed and 19.5%
significantly delayed).
Table I BSID-II Motor Scale-PDI scores for all children tested and analyzed
(n=437), for males and females and tested age groups (months)
Corrected age
tested, mo,
mean (SD)
All cohorts
29.0(3.3)
Males
28.9(3.3)
Females
29.2(3.3)
Tested age group, mo
20–22
23–25
26–28
29–31
32–34
35–37
38–42
21
24.5(0.6)
27.1(0.8)
30.2(0.8)
32.9(0.9)
35 0 (0.0)
38.5(0.7)
Number
of
children
PDI
PDIa
Median Mean (SD)
Developmental
motor age, mo,
mean (SD)
437
244
193
86
84
90
85(17.9)
82(17.3)
89(18.1)
26(5.6)
25(5.3)
27(5.8)
1
99
86
141
103
5
2
93
80
73
91
91
104
73
93
80(13.8)
74(15.5)
90(17.1)
92(18.9)
102(9.3)
73(22.6)
21
21(3.5)
23(3.5)
28(3.8)
30(5.3)
35(2.4)
26(13.4)
Mean psychomotor developmental index (PDI) 100 (SD 15). Abbreviation: BSID-II, Bayley Scales
of Infant Development, 2nd edition.
a
The Motor Scale classification of the study sample differed significantly from
that of the reference group (c2=625.0, df 3; p<0.01). Males performed significantly
worse than females (Mann–Whitney U 18108; p<0.01). Moreover, we expected
that the higher the corrected age of the children (and therefore, the higher the
tested age band), the better the motor performance would be. In four age
bands between 23 and 34 months (n=429), a sufficient number of children were
38
16
68
13.5
2.5
Accelerated
Within normal limits
Mildly delayed
Significantly delayed
3.4
50.1
27.0
19.5
Total
cohort
%
4.1
60.6
17.1
18.1
%
%
2.9
41.8
34.8
20.5
Females
Males
0
33.3
44.4
22.2
Age group
23–25(mo)
%
1.2
27.9
29.1
41.9
Age group
26–28(mo,
%
3.5
63.1
23.4
9.9
Age group
29–31(mo)
%
8.7
64.1
15.5
11.7
Age group
32–34(mo)
%
75
15
10
Within normal limits
Questionable
Non-optimal
68.6
14
17.4
Total
score
%
90.9
78.4
66.9
Mean
PDI
67.7
14.2
18.1
Orientation/
engagement
%
89.8
79.3
71.3
Mean
PDI
72.1
14.6
13.3
Emotional
regulation,
%
89.8
76.6
68.1
Mean
PDI
BSID-II=Bayley Scales of Infant Development, 2nd edition; BRS=Behavior Rating Scale; PDI=psychomotor developmental index.
Norm
group
%
Classification
(n=437)
64.3
19.2
16.5
Motor
quality,
%
91.0
78.5
68.9
Mean
PDI
Table III BSID-II classification for BRS and mean PDI scores for the BRS subscales for all children tested and analyzed
BSID-II, Bayley Scales of Infant Development, 2nd edition.
Norm group
expected
%
Classification
Table II BSID-II Motor Scale classification for males, females and tested age groups (n=437)
Influence of behavior and risk factors on motor performance in preterm infants
2
39
Chapter 2
tested and significant differences in motor performance were present (Kruskall–
Wallis ANOVA = 71.2, df 3; p <0.01). From the 29 to 31 months age band motor
performance was better.
Behavior rating scale
Ninety-one percent of the parents classified their child’s test-taking behavior as
(very) typical of their child, and 82% of the parents classified the test adequacy
as good or excellent. Table III shows the classification on the BRS. We observed
abnormal test-taking behavior in 31.4% of the children, and abnormal motor
quality in 36.7%. Outcomes for the total score (c2=26.5, df 2; p<0.01), and the
subscale scores orientation/engagement (c2=31.8, df 2; p<0.01) and motor
quality (c2=30.2, df 2; p<0.01) deviated significantly from the reference sample.
No deviation was observed for the subscale emotional regulation (c2=5.2, df 2;
p=0.07).
The PDI and BRS scores were significantly correlated (rs=0.62, p<0.01) as was
the Motor Scale classification with all three BRS subscale scores (orientation/
engagement: rs=0.50, p<0.01; emotional regulation: rs= 0.57, p<0.01; and
motor quality: rs= 0.50, p<0.01). Thus, motor performance tended to be better
if the BRS showed ‘normal’ test-taking behavior.
Risk factors
Table IV shows the distribution of risk factors in the final cohort and the results
of the binary logistic regression analysis for the association with delays in motor
performance. Forward and backward (likelihood ratio) regression analyses
(n=393) revealed the same four factors attributing to delayed motor performance:
neonatal convulsions, chronic lung disease, male sex, and low maternal educational
level, with the highest OR for neonatal convulsions.
40
Influence of behavior and risk factors on motor performance in preterm infants
Table IV Percentage of risk factors and results of binary logistic regression
analysis of risk factors for delayed motor performance for total
number of children tested and analyzed (n=393)a
Percentage in
cohort
Sepsis
54b
Meningitis/encephalitis/ventriculitis
4
Intracranial hemorrhage
14b
Post-hemorrhage ventricle dilatation
3b
Ischemic cerebral abnormality
6b
Neonatal convulsions
6b
Retinopathy of prematurity
5
Idiopathic respiratory distress syndrome
54
Chronic lung disease
13b
Respiratory problems
9
Hyperbilirubinemia
68b
Necrotizing enterocolitis
6
Male sex
56b
Gestational age
<28wks
≥ 28 < 30wks
≥ 30wks
Birthweight
< 1000g
≥ 1000g
Educational level of mother
Low education
Average education
High education
Multiple births
b
21
52
18
24b
Caesarean section
57
Apgar score <7 at 5min
34b
Congenital malformation
10
Endotracheal intubation after birth
33b
2
OR (95% CI)
4.5 (1.6–12.9)*
b
20
35
45
2.1 (1.1–4.1)*
2.8 (1.8–4.3)*
1
31
69
b
3.3 (1.7–6.5)*
1
9% of maternal education and 2% of Apgar scores missing
factors entered in binary stepwise logistic regression analysis, p < 0.5
* p<0.05
Abbreviations: OR=odds ratio; CI=confidence interval
a
b
41
Chapter 2
Discussion
We investigated test-taking behavior and risk factors for delayed motor performance
in a 6-year cohort (1996–2001) consisting of 437 preterm infants (244 males, 193
females) born before 32 weeks’ gestation at the corrected age of 2 to 3 years
(mean age 29mo [SD 3.3]). Nearly one-half (46.5%) of the children showed delayed
motor performance based on their BSID-II score. Test-taking behavior was not
optimal in nearly one third (31.4%) of the participants and related to their motor
performance. Four risk factors (i.e. neonatal convulsions, chronic lung disease,
male sex, and low maternal educational level) were found to be associated with
the observed motor delays.
In total, 70% of the children invited to participate in our study actually did so. This
is a substantial sample compared with 56% of a national cohort in Switzerland14
and 64% of a cohort in the Netherlands7, although it is considerably lower than the
82 to 97% in other studies.5;15;16 The children who were not invited to participate
(n=112) did not differ significantly from the participating children in terms of risk
factors and, therefore, selection bias was not introduced into this cohort study.
However, attrition bias may represent a problem since 30% did not participate
and this may have resulted in an underestimation of the presence of delayed
motor performance.17-19
The BSID-II is norm-referenced (standardized in 1991–1992 on 1700 US children)
with age-adjusted scores; as such, it allows individual test scores to be compared
with data of an age-related term norm group. Consequently, we considered a
norm group unnecessary. The use of US BSID-II reference values for a Dutch
population may be open to discussion. At the initiation of this study the Dutch
version of the BSID-II was not available, and at the time of its publication in 2002
had undergone substantial changes. We checked in our study the difference in
PDI between the US and Dutch versions and found a significant but low mean
difference (paired t-test) of -1.5 (SD 4.7, 95% CI -1.9 to -1.0). This difference is in
line with that of Westera et al.20 except our PDI scores for the Dutch PDI were
higher. We decided to use the US reference values to guarantee comparability
with international studies.
Mean PDI scores found in our study are lower and the percentage of infants with
delayed motor performance are higher than values published in earlier studies
using the BSID-I.6;7;21;22 As BSID-I scores in the normal population are higher than
BSID-II scores, underestimation with the BSID-I is plausible. The mean PDI in our
42
Influence of behavior and risk factors on motor performance in preterm infants
study was higher than that reported by Bayley (86 vs. 83.5) despite the inclusion
of preterm infants with a gestational age of less than 36 weeks in the latter study.11
Studies reporting BSID-II outcomes in infants with a birth weight below 1000g,15;23
or a gestational age of 20 to 25 weeks16 mentioned much higher motor-delay
proportions (as high as 64–86%). Although percentages differ in all these studies,
it can be concluded from our study and the literature that preterm infants are at
risk of delayed motor performance at age 2 to 3 years.
Our data show that the mean PDI score is higher from the 29 to 31 months age
group than in the lower age groups, which suggests a ‘catch-up’ pattern. However,
it could be argued that this is a test artifact. The BSID-II contains both gross and
fine motor items, as these skills cannot be arranged into two different subscale
scores, both domains were tested together. The pediatric physical therapists’
examinations revealed that gross motor functions (e.g. climbing stairs, running,
jumping) were the most delayed and that motor performance in these children
was similar to that of younger aged children, despite the correction for preterm
birth. This delay was detected in the lower age groups but disappeared from the
29 to 31 months age group. In the study of extremely preterm infants (EPICure
study), infants were also tested at a corrected age of 30 months.16;24 These results
underscore the need to implement the follow-up assessment at a corrected age
of 30 months as this may prevent over-referral to pediatric physical therapists.
Although delayed motor performance seems to be a mild disability, it may
influence the possibilities of the child exploring the environment and, consequently,
impede development in general.
We used the BRS in addition to the Motor Scale because the BRS can facilitate
interpretation of the test results.4 The BRS enables evaluation of the qualitative
aspects of the child’s test-taking behavior, such as task compliance, orientation
of the child to the examiner, emotional regulation, and the adequacy in control
and coordination in fine and gross motor tasks, and reduced muscle tone or
stiffness thereby enabling the examiner to assess how the child copes with
increased task load. Delay in motor performance can then better be placed in the
context of the question as to whether pediatric physical therapy or other
interventions are justified.
The BRS is scored after the Motor Scale and it takes about 5 minutes; the parents
have to be asked two additional questions, which takes approximately 1 minute.
Therefore, it is a relatively simple and inexpensive approach for assessing test-taking
behavior and determining whether the behavior was representative of the child.
43
2
Chapter 2
Thompson and his group investigated the construct validity of the BRS in 1996. 25
They reported that motor performance is a dominant component of the Total
score and that its structure becomes more complex at older ages as older
children show a more heterogeneous behavior.
As 90% of the parents stated that the test-taking behavior of their child was
typical of the child and not related to the test situation, one may assume that in
daily life these children cope in the same manner when faced with new or
challenging situations. Although we had excluded children who refused to
perform tasks from our analysis, we found a relationship between the BRS total
and subscale scores and motor performance. Future studies in which both
the BSID-II Motor Scale and BRS are used as assessment tools could provide
more information on the relationship between test-taking behavior and motor
performance.
Male sex and chronic lung disease were found to be risk factors for delayed
motor performance, which is in accordance with other studies. However, the
ORs in our study were lower than those reported in the literature.16;23;24
While low maternal educational level is a known risk factor for cognitive
development in preterm infants, we also found it to be a risk factor for motor
performance. Neonatal convulsions was the highest risk factor. In our sample,
28 children had convulsions, 17 of whom also had cerebral damage (11 hemorrhage,
one meningitis/encephalitis, four ventricle dilation, and one ischemic damage).
‘Severe cranial ultrasound abnormalities have been found to strongly predict motor
disability, but one third of infants with CP had no ultrasound abnormalities’.26(p828)
In contrast to the EPICure study, where a regression analysis was performed on
the mean PDI score, 24 we applied binary logistic regression analysis on defined
delayed and normal motor performance. The EPICure study included only
children under 25 weeks of gestation, which was the most plausible explanation
for the lower OR we found for sex and chronic lung disease and a difference in
significant factors.16;23;24
Conclusion
Based on our findings, we conclude that it is worthwhile to quantify the behavior
of preterm infants during motor assessment to determine its effect on their motor
performance. Additionally, we recommend that future studies assess motor quality
44
Influence of behavior and risk factors on motor performance in preterm infants
separately from the behavior scores. More longitudinal studies are required to
establish the predictability of delayed motor performance at the age of 2 to 3
years as a tool for assessing developmental problems at older age.
Acknowledgements
We thank all of the children and their parents for participating in this study. We
also thank the staff of the neonatal intensive care unit and the pediatric physical
therapists: S Haring, B Kolzer, A Langelaan, G Peters, and I Wouters. We
especially thank T Plamont, D Vrins, and H Verbeet for secretarial assistance
and planning.
45
2
Chapter 2
References
1. De Kleine MJ, den Ouden AL, Kollée LA, Nijhuis-van der Sanden MW, Sondaar M, van
Kessel-Feddema BJ, et al. Development and evaluation of a follow up assessment of preterm
infants at 5 years of age. Arch Dis Child 2003; 88: 870-75.
2. Vohr BR, O’Shea M, Wright LL. Longitudinal multicenter follow-up of high-risk infants: why, who,
when, and what to assess. Semin Perinatol 2003; 27: 333-42.
3. Msall ME. Neurodevelopmental surveillance in the first 2 years after extremely preterm birth:
evidence, challenges, and guidelines. Early Hum Dev 2006; 82: 157-66.
4. Johnson S, Marlow N. Developmental screen or developmental testing? Early Hum.Dev. 2006; 82:
173-83.
5. van Zeben-van der Aa TM, Verloove-Vanhorick SP, Brand R, Ruys JH. Morbidity of very low
birthweight infants at corrected age of two years in a geographically defined population. Report
from Project on Preterm and Small for gestational age infants in The Netherlands. Lancet 1989;
333: 253-55.
6. Rijken M, Stoelhorst GM, Martens SE, van Zwieten PH, Brand R, Wit JM, et al. Mortality and
neurologic, mental, and psychomotor development at 2 years in infants born less than 27 weeks’
gestation: the Leiden follow-up project on prematurity. Pediatrics 2003; 112: 351-58.
7. Stoelhorst GM, Rijken M, Martens SE, van Zwieten PH, Feenstra J, Zwinderman AH, et al.
Developmental outcome at 18 and 24 months of age in very preterm children: a cohort study from
1996 to 1997. Early Hum Dev 2003; 72: 83-95.
8. Vohr BR, Wright LL, Poole WK, McDonald SA. Neurodevelopmental outcomes of extremely low
birth weight infants <32 weeks’ gestation between 1993 and 1998. Pediatrics 2005; 116: 635-43.
9. Shumway-Cook A, Woollacott MH. Motor control: issues and theories. Motor Control; Translating
research into clinical practice, 3rd edn. Philadelphia, Pennsylvania: Lippincott Williams & Wilkens,
2005: 3-20.
10. World Health Organization. The International Classification of Functioning, Disability and Health
(ICF). Geneva: World Health Organization, 2001.
11. Bayley N. Bayley Scales of Infant Development, second edition. San Antonio: The Psychological
Corporation, 1993.
12. Provost B, Heimerl S, McClain C, Kim N, Lopez B, Kodituwakku P. Concurrent validity of the Bayley
Scales of Infant Development II motor scale and the Peabody Developmental Motor Scales-2 in
children with developmental delays. Pediatr Phys Ther 2004; 16: 149-56.
13. Van der Meulen BF, Ruiter SAJ, Spelberg HC, Smrkovský M. BSID-II-NL Dutch manual. Lisse:
Swets Test Publishers, 2002.
14. Bucher HU, Ochsner Y, Fauchere JC. Two years outcome of very pre-term and very low birthweight
infants in Switzerland. Swiss Med Wkly 2003; 133: 93-99.
15. Vohr BR, Wright LL, Dusick AM, Mele L, Verter J, Steichen JJ, et al. Neurodevelopmental and
functional outcomes of extremely low birth weight infants in the National Institute of Child Health
and Human Development Neonatal Research Network, 1993-1994. Pediatrics 2000; 105: 1216-26.
16. Wood NS, Marlow N, Costeloe K, Gibson AT, Wilkinson AR. Neurologic and developmental
disability after extremely preterm birth. EPICure Study Group, N Engl J Med 2000; 343: 378-84.
17. Hille ET, den Ouden AL, Stuifbergen MC, Verrips GH, Vogels AG, Brand R, et al. Is attrition bias a
problem in neonatal follow-up? Early Hum Dev 2005; 81: 901-08.
18. Pennefather PM, Tin W, Clarke MP, Dutton J, Fritz S, Hey EN. Bias due to incomplete follow up in a
cohort study. Br J Ophthalmol. 1999; 83: 643-45.
19. Wolke D, Sohne B, Ohrt B, Riegel K. Follow-up of preterm children: important to document
dropouts. Lancet 1995; 345: 447.
20. Westera JJ, Houtzager BA, Overdiek B, van Wassenaer AG. Applying Dutch and US versions of the
BSID-II in Dutch children born preterm leads to different outcomes. Dev Med Child Neurol 2008;
50: 445-49.
46
Influence of behavior and risk factors on motor performance in preterm infants
21. Tommiska V, Heinonen K, Kero P, Pokela ML, Tammela O, Jarvenpaa AL, et al. A national two year
follow up study of extremely low birthweight infants born in 1996-1997. Arch Dis Child Fetal
Neonatal Ed 2003; 88: F29-F35.
22. Pinto MJ, Whitaker AH, Feldman JF, Van-Rossem R, Paneth N. Relation of cranial ultrasound
abnormalities in low-birthweight infants to motor or cognitive performance at ages 2, 6, and 9
years. Dev Med Child Neurol 1999; 41: 826-33.
23. Hack M, Wilson-Costello D, Friedman H, Taylor GH, Schluchter M, Fanaroff AA. Neurodevelopment and predictors of outcomes of children with birth weights of less than 1000 g: 1992-1995.
Arch Pediatr Adolesc Med 2000; 154: 725-31.
24. Wood NS, Costeloe K, Gibson AT, Hennessy EM, Marlow N, Wilkinson AR. The EPICure study:
associations and antecedents of neurological and developmental disability at 30 months of age
following extremely preterm birth. Arch Dis Child Fetal Neonatal Ed 2005; 90: F134-F140.
25. Thompson B, Wasserman JD, Matula K. The factor structure of the behavior rating scale of the
Bayley scales of infant development-II. Educ Psychol Meas 1996; 56: 460-74.
26. Ancel PY, Livinec F, Larroque B, Marret S, Arnaud C, Pierrat V, et al. Cerebral palsy among very
preterm children in relation to gestational age and neonatal ultrasound abnormalities: the EPIPAGE
cohort study. Pediatrics 2006; 117: 828-35.
47
2
Chapter 3
Unstable longitudinal motor
performance in preterm infants from
6 to 24 months on the Bayley Scales of
Infant Development – second edition
Anjo JWM Janssen MSc, PT1,2, Reinier P Akkermans MSc2,
Katerina Steiner MD3, Olga AM de Haes MD4, Rob AB Oostendorp PhD, PT2,
Louis AA Kollée PhD, MD3, Maria WG Nijhuis-van der Sanden PhD, PT1,2
Radboud University Nijmegen Medical Centre, Department of Pediatric Physical Therapy, Geert
Grooteplein zuid 10, 6525 GA Nijmegen, The Netherlands
2
Radboud University Nijmegen Medical Centre, Scientific Institute for Quality of Healthcare, Geert
Grooteplein noord 21, 6525 EZ Nijmegen, The Netherlands
3
Radboud University Nijmegen Medical Centre, Department of Pediatrics, Geert Grootplein zuid 10,
6525 GA Nijmegen, The Netherlands
4
R adboud University Nijmegen Medical Centre, Department of Primary and Community Care,
Kapittelweg 54, 6525 EP Nijmegen, The Netherlands
1
Res Dev Disabil. 2011; 32: 1902-1906
Chapter 3
Abstract
Preterm birth increases the risk for neurologic and developmental disabilities
and therefore long-term follow-up is important. This prospective follow-up study
aims to describe longitudinal motor performance in preterm infants from 6 to 24
months and to detect the influence of risk factors on motor performance
trajectories.
We included preterm infants (n = 348) with a gestational age of ≤ 32 weeks. The
Bayley Scales of Infant Development, 2nd edition (BSID-II) Motor Scale and the
Behavior Rating Scale were recorded at the corrected ages of 6, 12 and 24
months. The Motor Scale raw score was the dependent variable in random
coefficient analysis for risk factors in the cohort if infants with cerebral damage
were in- and excluded.
The raw score increased, showed the highest correlation (rp= 0.48–0.67) and
was more stable than the PDI and its classification. Fifteen percent of the infants
had a stable classification, while 45% changed one class. Male sex and intraventricular hemorrhage (IVH) were related to lower raw scores. Higher motor
quality scores and height were related to higher raw scores, while the influence
of maternal education varied at different time points. Removal of infants with
cerebral damage from the cohort did not change the risk factors. The results
showed that the raw score trajectories were more stable, but after corrections
for norm data, the measurements became highly unstable. This is clinically
important when reporting results to parents, guiding intervention and in
randomized trials. The risk factors predominantly influenced the level of motor
performance raw scores. Maternal education additionally influenced the
trajectory and should be included in randomization procedures.
50
Unstable longitudinal motor performance in preterm infants from 6 to 24 months
Introduction
Preterm birth is defined as childbirth before 37 completed weeks of gestation as
measured from the first day of the last menstrual period.1 Preterm birth increases
the risk for neurologic and developmental disabilities2-4 and therefore, regular
and long-term follow-up is important.5 The evaluation of motor performance in
preterm infants is important because motor impairment influences motor and
cognitive ability in later life.6;7 However, few studies have investigated motor
performance trajectories in follow-up practice.
Such evaluations are especially relevant from 6 to 24 months, when the motor
performance of preterm infants differs both qualitatively and quantitatively from
that of term infants.8-10 Qualitative longitudinal motor performance is influenced
by the spontaneous recovery of suspected or abnormal motor signs observed
in some preterm infants.11;12 Quantitative motor performance shows delayed
acquisition of motor milestones8;9 and an indication of unstable longitudinal
motor performance in preterm infants.13
Several cross-sectional studies have investigated the influence of risk factors on
motor performance,3;14;15 but only one study has investigated the influence of
risk factors on longitudinal motor performance.16
The aims of this study were to describe longitudinal motor performance in
neonatal follow-up practice in individual preterm infants from 6 to 24 months
and to detect the influence of risk factors on motor performance raw score
trajectories, in the total cohort (T) and also in the same cohort if infants with
cerebral damage were excluded (E).
Methods
In this prospective longitudinal study, we assessed motor performance at three
different time points from 6 to 24 months as part of a standardized multidisciplinary follow-up program for preterm infants at the Radboud University
Nijmegen Medical Centre, the Netherlands. Medical ethical committee approval
was not required, since the protocol is part of accepted medical practice.
51
3
Chapter 3
Participants
The study cohort consisted of surviving infants born between March 2003 and
December 2006 at less than 32 weeks of gestation and who had been admitted
to the neonatal intensive care unit. All infants were assessed at the corrected
age (CA), which is the chronological age corrected for the days born too early,
at 6 (t 0), 12 (t1) and 24 (t 2) months by eight experienced pediatric physical
therapists (PPTs). Parents and/or caregivers were present throughout the test
procedures.
Instruments
Motor performance and behavior were assessed using the Bayley Scales of
Infant Development, 2nd edition (BSID-II). The BSID-II Motor Scale (BSID-II-MS)
and the BSID-II Behavior Rating Scale (BSID-II-BRS) were recorded at t 0 and t1
and t 2.17 Based on content, construct and concurrent validity, the BSID-II has
been shown to be a comprehensive and appropriate instrument for assessing
motor performance.18 We used American norm references because the Dutch
version includes modifications to the test procedure that hamper comparison
with international studies.19
The BSID-II-MS assesses gross and fine motor skills. Raw scores (range 0–111)
are recalculated into the Psychomotor Developmental Index (PDI; mean of 100
[SD 15]) using norm-referenced tables. The PDI scores are subsequently
transformed into the following classifications, with scores of ≤ 69 (-2 SD) indicating
“significantly delayed”, 70–84 (-1 SD) indicating “mildly delayed”, 85–114 indicating
“within normal limits” and ≥115 (+1 SD) indicating “accelerated” motor performance.
Developmental age equivalents are determined from the raw scores. A difference
of 1–3 points on the BSID-II raw score means a 1-point difference in PDI scores.
The BSID-II-BRS assesses the qualitative aspects of the infant’s behavior during
testing (“test-taking behavior”) and comprises 30 questions, each scored on a
5-point scale. At the ages assessed, the questions are related to three subscales:
(1) orientation/engagement, (2) emotional regulation and (3) motor quality. The
total raw score and subscale raw scores can be calculated.
The Gross Motor Function Classification System (GMFCS) was used to classify
the subgroup of infants with cerebral palsy (CP) at t 2. It consists of a 5-level
pattern recognition system in which levels I and II indicate that there is potential
to walk and levels III through to V represent an increasing limitation in the
self-mobility of infants. 20
52
Unstable longitudinal motor performance in preterm infants from 6 to 24 months
Factors included
The neonatologist recorded individual risk factors at hospital discharge. The
time-independent factors were: sex, gestational age (GA, days), birth weight
(grams), neonatal convulsions, retinopathy of prematurity (ROP), necrotizing
enterocolitis (NEC), intraventricular hemorrhage (IVH)21, periventricular leukomalacia
(PVL)22;23 and chronic lung disease (CLD). Parents reported maternal education
at t 2 which was included as an indicator of socioeconomic status (SES)24;25,
categorized as: low (no/elementary school/vocational training); average (lower
general secondary education/secondary vocational education) or high (upper
general secondary education, pre-university education/bachelor degree or
higher).
At t 0, t1 and t 2, a nurse measured the time-dependent factors weight (kilograms),
height and head circumference (centimeters). These data were included since
there is some evidence that growth in preterm infants predicts delayed motor
performance. 26;27 The PPT registered PPT treatment (yes/no).
From the assessment CA tested (months), the BSID-II-BRS total raw scores and
subscale motor quality raw scores were included because behavior influences
motor performance. Only these two were selected because the correlation at t 0,
t1, t 2 between the BSID-II-BRS total score and the subscale scores orientation/
engagement (rs = 0.80/0.84/0.79) and emotional regulation (r s = 0.73/0.52/0.85)
was higher than that for the motor quality scale (r s = 0.58/0.36/0.53).
Statistical analysis
We described cohort demographics and motor performance outcomes of the
BSID-II-MS. PDI scores <50 were set at 49. Correlations (Pearson and Spearman,
where appropriate) were calculated for the BSID-II-MS scores at t 0/t1, t1/t 2 and
t 0/t 2. The distribution of the classification over the three assessments was
visualized. A stable classification was defined as no change and relatively stable
as only one class change over the three assessments.
The empirical individual growth trajectories were visually examined for all infants
and presented for every 20th infant in the cohort with three measurements for the
raw and PDI scores. Because of the hierarchical structure of the study (repeated
assessments nested within the individual children), we performed multilevel
analyses. 28 A random coefficient analysis with the raw score as the dependent
variable was used to describe longitudinal trajectories for individual infants and
to detect the influence of risk factors. The best fitting model included both linear
53
3
Chapter 3
and quadratic components of age. The -2log likelihood and Akaike information
criteria statistics compared the fit of the different models. Visual inspection
revealed no differences for the trajectories of the infants with and without
missing data.
Risk factors were univariately explored, and factors with a significance level of
< 0.2 were entered in the multivariate analysis, in the cohort T and E. Excluded
were infants with CP and infants with IVH grades III and IV, PVL or other
intracerebral hemorrhage, the latter’s because these infants have an increased
risk of CP.
In all cases, two-sided p-values < 0.05 were considered to be statistically
significant. Data were analyzed using SPSS (ver. 16) and multilevel analysis was
performed using SAS PROC MIXED, using the full maximum likelihood method
(ver. 9.2).
Results
The flow chart in Figure 1 provides an overview of the infants assessed. A total
of 407 infants were eligible, of whom 59 (14.5%) did not participate. The final
study cohort consisted of 348 infants (194 males, 56%) with a mean birth weight
of 1268 grams (SD 364.8) and a gestational age of 29+4 weeks (SD 1+5). There
were three reliable assessments for 272 (78.2%) of the infants. The mean CA at
6 months was 6.3 months (SD 0.5), at 12 months, 12.0 months (SD 0.5) and at
24 months, 24.6 months (SD 0.7). The patient characteristics and risk factors of
the 59 non-participants did not differ significantly from those of the participants.
Of the 21 (6.3%) infants diagnosed with CP, five had unilateral (GMFCS level I)
and 16 bilateral CP (GMFCS: 8 level I; 3 level II; 3 level III and 2 level IV). One
infant was diagnosed with Down’s syndrome, one with spinal muscular atrophy
and one with severe psychomotor retardation. Three infants were visually
impaired and one infant was blind.
Table 1 shows the number of infants and the mean outcome on all BSID-II-MS
scores for the cohort (T and E) at t0, t1 and t2. For the T cohort, the highest
correlation was between the raw scores (rp= 0.48–0.67) and the lowest was between
the classification (rs = 0.24–0.41). The percentage of infants with motor performance
within normal limits increased from 24% at t0 to 47% at t2. The percentage of
infants with significantly delayed motor performance was highest at t1.
54
Unstable longitudinal motor performance in preterm infants from 6 to 24 months
Figure 1 Flow chart of infants assessed
May
2003
n
90
2004
2005
2006
Total cohort
n
137
n
139
n
121
17
13
24
20
Deaths
73
2
1
124
2
115
1
-
101
-
Discharged
Deaths after discharge
Not invited
413
3
3
Eligible
407
n
487
No assessment 14.5%
74
100%
3
59
Untraceable / living abroad
16
Reasons for refusal to participate:
Too problematic
5
Unknown
3
Parents report no problems
6
Recently checked by GP
4
Follow-up elsewhere
22
Did not show up / ill / cancelled
3
Included
Unreliable assessment 3.2%
One assessment 6.6%
23
85.5%
348
11
Two assessments 12.1% 42
t0
t1
t2
t0 /t1
t0 /t2
t1 /t2
14
6
3
20
10
12
Three assessments 78.2% 272
Figure 2a graphically depicts the distribution of classifications over the three
assessments for all infants. The most common classification was that of mildly
delayed at t 0 and t1 and normal at t 2 in 43 infants (12.4%). Among the infants with
normal motor performance at t 0 or t1, eight (2.3%) infants deteriorated to a
55
56
0.35*
0.32
0.31*
0.30
81.2 (5.0)
0.51*c
*c
0.39
0.67*c
0.49*c
0.48*c
0.32*c
266
297
261
286
254
283
249
T
E
T
E
T
E
0.26*
0.28*
0.30*
0.41*
*
21.9 (3.0)
b
a
74.3 (12.9)
0.26*
0.30*
0.43*
0.45*
*
Correlationb
85.3 (15.4)
84.0 (15.9)
E
21.5 (3.5)
9.7 (1.4)
73.3 (13.2)
57.6 (4.9)
E
9.6 (1.5)
77.2 (11.2)
76.0 (11.6)
80.5 (6.7)
57.1 (5.8)
318
T
4.7 (0.94)
4.8 (0.86)
282
33.7 (4.2)
Mean (SD)
PDI
300
33.3 (4.6)
322
283
T
E
Mean (SD)
Developmental age
T
Mean (SD)
n
Raw score
Cerebral damage excluded CP/IVH III, IV/PVL or other intracerebral hemorrhage
Spearman correlation coefficient
c
Pearson correlation coefficient
* p<0.05
t0/t2
t1/t2
t0/t1
t2 (24 mo)
t1 (12 mo)
t0 (6 mo)
BSID-II-MS
131 (49.3)
140 (46.7)
66 (23.4)
69 (21.7)
75 (26.5)
77 (23.9)
n (%)
Within
normal limits
0.20*
0.24*
0.39*
0.41*
0.25*
0.29*
87 (32.7)
98 (32.7)
122 (43.3)
135 (42.5)
146 (51.6)
162 (50.3)
n (%)
Mildly
delayed
Classification
48 (18.1)
62 (20.7)
94 (33.3)
114 (35.8)
62 (21.9)
83 (25.8)
n (%)
Significantly
delayed
mo) and t2 (24 mo) given as the frequency and percentage for the total cohort (T) and infants with cerebral damage
excluded (E)a. Correlation for the three follow-up assessments between t0 and t1 and t2
Table 1 Motor performance on raw score, developmental age, PDI mean (SD) and classification for PDI at t0 (6 mo), t1 (12
Chapter 3
Unstable longitudinal motor performance in preterm infants from 6 to 24 months
significantly delayed motor performance at t 2. Of the 272 infants with three
reliable assessments, 42 (15.4%) had a stable classification. The classification
was relatively stable (one class change) in 123 (45.2%) infants. An unstable
classification was seen in 107 (39.3%).
Figure 2 Percentage of the PDI classification over the three assessments
for all infants (n=348)
3
Classifying the PDI score in significantly delayed as 1, and mildly delayed as 2 and normal motor
performance as 3.
1
In the figure, 2
s hould be read as significantly delayed motor performance at t0, mildly delayed
at t1 and normal motor performance at t2.
3
}
Figure 3 shows the instability in the rate and the shape of the individual PDI
scores and the more stable raw scores. The difference (t1-t 0 and t 2-t1) of the raw
scores and PDI scores vary from 0 to 42 and -40 to 51, respectively. The 95%
limits of agreement vary from 12.88 to 33.91 and -29.9 to 40.49, respectively.
This indicated highly unstable PDI scores.
Table 2 presents the results of the multilevel analyses. Five factors in the T
cohort and four factors in the E cohort had a significant effect on the level of
motor performance, but only maternal education showed additionally a
significant effect on motor performance trajectories. In the T cohort, the mean
raw score was 1.57 lower for infants with IVH and that for males was 0.95 lower
than that for females. Higher motor quality subscale scores and height increased
57
Chapter 3
Figure 3 Random individual preterm infant’s trajectories (n = 14) of BSIDII-MS raw score and PDI for the same infants and mean raw score
and mean PDI (Expected mean 100, SD 15)
MS raw score individual infants
PDI individual infants
90
115
85
110
80
105
infant
75
70
infant
100
infant
95
infant
65
infant
60
infant
infant
55
infant
50
infant
45
infant
90
85
80
75
70
infant
40
35
infant
65
infant
60
infant
30
mean
25
55
50
20
45
6 months
12 months
18 months
24 months
6 months
12 months
18 months
24 months
the raw scores. For example, in the T cohort, a 1-point higher score on the motor
quality subscale was associated with a 0.42 higher raw score.
The effect of maternal education was more difficult to interpret because there
was a double effect, which varied the score at different time points. In the T and
E cohort, at t 0, infants from mothers with a low educational level started with a
lower mean raw score (3 points) than those with a medium or high educational
level. A lower educational level showed the fastest rate of increase up to t1
(1 point higher), but the lowest rate of increase from t1 to t 2 (2 points lower).
The mean SD of the raw scores was 5.8 over the different assessments.
Therefore, the effect size of the significant risk factors ranged from 0.16 to 0.52,
which indicates a small to medium effect size according to the definition of
Cohen. 29 Maternal education was the only factor with a medium effect size.
58
Unstable longitudinal motor performance in preterm infants from 6 to 24 months
Table 2 Multilevel, multivariate estimates of fixed effects from individual
growth models in which risk factors influence the intercept, linear
and quadratic rate of change in infants’ motor performance on
the BSID-II-MS raw scores between 6 and 24 months in the cohort
(T and E)
Estimate (SE) of motor performance
Total cohort (T)
n = 348
Excluded cerebral
damage (E)a
n = 300
3
Fixed effects
Intercept
Age b
Age2 b
36.96 (0.72)
4.20 (0.15)
-0.11 (0.01)
-0.81 (0.37)
****
Male sex
37.00 (0.75)
4.21 (0.15)
-0.11 (0.01)
-0.95 (0.39)
IVH, dichotomousc
-1.57 (0.62)
****
–
Height
0.29 (0.05)
****
0.28 (0.05)
****
BRS “motor quality”
0.42 (0.05)
****
0.38 (0.04)
****
****
****
**
****
****
****
**
Maternal Education
Low
-0.88 (0.77)
Low x Age
Low x Age2
Average
0.64 (0.23)
-0.03 (0.01)
0.79 (0.53)
-0.81 (0.76)
Average x Age
-0.14 (0.16)
-0.12 (10.16)
Average x Age2
0.01 (0.01)
0.01 (0.01)
***
***
0.58 (0.22)
-0.03 (0.01)
0.77 (0.53)
***
***
Cerebral damage: excluded CP/IVH III, IV/PVL or other intracerebral hemorrhage
Centered corrected age infant (months) [= corrected age infant (months) - 6 months]
c
IVH, Dichotomous combining Grades I and II, and combing grades III and IV (grades according
to Papile et al 21) with PVL and other cerebral hemorrhagic.
Abbreviations: BSID-II-MS, Bayley Scales of Infants Development, second edition, motor scale,
x = interaction effect
**
p<0.05 ***p<0.01 **** p< 0.001
a
b
59
Chapter 3
Discussion
In our study, longitudinal motor performance on the BSID-II-MS PDI of preterm
infants from 6 to 24 months was unstable, which resulted in an unstable classification
in about 39% of the infants. The BSID-II-MS raw scores improved, were more
stable and the correlation over the three assessments was sufficient to detect
the influence of risk factors. Male sex, height, IVH and motor quality influenced
the level of the raw score trajectories, while the influence of maternal education
varied at different time points. Removal of infants with cerebral damage from the
cohort did not change the risk factors.
The unstable PDI results are discussed by three arguments: (1) Would the use
of the Dutch reference values lead to different results? The mean PDI scores at
t 0 and t1 in our study are lower than those previously reported,13 and the PDI
scores at t 2 are close to reported values.6 Comparing the PDI scores using the
U.S. and Dutch norm references revealed that our mean PDI differences are in
line with previously reported data.19 We agree with these authors that large
differences at t 0 and t1 are likely to have been caused by a bias in the Dutch
normative data, although a slower developmental pace of Dutch infants in
general could also play a role.19 In addition, the mean PDI score at t1 was lower
than that at t 0 and t 2, possibly due to the above-mentioned factors. The range of
the individual differences in PDI trajectories based on the Dutch reference
values were also checked and were the same as the individual differences in
PDI trajectories using the U.S. reference values. Therefore, the instability was
the same when the Dutch reference values were used. (2) There could also be
an increased variation of motor performance in preterm infants compared to
term infants. However, our standard deviation at t 0 and t1 was lower and at t 2
similar as in the norm data. The cross-sectional established reference values for
Dutch preterm infants on the Alberta Infant Motor Scale from 1 to 18 months
show a significantly delayed performance, but the pattern of the standard
deviation mirrors the pattern of the standard deviation of full-term infants.8 (3)
Also, the longitudinal motor performance of typically developing term infants is
variable and nonlinear rather than constant.30;31 The question is whether this
longitudinal instability is the result of normal variation in motor performance or
due to correction of the raw scores by norm reference values for motor tests
gathered by cross-sectional studies. Given the fact that our raw data show a
much more reliable pattern, an increase at all consecutive measurements and
60
Unstable longitudinal motor performance in preterm infants from 6 to 24 months
motor performance in term infants on cross-sectional norm scores shows a
similar variation, suggests the latter explanation to be most valid. This raises
doubts about the usually adopted method of gathering norm reference data.
The only study that has investigated risk factors on longitudinal motor performance16
uses “stability of longitudinal motor performance” as the dependent variable.
They divided the raw scores in four classes and the group characteristics of
stable and unstable performers were investigated. This is a limitation of this
study because children with severe cerebral palsy and well performers are
stable in their longitudinal motor performance, but their risk factors differ. As a
consequence, the risk factors cannot be compared.
Among the factors examined in our study, those of male sex, IVH and low
maternal education are established risk factors in cross-sectional studies in
preterm infants.3 The factor male sex needs reconsideration, namely, since the
variation in longitudinal associated movements is higher in term boys than in
girls.32;33 In our study, the SD of the raw scores in boys was larger than in girls
(data not shown) and male sex may not be a factor solely related to preterm
infants. The influence of movement quality has been relatively less investigated
in risk factor analysis studies. Prediction studies are available and there is
increasing evidence that movement quality is important in the prediction of
delayed motor performance.34-36 Unexpectedly, we found that height, but not
head circumference, was a risk factor. In most infants with CP, head circumference
has been found to be a predictor of delayed motor performance. 26 The total
cohort included a low percentage (6%) of infants with CP, which might explain
our findings. Height has also been found to predict delayed motor performance
in a cross-sectional study in boys, but not in girls. 27 Furthermore, we looked into
the relation in infants who were born small for their gestational age (SGA); this
factor could not be included in the analyses because the frequency of SGA
(-2SD)37 was only 2.6%. There was a small significant correlation with height at
t 0 (rp=0.21) and t1 (rp =0.14). Therefore, SGA is unlikely to explain our present
findings.
The risk factors influencing longitudinal motor performance trajectories are
similar to those found in cross-sectional studies. The risk factors influenced
mainly the level of motor performance, and only maternal education influenced
the level and motor performance trajectory. The influence of the risk factors was
significant on the raw scores, but they were relatively small. In practice, effect
sizes larger than 0.5 are usually taken to be clinically relevant, making the effect
61
3
Chapter 3
of the factors clinically limited and relevant for maternal education level.
However, an IVH grade IV can mean that an infant has cerebral palsy, which is
for the infant very relevant despite the small effect size. Another limitation of our
study is that we had only three measurements with large time intervals and the
inclusion of only maternal education level as indictor of SES. Coping of the
parents and child, as well as expectancy of the parents regarding the outcome
and learning environment might also influence the outcome.
Longitudinal motor performance from 6 to 24 months on the BSID-II-MS raw
score seemed more reliable, but after correction for norm data, the PDI and to a
lesser extent the classification became highly unstable. This has clinical implications
when reporting BSID-II-MS outcomes to parents, guiding intervention and in
randomized trials. The factors IVH, male sex and, where possible, height and
motor quality should be equally divided over the cohorts in preterm infants to
avoid score differences at the beginning of an intervention study, and maternal
education should be included in the randomization.
Acknowledgements
We thank all of the participating infants and their parents, the staff of the
Neonatal Intensive Care Unit and the Department of Pediatric Physical Therapy.
Special thanks are given for the secretarial assistance from both the Departments
of Neonatology and Pediatric Physical Therapy.
Competing interests
None
Funding
None
62
Unstable longitudinal motor performance in preterm infants from 6 to 24 months
References
(1) WHO. International Classification of Functioning, Disability and Health, Children and Youth Version
(ICF-CY). Geneva: World Health Organization; 2007.
(2) Vohr BR, Wright LL, Poole WK, McDonald SA. Neurodevelopmental outcomes of extremely low
birth weight infants <32 weeks’ gestation between 1993 and 1998. Pediatrics 2005; 116: 635-643.
(3) Wood NS, Costeloe K, Gibson AT, Hennessy EM, Marlow N, Wilkinson AR. The EPICure study:
associations and antecedents of neurological and developmental disability at 30 months of age
following extremely preterm birth. Arch Dis Child Fetal Neonatal Ed 2005; 90: F134-F140.
(4) Arpino C, Compagnone E, Montanaro ML, Cacciatore D, De LA, Cerulli A et al. Preterm birth and
neurodevelopmental outcome: a review. Childs Nerv Syst 2010; 26: 1139-1149.
(5) Allen MC. Neurodevelopmental outcomes of preterm infants. Curr Opin Neurol 2008; 21: 123-128.
(6) de Kieviet JF, Piek JP, arnoudse-Moens CS, Oosterlaan J. Motor development in very preterm and very
low-birth-weight children from birth to adolescence: a meta-analysis. JAMA 2009; 302: 2235-2242.
(7) Piek JP, Dawson L, Smith LM, Gasson N. The role of early fine and gross motor development on
later motor and cognitive ability. Hum Mov Sci 2008; 27: 668-681.
(8) van Haastert I, de Vries LS, Helders PJ, Jongmans MJ. Early gross motor development of preterm
infants according to the Alberta Infant Motor Scale. J Pediatr 2006; 149: 617-622.
(9) Jeng SF, Lau TW, Hsieh WS, Luo HJ, Chen PS, Lin KH et al. Development of walking in preterm and
term infants: age of onset, qualitative features and sensitivity to resonance. Gait Posture 2008; 27:
340-346.
(10) Fallang B, Hadders-Algra M. Postural behavior in children born preterm. Neural Plast 2005; 12: 175-182.
(11) Coolman RB, Bennett FC, Sells CJ, Swanson MW, Andrews MS, Robinson NM. Neuromotor
development of graduates of the neonatal intensive care unit: patterns encountered in the first two
years of life. J Dev Behav Pediatr 1985; 6: 327-333.
(12) Piper MC, Mazer B, Silver KM, Ramsay M. Resolution of neurological symptoms in high-risk infants
during the first two years of life. Dev Med Child Neurol 1988; 30: 26-35.
(13) Harris SR, Megens AM, Backman CL, Hayes VE. Stability of the Bayley II scales of infant
development in a sample of low-risk and high-risk infants. Dev Med Child Neurol 2005; 47: 820-823.
(14) Hack M, Wilson-Costello D, Friedman H, Taylor GH, Schluchter M, Fanaroff AA. Neurodevelopment and predictors of outcomes of children with birth weights of less than 1000 g: 1992-1995.
Arch Pediatr Adolesc Med 2000; 154: 725-731.
(15) Vohr BR, Wright LL, Dusick AM, Mele L, Verter J, Steichen JJ et al. Neurodevelopmental and
functional outcomes of extremely low birth weight infants in the National Institute of Child Health
and Human Development Neonatal Research Network, 1993-1994. Pediatrics 2000; 105: 1216-1226.
(16) Erikson C, Allert C, Carlberg EB, Katz-Salamon M. Stability of longitudinal motor development in
very low birthweight infants from 5 months to 5.5 years. Acta Paediatr 2003; 92: 197-203.
(17) Bayley N. Bayley scales of infant development, second edition. San Antonio: The Psychological
Corporation; 1993.
(18) Provost B, Crowe TK, McClain C. Concurrent validity of the Bayley scales of infant development II
motor scale and the Peabody developmental motor scales in two-year-old children. Phys Occup
Ther Pediatr 2000; 20: 5-18.
(19) Westera JJ, Houtzager BA, Overdiek B, van Wassenaer AG. Applying Dutch and US versions of the
BSID-II in Dutch children born preterm leads to different outcomes. Dev Med Child Neurol 2008;
50: 445-449.
(20) Wood E, Rosenbaum P. The gross motor function classification system for cerebral palsy: a study
of reliability and stability over time. Dev Med Child Neurol 2000; 42: 292-296.
(21) Papile LA, Burstein J, Burstein R, Koffler H. Incidence and evolution of subependymal and intraventricular hemorrhage: a study of infants with birth weights less than 1,500 gm. J Pediatr 1978; 92:
529-534.
(22) de Vries LS, Dubowitz LM, Dubowitz V, Kaiser A, Lary S, Silverman M et al. Predictive value of
cranial ultrasound in the newborn baby: a reappraisal. Lancet 1985; 2: 137-140.
63
3
Chapter 3
(23) de Vries LS, Regev R, Pennock JM, Wigglesworth JS, Dubowitz LM. Ultrasound evolution and later
outcome of infants with periventricular densities. Early Hum Dev 1988; 16: 225-233.
(24) Sonnander K, Claesson M. Predictors of developmental delay at 18 months and later school
achievement problems. Dev Med Child Neurol 1999; 41: 195-202.
(25) Thompson RJ, Jr., Gustafson KE, Oehler JM, Catlett AT, Brazy JE, Goldstein RF. Developmental
outcome of very low birth weight infants at four years of age as a function of biological risk and
psychosocial risk. J Dev Behav Pediatr 1997; 18: 91-96.
(26) Badr LK, Bookheimer S, Purdy I, Deeb M. Predictors of neurodevelopmental outcome for preterm
infants with brain injury: MRI, medical and environmental factors. Early Hum Dev 2009; 85:
279-284.
(27) Cooke RW, Foulder-Hughes L. Growth impairment in the very preterm and cognitive and motor
performance at 7 years. Arch Dis Child 2003; 88: 482-487.
(28) Twisk JW. Longitudinal data analysis. A comparison between generalized estimating equations
and random coefficient analysis. Eur J Epidemiol 2004; 19: 769-776.
(29) Cohen J. A power primer. Psychol Bull 1992; 112: 155-159.
(30) Darrah J, Hodge M, Magill-Evans J, Kembhavi G. Stability of serial assessments of motor and
communication abilities in typically developing infants−implications for screening. Early Hum Dev
2003; 72: 97-110.
(31) Roze E, Meijer L, Van Braeckel KN, Ruiter SA, Bruggink JL, Bos AF. Developmental trajectories
from birth to school age in healthy term-born children. Pediatrics 2010; 126: e1134-e1142.
(32) Largo RH, Fischer JE, Rousson V. Neuromotor development from kindergarten age to adolescence:
developmental course and variability. Swiss Med Wkly 2003; 133: 193-199.
(33) Largo RH, Caflisch JA, Hug F, Muggli K, Molnar AA, Molinari L. Neuromotor development from 5 to
18 years. Part 2: associated movements. Dev Med Child Neurol 2001; 43: 444-453.
(34) Hemgren E, Persson K. Quality of motor performance in preterm and full-term 3-year-old children.
Child Care Health Dev 2004; 30: 515-527.
(35) Fallang B, Oien I, Hellem E, Saugstad OD, Hadders-Algra M. Quality of reaching and postural
control in young preterm infants is related to neuromotor outcome at 6 years. Pediatr Res 2005; 58:
347-353.
(36) Janssen AJ, Nijhuis-van der Sanden MW, Akkermans RP, Tissingh J, Oostendorp RA, Kollee LA. A
model to predict motor performance in preterm infants at 5 years. Early Hum Dev 2009; 85:
599-604.
(37) Visser GH, Eilers PH, Elferink-Stinkens PM, Merkus HM, Wit JM. New Dutch reference curves for
birthweight by gestational age. Early Hum Dev 2009; 85: 737-744.
64
Unstable longitudinal motor performance in preterm infants from 6 to 24 months
3
65
Part II
Chapter 4
A model to predict motor performance
in preterm infants at 5 years
Anjo JWM Janssen MSc, PT1,3, Maria WG Nijhuis-van der Sanden PhD, PT 1,3,
Reinier P Akkermans MSc 2, Joke Tissingh MSc, PT 1,3, Rob AB Oostendorp
PhD, PT 3,5, Louis AA Kollée PhD, MD 4
Radboud University Nijmegen Medical Centre, Nijmegen,Department of Pediatric Physical Therapy,
The Netherlands
2
Radboud University Nijmegen Medical Centre, Nijmegen, Scientific Institute for Quality of
Healthcare,
Information Processing and Statistical Support, The Netherlands
3
Radboud University Nijmegen Medical Centre, Nijmegen, Scientific Institute for Quality of
Healthcare,
Research Centre of Allied Health Sciences, The Netherlands
4
Radboud University Nijmegen Medical Centre, Nijmegen, Department of Pediatrics, The
Netherlands
5
Dutch Institute of Allied Health Care, Amersfoort, The Netherlands
1
Early Hum Dev. 2009; 85: 599-604
Chapter 4
Abstract
Background: Approximately 60% of preterm infants who are assessed at 5
years for motor performance in a standardized multidisciplinary follow-up
program are found to have normal results, indicating that, for these children,
routine motor assessment at this age is unnecessary.
AIM: To improve the efficiency of our follow-up practice for motor assessment
by developing a model to predict motor performance of preterm infants at 5
years with a maximal sensitivity (≥ 90%).
Study design: Longitudinal design.
Subjects: We included preterm infants (n = 371) with a gestational age of ≤ 32
weeks; children with severe disabilities were excluded.
Outcome measures: The Movement Assessment Battery for Children (M-ABC)
at 5 years with ‘delayed’ motor performance (< 15 percentile) was the dependent
variable. As factors in the model, we used twenty neonatal risk factors, the
maternal education level, the Motor Scale and the Behavior Rating Scale (BRS)
of the Bayley Scales of Infant Development, 2nd edition, at 2 years.
Results: Binary logistic regression analysis revealed that the prediction model
(n = 345) reached a sensitivity of 94%. Five factors contributed significantly (p
< 0.05) to the model: a Motor Scale PDI < 90 and a BRS ‘motor quality’ < 26
percentile, and the neonatal risk factors gestational age < 30 weeks, male
gender and intra-ventricular hemorrhage.
Conclusion: The prediction model can improve the efficiency of follow-up
practice for motor assessment by 37% at 5 years. Applying this model, we would
not have assessed 129 children and would have missed six children.
70
A model to predict motor performance in preterm infants at 5 years
Introduction
Regular and long-term follow-up after a preterm birth is important, but there is
currently no international guideline for such assessments.1;2 The Dutch Neonatal
Follow-up Study Group recommends multidisciplinary assessments at 6 and 12
months and 2 and 5 years. The pediatric physical therapist assesses motor
performance to evaluate neonatal care and to determine whether pediatric physical
therapy intervention is indicated. For about 60% of preterm infants, motor performance
is normal at 5 years,3;4 indicating that the routine follow-up of motor assessment at
this age is unnecessary in these children. A tool by which to predict normal and
delayed motor performance in preterm infants would improve the efficiency of the
follow-up program. It has been suggested that prediction can be improved by
taking both the medical history and environmental factors into account.5;6
The prediction of motor performance in infants based on motor assessment in the
first year of life has been documented.6-10 Given the infants young age, such tests
are designed to primarily determine the integrity of the central nervous system;
consequently, any assessment of the influence of task constraints and
environmental factors is intrinsically limited. Early prediction is also complicated
by the spontaneous recovery of suspect or abnormal motor signs observed in
some preterm infants during the first two years of life.11-15 The large random
variation in longitudinal motor performance during a child’s first two years also
negatively affects prediction.4;5;16 For these reasons, we have chosen to develop a
prediction model based on data obtained at a corrected age of 2 years.
Prediction of developmental outcome based on one overall dimension is not
advisable, because motor, cognitive and behavioral outcome are not inter-related
in all children. Insights into the domain that shows delayed development is
important for an adequate intervention. Moreover, motor and cognitive outcomes
are associated with different sets of risk factors. As motor performance has been
less intensively investigated than cognitive development, we focused on motor
performance. We wanted to reach maximal sensitivity to prevent that children
who need intervention will be missed.
The aim of our study was to develop a prediction model for delayed motor
performance in children at 5 years that has a sensitivity of at least 90%. This
model was developed using data sets of preterm infants based on the outcome of
the assessment (motor and behavior scale) at the corrected age of 2 years,
neonatal risk factors and maternal education.
71
4
Chapter 4
Methods
The motor assessments used in this prospective longitudinal study constitute
part of a standardized multidisciplinary follow-up program for preterm infants at
the Radboud University Nijmegen Medical Centre, the Netherlands. The children
were assessed by two teams, of eight highly trained and experienced pediatric
physical therapists each, who were blinded to the medical histories.
The study cohort consisted of preterm infants born between January 1996 and
December 2001 at a gestational age (GA) of ≤ 32 weeks who had been admitted
to the neonatal intensive care unit. Of these preterm infants, children who had
been assessed at the corrected age of 2 years and at the chronological age of
5 years were considered for enrollment in the study. Children with known
chromosomal disorders, neuromuscular diseases or cerebral palsy (CP) and
children unable to complete the test were excluded. Parents and/or caregivers
were present throughout the assessment procedures.
The standard assessment protocol conforms to national consensus standards
on neonatal follow-up. The medical ethical committee stated that approval was
not required since the protocol is part of accepted medical practice.
Instruments
Assessment at a corrected age of 2 years
Motor performance was assessed using the Bayley Scales of Infant Development,
2nd edition (BSID-II). The Motor Scale and the Behavior Rating Scale (BRS) were
recorded.17 The BSID-II has been shown to be a comprehensive and appropriate
instrument for assessing motor performance based on content, construct and
concurrent validity.18;19 Its predictive validity increases with increasing age of the
child after 2 years.17 We used the American norm references since the Dutch
norm references were not available at the time we started our study. 20
The Motor Scale assesses gross and fine motor skills. Raw scores are
recalculated into the Psychomotor Developmental Index (PDI; mean 100 [SD
15]; range 50-129) using norm-referenced tables. The PDI scores are transformed
into the following classifications: scores ≤ 69 (-2 SD) indicate ‘significantly
delayed’; scores 70–84 (-1 SD) indicate ‘mildly delayed’; scores 85–114 indicate
‘within normal limits’; scores ≥ 115 (> +1 SD) are classified as ‘accelerated’
motor performance.
72
A model to predict motor performance in preterm infants at 5 years
The BRS assesses the qualitative aspects of the child’s behavior during testing
(‘test-taking behavior’) and comprises 30 questions, each scored on a 5-point
scale. When a 2-year-old child is being assessed, the questions are related to
three subscales: (1) orientation/engagement, (2) emotional regulation and (3)
motor quality. The total and subscale scores are transformed into percentiles.
Scores ≤ 10 th percentile reflect ‘non-optimal’ behavior; scores 11th –25th
percentile, ‘questionable’ behavior; scores ≥ 26 th percentile, behavior ‘within
normal limits’.
Assessment at 5 years of age
The Movement Assessment Battery for Children (M-ABC) was designed to
identify impairments of motor function in children aged 4–12 years. 21 Test results
are expressed as a total motor impairment score using age-related norm tables
and then divided into three classes. 21 Scores ≤ 5th percentile are considered a
movement difficulty, scores between the 6 th and 14th percentile are considered
‘at risk’; scores ≥ 15th percentile are considered ‘normal’. The inter-rater reliability
and validity of the M-ABC is considered to be sufficient. 21;22 The Dutch version of
the M-ABC does not differ from the American version in terms of norm scores. 23
Factors included
The Dutch Neonatal Follow-up Study Group has developed a database of
documented risk factors. We selected, based on a literature search and clinical
experience, 20 of the registered factors expected to influence motor performance.
During the neonatal period, cranial ultrasound was performed soon after
admission, at least twice in the first week of life and once weekly thereafter. We
added maternal education as an indicator of social economic status24-26 and
grouped this indicator into three categories: low (no/elementary school/
vocational training), average (lower and upper general secondary education/
secondary vocational education) or high (pre-university education/bachelor
degree or higher) (Table 2).
Statistical analysis
All outcomes at 2 and 5 years were dichotomized in ‘normal’ and ‘delayed’
performance. Delayed performance was indicated by: (1) PDI scores < 85; (2)
73
4
Chapter 4
BRS scores (all subscale scores) < 26 th percentile; (3) a M-ABC total score <
15th percentile. The outcome of the M-ABC was the dependent variable, and
cross-tabulations were performed. We used the Spearman’s rank correlation
coefficient (rs) to estimate the correlation between categories of outcome
variables. To demarcate the most adequate PDI cutoff score for delayed or
normal motor performance, we computed the balance between sensitivity and
specificity using a receiver operating characteristics (ROC) curve. The area
under the curve (AUC) was used as an estimate of diagnostic accuracy.
We calculated odds ratios (ORs) by univariate regression and performed a
backward binary logistic stepwise regression. The final model included only
factors with p values ≤ 0.05. The probability of ‘delayed’ motor performance for
each child was calculated based on the regression coefficients from the
multivariate logistic model. The predicted probability cutoff point was chosen in
the model when the ROC curve reached a sensitivity of 90% or more. We then
calculated the number of children we would not have asked to come for the
5-year motor assessment and the number of children we would have missed.
Finally, the positive and negative predictive values were computed. Where
appropriate, 95% confidence (CI) are given. Two-sided p values ≤ 0.05 were
considered statistically significant. SPSS 16.0 was used for data analysis.
Results
The flow chart of the cohort and children assessed at 2 and 5 years of age is
given in Fig.1. A total of 371 children, (207 boys, 56%) were enrolled in the
analysis with a mean birth weight of 1221.0 grams (SD 326.6) and a gestational
age of 29+6 weeks (SD 2+0). The mean corrected age at 2 years was 29 months
(SD 4.9), and the mean chronological age at 5 years was 64 months (SD 2.3).
Table 1 shows the number of children in each class of the BSID-II at 2 years and
the M-ABC at 5 years. The correlations between the classes of the M-ABC and
the BSID-II Motor Scale (r s = 0.26), the M-ABC and BRS ‘total score’ (rs = 0.17),
and the M-ABC and BRS subscales ‘emotional regulation’ (rs = 0.20) and ‘motor
quality’ (rs = 0.27) were significant, but poor.
The ROC curve for different PDI cutoff points between 70 and 105, at 5-point
intervals was examined and is shown in Fig. 2. The AUC for the cutoff point 85
was 0.65 (CI 0.58–0.71) and for 90 the AUC was 0.66 (CI 0.60 – 0.71). The PDI
74
A model to predict motor performance in preterm infants at 5 years
Figure 1 Flow chart of children assessed at 2½ and 5 years of age
n = 787
1996-2001 cohort
Excluded:
112 Random not invited
11 Disability
81 Untraceable
94 Refused
23 Cancelled/did not show up
Motor assessment at 2 years
n =466
Excluded:
12 Disability
4 Wrong age band
9 Uncooperative
3 language problems
1 ill during testing
Eligible motor assessment at 2 years
n = 321
n = 29
4
n = 437
No motor assessment at 5 years n = 56
Motor assessment at 5 years
n = 381
n = 10
Excluded:
6 Cerebral Palsy,
2 visual disability
2 ill during testing,
Eligible motor assessment at 2 and 5 years n =371
Missing data independent variable
maternal education
n = 26
Complete data set for
multivariate logistic model
n = 345
75
Chapter 4
Table 1 Motor Performance outcome at a corrected age of 2 years
(BSID-II) and at chronological age of 5 years (M-ABC) for n = 371
and n = 345*
2 years
5 years
< 15 percentile
≥ 15 percentile
Total
n
(%)
n
(%)
n
(%)
PDI < 85
70/168
(19)
98/168
(26)
168/371
(45)
PDI ≥ 85
Total
36/203
(10)
167/203
(45)
203/371
(55)
106/371
(29)
265/371
(71)
371
(100)
(18)
(10)
(28)
90/152
160/193
250/345
(26)
(46)
(72)
152/345
193/345
345
(44)
(56)
(100)
Maternal education missing
PDI < 85
PDI ≥ 85
Total
62/152
33/193
95/345
*maternal education missing (n = 26)
Figure 2 ROC curve based on eight PDI cut-off points between 70-105
1
95
Sensitivity
0,8
100
105
90
85
0,6
80
0,4
70
0,2
Cut-off point
PDI
75
0
0
0,2
0,4
0,6
1-Specificity
76
0,8
1
A model to predict motor performance in preterm infants at 5 years
with a cutoff point of < 85 predicted motor performance at 5 years with a
sensitivity of 66% and a specificity of 63%. The sensitivity and specificity for a
PDI cutoff point < 90 was 79% and 52%, respectively. Based on these results,
we decided to define a PDI < 90 as the cutoff point for ‘normal’ and ‘delayed’
motor performance at 2 years.
The BRS subscales correlated significantly with the BRS ‘total score’, ‘orientation/
engagement’ (r s = 0.69) and ‘emotional regulation’ (rs = 0.67) and ‘motor
quality’ (rs = 0.26). To avoid multicollinearity, we included only the ‘total score’
and the subscale ‘motor quality’. Among all of the risk factors tested, only the
Apgar score and endotracheal intubation showed a correlation rs > 0.4 (rs =
0.47). Therefore, we included all of the risk factors in the analysis. Table 2 shows
the distribution of risk factors and the results of the univariate and multivariate
binary logistic regression analysis for the association with delay in motor
performance at 5 years. Multivariate analysis (n = 345) revealed five contributing
factors: PDI < 90; subscale ‘motor quality’ < 26 percentile; GA < 30 weeks;
male gender; intra-ventricular hemorrhage (IVH).
Our calculation of the probability that a follow-up evaluation is indicated for an
individual child is based on the coefficients from the logistic regression equation,
with the formula:
P(x) = 1
1 + eZ
where Z = -3.124 + 0.855 (IVH) + 0.662 (male gender) + 0.868 (GA < 28 weeks)
+ 1.234 (GA ≥ 28 < 30 weeks) + 1.152 (PDI < 90) + 0.809 (motor quality < 26
percentile).
Based on the results of the ROC curve of the predictive probability of the five
factors we defined a cutoff point of 0.19 with a sensitivity and specificity of 94%
and 50%, respectively (AUC: 0.78; CI 0.73–0.83). Children scoring below the
predicted probability cut-off point should not be invited for follow-up. According
to this model (n = 345), 129 children (36.7%) would not have been invited for
motor performance follow-up, indicating that the efficiency of the follow-up can
be improved by 37%. We would have missed six (6.3%) of the 95 children who
had a delayed motor performance at 5 years, and we would have tested 127
children with normal motor performance in an unnecessary assessment (36.8%).
The positive and negative predictive value of the model is 41% and 95%,
respectively.
77
4
78
7
1.9
4.0
24
15
Neonatal convulsions
Idiopathic respiratory distress syndrome
Retinopathy of prematurity
200
6.5
22
Ischemic cerebral abnormality
+++
12
53.9
5.9
3.2
12.4
46
Intraventricular hemorrhage++
Post-hemorrhage ventricle dilatation
54.7
25.1
55.5
4.3
203
93
206
32.1
16
Meningitis/encephalitis/ventriculitis
Sepsis+
Multiple births
Caesarean section
Apgar missing
119
119
Endotracheal intubation after birth
Apgar score < 7 at 5 min
32.1
28.8
107
Neonatal factors
47.4
176
GA ≥ 30
Birth weight < 1000 grams
33.2
19.4
55.8
123
72
207
%
GA ≥ 28 < 30 weeks
GA < 28 weeks
Male gender
Patient characteristics
n
1.3 (0.8-2.1)
3.0 (1.1-8.5)*
2.3 (1.0-5.2)
0.7 ( 0.3-2.0)
a
a
a
a
a
a
2.4 (1.2-4.7)*
0.855
1.234
0.868
2.4 (1.2-4.7)*
3.4 (1.9-6.2)*
0.622
1.9 (1.1-3.2)*
B
a
OR (95% CI)
Multivariate
a
3.7 (1.2-11.6)*a
3.2 (1.6-5.5)*
1.1 (0.4-3.4)
1.5 (1.0-2.5)
0.8 (0.5-1.3)
0.7 (0.5-1.2)
0.9 (0.6-1.5)
1.1 (0.7-1.8)
1.8 (1.1-2.9)*
3.4 (2.0-5.8)*
3.2 (1.7-6.0)*
2.0 (1.3-3.5)*
OR (95% CI)
Univariate
regression analysis for motor performance at age 5 years (n = 345)
Table 2 Distribution of risk factors and results of univariate regression (n = 371) and results of binary multivariate logistic
Chapter 4
33
Congenital malformation
121
BRS ´motor quality´ < 26
26
Missing
7.0
17.5
54.2
21.0
32.6
28.6
57.1
8.9
5.1
69.3
7.6
11.9
a
a
a
a
1.7 (0.8-3.6)
4.1 (2.4-6.9)*
2.2 (1.4-3.5)*
3.5 (2.2-5.5)*
2.4 (1.1-5.2)*
1.5 (0.7-2.9)
a
a
1.0 (1.0-2.7)
2.4 (0.9-6.0)
a
a
1.7 (0.8-3.7)
a
2.6 (1.4-4.9)*
++
+
Sepsis included both culture-proven and probable sepsis based on biochemical and clinical signs
IVH included all grades according to Papile et. al. (J Ped 1978;92: 529-534)
+++
ROP included all stages of the disease
++++
Defining chronic lung disease as mechanical ventilation or oxygen at 36 weeks postmenstrual age
* p < 0.05
a
Factors (18) entered in binary multivariate logistic regression analysis p < 0.3
B = regression coefficient for each variable in the equation
65
201
High
Average
Low
78
106
BRS ´total score´ < 26
Maternal Education
212
Motor Scale PDI < 90
Test related factors BSID-II
19
Necrotizing enterocolitis
257
28
Pneumonia/pneumothorax/atelectasis
Hyperbilirubinemia
44
Chronic lung disease++++
2.3 (1.3-3.8)*
3.2 (1.8-5.7)*
0.809
1.152
A model to predict motor performance in preterm infants at 5 years
4
79
Chapter 4
Figure 3 Algorithm to indentify children for no retest or retest of motor
performance at 5 years of age
No retest
1 yes
Is the PDI < 90?
Is the Motor Quality< 26 percentile?
Does an IVH exist?
GA < 30 weeks?
Male gender?
No/Yes
No/Yes
No/Yes
No/Yes
No/Yes
Male Gender and motor quality < 26 percentile
Male Gender and IVH
Male Gender and GA < 28 weeks
Motor quality < 26 percentile and IVH
Except
2 yes
Retest
3 yes
Retest
Three of the missed children had combinations of risk factors that fell outside of
the constraints of the model. One child had a necrotizing enterocolitis (NEC),
hyperbilirubinemia, and idiopathic respiratory distress syndrome (IRDS), sepsis
and a birth weight of < 1000 grams; the second child had sepsis, meningitis and
IRDS; the third child had sepsis, IRDS and hyperbilirubinemia, in combination
with a birth weight of < 1000 grams.
An algorithm for clinical identification of children for no retest or retest is shown
in Fig. 3. This figure shows that children with one factor and four combination of
two factors should not be retested.
Discussion
We developed a prediction model whether in preterm infants at 2 years a routine
motor performance assessment is indicated at 5 years. Our model, which
incorporates the BSID-II Motor Scale (PDI <90), the BRS subscale ‘motor
quality’ (< 26 percentile) and the risk factors GA < 30 weeks, male gender and
IVH, has a sensitivity of 94% in predicting delayed motor performance at 5
80
A model to predict motor performance in preterm infants at 5 years
years. We chose maximal sensitivity to reduce the risk of missing children with
motor performance problems. Sensitivity improved by shifting the cutoff point
of the BSID-II Motor Scale. Using our prediction model, 129 (37%) children
would not have been assessed at 5 years, indicating that the model provides a
substantial improvement in the efficiency of motor performance assessment in
the follow-up program; in comparison, six children who required testing would
not have received it.
In general, using a clinical prediction model has its limitations. Risk factors
known to influence motor performance, but not included in this model, should
be taken into account when evaluating the question whether children should be
offered a motor performance evaluation.
Although 112 of the total number of children eligible for entry to the study were
not invited to participate, the findings can still be generalized because the group
did not differ significantly from the children that participated in terms of risk
factors and outcome at 5 years. However, attrition bias may be present and may
have influenced the findings that maternal education was not a risk factor, since
the proportion of parents with a low education could be overrepresented in the
non-participants.
We used the U.S. version and norm references of the BSID-II. Comparing the
PDI outcomes for the children using the U.S. or Dutch versions, the difference
was significant but low: mean difference (paired t-test) of –1.5 (SD 4.7, 95% CI
–1.9 to –1.0). The absolute difference is in line with that reported by Westera et
al., with the exception that our PDI scores for the Dutch version were higher. 27
Prediction depends on the test and cutoff point selected, the time interval of
prediction, the gold standard and the factors included. We found that the Motor
Scale was more predictive than the Mental Scale of the BSID-II for motor
performance outcome at 5 years. Therefore, we included only the Motor Scale.
We shifted the cutoff point of the BSID-II Motor Scale to 90. Although, the official
cutoff point of the BSID-II, 85, revealed the same five significant factors, the
sensitivity at a cutoff point of 85 was 4% lower than that with a cutoff point of 90.
There was no substantial difference in the AUC. However, we wanted to reach
maximal sensitivity to prevent that children who need intervention, will be
missed.
It has been possible to predict motor performance from the first year of life
onwards for different time intervals. Short-term prediction (≤ 1 year) has been
reported at the same sensitivity as that achieved by our model. However, the
81
4
Chapter 4
gold standard of the pediatrician’s classification of normal, suspect or abnormal
motor performance is open to discussion.9 Two studies have shown that
long-term prediction based on motor tests in the first year of life has a lower
sensitivity.6;8
A literature search revealed the complete absence of studies predicting motor
performance from the BSID-II in children at 2 years. One study predicted motor
performance in full-term healthy infants from the BSID-I and the Bruininks-Oseretsky Test of Motor Proficiency (BOTMP). 28 Because of the poor gold standard
of the BOTMP, this study provided only weak evidence that prediction is better
for female infants. Separate analysis for male and female infants confirmed the
better prediction for female infants (data not shown). Barnet et. al. found that the
Griffiths Mental Development Scale at 2 years had a sensitivity of 67% and a
specificity of 100% (n = 45) to predict outcome on the M-ABC at 5–6 years. 29
They included full-term infants with abnormal motor signs who were likely to
have neurological impairments. In contrast with some preterm infants, who
show spontaneous recovery of abnormal motor signs during the first two years
of life.11-14 This difference can clarify why an abnormal score in full-term infants
can be highly predictive of poor motor performance at 5 years. The Griffiths
Scale has been shown to have a higher correlation with the M-ABC than with the
WPPSI intelligence test. 29 Research is necessary to establish whether the
Griffiths Scale is capable of reaching a better predictive power than the BSID-II
Motor Scale in preterm infants.
Erikson et. al. investigated the influence of risk factors on the stability of
longitudinal motor performance.4 These authors based their definition of ‘Stable
motor performance over time’ on an outcome that changed in fewer than three
of four defined classes. Motor performance in 50% of the children was
characterized by longitudinal variability, and periventricular leucomalacia (PVL)
and body weight (BW) were found to contribute significantly. A comparison
between our study and that of Erikson et al. is difficult principally because that
latter included children who performed stably but delayed and those who
performed stably but well in the stable group; in contrast we had delayed
performers in the delayed group and good performers in the normal group.
We compared the children who performed normally at 2 years and delayed at 5
years to the children who were first delayed and then normal at 5 years. The
group of children who showed deterioration were found to include more children
with GA < 30 weeks sepsis, and IVH; in addition, the children were tested at a
82
A model to predict motor performance in preterm infants at 5 years
younger age (corrected age < 31 months), and a higher proportion had hyperbilirubinemia. This finding provides a reason to test these children after 2 years.
Furthermore, the children who improved had a delayed BRS total- and subscale
scores of orientation/engagement, emotional regulation, motor quality; they
also had a low APGAR score. In addition, more children in this group had been
delivered by caesarean section. The cooperativeness of the child during testing
at 2 years influences the motor outcome and the BRS helps to interpret the
outcome of the BSID-II Motor Scale.
Of the factors contributing to our results, GA has been found to be a significant
predictive factor of the social consequences of preterm birth;30 in addition,
decreasing GA is correlated with an increased risk for minor neurological
dysfunction31 and linked to delayed M-ABC performance. 32 Male gender is a risk
factor for delayed motor performance,31;33;34 and IVH correlates with CP35 and
with minor neurological dysfunction.31;36 We excluded children with CP but still
found that IVH, the milder grades, is important in predicting delayed motor
performance. Motor quality is not investigated in most studies, but it would
appear to be a promising predictive factor.36
Unexpectedly, we were unable to confirm whether maternal education has a
significant influence in predicting motor performance, although it has been
proposed that motor performance ‘emerges from the interactions between the
individual, the task and the environment’.37 In the Netherlands, maternal
education usually is used for defining social economic status.38 Apart from the
possible attrition bias, we can hypothesize that maternal education has already
influenced motor outcome on the BSID-II. Even more remarkable was the poor
correlation between motor outcome at 2 and 5 years. It is possible that the tests
used in our study measure a different construct: at 2 years, the outcomes of the
tests are more dependent on the cooperation and orientation of the child to the
task and examiner, than at older ages. It is also possible that overestimation of
the gross motor skills exist because of lack of division in the fine and gross
motor scale. Therefore, the results of the assessment at 2 years may be a better
predictor of problems relating to participation in subsequent age-related daily
life activities and not only of motor performance problems.
This was a single-center study in which a standardized follow-up protocol was
closely followed. However, we cannot exclude the possibility of differences in
perinatal care, data recording and test conditions in other centers. Consequently,
this model needs to be validated before it can be generalized.
83
4
Chapter 4
The method described can be implemented in future research, including
longitudinal assessments in combination with risk factors. When the prediction
factors are known for each domain separately, the next step may be to develop
a combined algorithm.
Acknowledgements
We thank all of the participating children and their parents, the staff of the
Neonatal Intensive Care Unit and the Department of Pediatric Physical Therapy.
Special thanks are given to the secretarial assistance from both the Department
of Neonatology and Pediatric Physical Therapy.
84
A model to predict motor performance in preterm infants at 5 years
References
1
2
3
4
5
6
7
8
9
10 11 12 13 14 15 16 17 18 19 20 21 Johnson S, Marlow N. Developmental screen or developmental testing? Early Hum Dev 2006; 82:
173-83.
Allen MC. Neurodevelopmental outcomes of preterm infants. Curr Opin Neurol 2008; 21: 123-8.
De Kleine MJ, den Ouden AL, Kollée LA, Nijhuis-van der Sanden MW, Sondaar M, van
Kessel-Feddema BJ, et al. Development and evaluation of a follow up assessment of preterm
infants at 5 years of age. Arch Dis Child 2003; 88: 870-5.
Erikson C, Allert C, Carlberg EB, Katz-Salamon M. Stability of longitudinal motor development in
very low birthweight infants from 5 months to 5.5 years. Acta Paediatr 2003; 92: 197-203.
Largo RH, Graf S, Kundu S, Hunziker U, Molinari L. Predicting developmental outcome at school
age from infant tests of normal, at-risk and retarded infants. Dev Med Child Neurol 1990; 32: 30-45.
Kolobe TH, Bulanda M, Susman L. Predicting motor outcome at preschool age for infants tested at
7, 30, 60, and 90 days after term age using the Test of Infant Motor Performance. Phys Ther 2004;
84: 1144-56.
Darrah J, Piper M, Watt MJ. Assessment of gross motor skills of at-risk infants: predictive validity
of the Alberta Infant Motor Scale. Dev Med Child Neurol 1998; 40: 485-91.
Flegel J, Kolobe TH. Predictive validity of the test of infant motor performance as measured by the
Bruininks-Oseretsky test of motor proficiency at school age. Phys Ther 2002; 82: 762-71.
Campbell SK, Kolobe TH, Wright BD, Linacre JM. Validity of the Test of Infant Motor Performance
for prediction of 6-, 9- and 12-month scores on the Alberta Infant Motor Scale. Dev Med Child
Neurol 2002; 44: 263-72.
Groen SE, de Blecourt AC, Postema K, Hadders-Algra M. General movements in early infancy
predict neuromotor development at 9 to 12 years of age. Dev Med Child Neurol 2005; 47: 731-8.
Coolman RB, Bennett FC, Sells CJ, Swanson MW, Andrews MS, Robinson NM. Neuromotor
development of graduates of the neonatal intensive care unit: patterns encountered in the first two
years of life. J Dev Behav Pediatr 1985; 6: 327-33.
Piper MC, Mazer B, Silver KM, Ramsay M. Resolution of neurological symptoms in high-risk infants
during the first two years of life. Dev Med Child Neurol 1988; 30: 26-35.
Wildin SR, Anderson A, Woodside M, Swank P, Smith K, Denson S, et al. Prediction of 12-month
neurodevelopmental outcome from a 6-month neurologic examination in premature infants. Clin
Pediatr (Phila) 1995; 34: 290-9.
de Vries AM, de GL. Transient dystonias revisited: a comparative study of preterm and term
children at 2 1/2 years of age. Dev Med Child Neurol 2002; 44: 415-21.
Aylward GP, Verhulst SJ. Predictive utility of the Bayley Infant Neurodevelopmental Screener
(BINS) risk status classifications: clinical interpretation and application. Dev Med Child Neurol
2000; 42: 25-31.
Darrah J, Hodge M, Magill-Evans J, Kembhavi G. Stability of serial assessments of motor and
communication abilities in typically developing infants--implications for screening. Early Hum Dev
2003; 72: 97-110.
Bayley N. Bayley Scales of Infant Development: second edition. San Antonio: The Psychological
Corporation; 1993.
Provost B, Crowe TK, McClain C. Concurrent validity of the Bayley Scales of Infant Development II
Motor Scale and the Peabody Developmental Motor Scales in two-year-old children. Phys Occup
Ther Pediatr 2000; 20: 5-18.
Provost B, Heimerl S, McClain C, Kim NH, Lopez BR, Kodituwakku P. Concurrent validity of the
Bayley Scales of Infant Development II Motor Scale and the Peabody Developmental Motor
Scales-2 in children with developmental delays. Pediatr Phys Ther 2004; 16: 149-56.
Van der Meulen B, Ruiter S, Spelberg H, Smrkovsky M. BSID-II-NL manuel. Lisse, Swets en
Zeitlinger. Lisse: Swets Test Publishers; 2002.
Henderson SE, Sugden DA. Movement Assessment Battery for Children. London: Psychological
Corporation, Harcourt Brace Jonanovich; 1992.
85
4
Chapter 4
22 Smits-Engelsman BC, Fiers MJ, Henderson SE, Henderson L. Interrater reliability of the Movement
Assessment Battery for Children. Phys Ther 2008; 88: 286-94.
23 Henderson SE, Sugden DA, Smits-Engelsman BCM. Movement Assessment Battery for Children,
Dutch Manuel.Lisse: Swets & Zeitlinger B.V.; 1998. p. 103-12.
24 Sonnander K, Claesson M. Predictors of developmental delay at 18 months and later school
achievement problems. Dev Med Child Neurol 1999; 41: 195-202.
25 Bradley RH, Corwyn RF. Socioeconomic status and child development. Annu Rev Psychol 2002;
53: 371-99.
26 Thompson RJ, Jr., Gustafson KE, Oehler JM, Catlett AT, Brazy JE, Goldstein RF. Developmental
outcome of very low birth weight infants at four years of age as a function of biological risk and
psychosocial risk. J Dev Behav Pediatr 1997; 18: 91-6.
27 Westera JJ, Houtzager BA, Overdiek B, van Wassenaer AG. Applying Dutch and US versions of the
BSID-II in Dutch children born preterm leads to different outcomes. Dev Med Child Neurol 2008;
50: 445-9.
28 MacCobb S, Greene S, Nugent K, O’Mahony P. Measurement and prediction of motor proficiency
in children using Bayley infant scales and the Bruininks-Oseretsky test. Phys Occup Ther Pediatr
2005; 25: 59-79.
29 Barnett AL, Guzzetta A, Mercuri E, Henderson SE, Haataja L, Cowan F, et al. Can the Griffiths
scales predict neuromotor and perceptual-motor impairment in term infants with neonatal
encephalopathy? Arch Dis Child 2004; 89: 637-43.
30 Moster D, Lie RT, Markestad T. Long-term medical and social consequences of preterm birth. N
Engl J Med 2008; 359: 262-73.
31 Arnaud C, ubisse-Marliac L, White-Koning M, Pierrat V, Larroque B, Grandjean H, et al. Prevalence
and associated factors of minor neuromotor dysfunctions at age 5 years in prematurely born
children: the EPIPAGE Study. Arch Pediatr Adolesc Med 2007; 161: 1053-61.
32 Cooke RW. Perinatal and postnatal factors in very preterm infants and subsequent cognitive and
motor abilities. Arch Dis Child Fetal Neonatal Ed 2005; 90: F60-F63.
33 Marlow N, Hennessy EM, Bracewell MA, Wolke D. Motor and executive function at 6 years of age
after extremely preterm birth. Pediatrics 2007; 120: 793-804.
34 Janssen AJ, Nijhuis-van der Sanden MW, Akkermans RP, Oostendorp RA, Kollee LA. Influence of
behaviour and risk factors on motor performance in preterm infants at age 2 to 3 years. Dev Med
Child Neurol 2008; 50: 926-31.
35 Roze E, Kerstjens JM, Maathuis CG, ter Horst HJ, Bos AF. Risk factors for adverse outcome in
preterm infants with periventricular hemorrhagic infarction. Pediatrics 2008; 122: e46-e52.
36 Schmidhauser J, Caflisch J, Rousson V, Bucher HU, Largo RH, Latal B. Impaired motor performance
and movement quality in very-low-birthweight children at 6 years of age. Dev Med Child Neurol
2006; 48: 718-22.
37 Shumway-Cook a., Woollacott MH. Motor control: issues and theories. Motor Control; Translating
research into clinical practice. Third ed. Philadelphia, Pennsylvania: Lippincott Williams & Wilkins;
2007. p. 3-20.
38 Driessen GWJM. Family and child characteristics, child-rearing factors, and cognitive competence
of young children. Early Child Dev Care 2003; 173: 323-39.
86
A model to predict motor performance in preterm infants at 5 years
4
87
Chapter 5
Comparing the predictive validity of two
infant motor tests for motor performance
at 5 years in preterm infants
Anjo JWM Janssen MSc, PPT1, Rob AB Oostendorp PhD, PT2, Reinier P
Akkermans MSc2, Katerina Steiner MD3, Louis AA Kollée PhD, MD3, Maria WG
Nijhuis-van der Sanden PhD, PPT2
Radboud University Nijmegen Medical Centre, Department of Pediatric Physical therapy,
Rehabilitation Nijmegen,
The Netherlands,
2
Radboud University Nijmegen Medical Centre, Scientific Institute for Quality of Healthcare,
Nijmegen, The Netherlands,
3
Radboud University Nijmegen Medical Centre, Department of Pediatrics, Nijmegen, The
Netherlands
1
Submitted for publication
Chapter 5
Abstract
AIMS: To compare the predictive validity of two motor tests used at 6 (t 0), 12
(t1), and 24 (t 2) months corrected age for motor performance at 5 years (t3).
Method: Longitudinal study involving 201 preterm infants (<32 weeks’ gestation,
106 boys, 95 girls) without known disabilities. Infants were assessed at t 0, t1,
and t 2 with the Bayley Scales of Infant Development-second edition, Motor
Scale (BSID-II-MS), Behavior Rating Scale (BSID-II-BRS) and its subscale motor
quality (MQ), at t 0 and t1 with the Alberta Infant Motor Scale (AIMS), and at t3 with
the Movement Assessment Battery for children-second edition (MABC-2-NL).
Normal versus abnormal scores were examined.
Results: The MABC-2-NL categorized 74 (37%) children as abnormal. The AIMS
≤10 percentile showed the best combination of sensitivity and specificity
(t1=65%, 68%). The BSID-II-MS ≤84 had the best sensitivity (t 0=79%, t1=93%,
t 2=63%) and the BSID-II-BRS ≤25 percentile had the best specificity (t 0=87%,
t1=90%, t 2=81%). At t 0 the combination AIMS ≤10/BSID-MS ≤69/BRS ≤25, at t1
the AIMS ≤10/BSID-MS ≤69/MQ ≤25 and at t 2 the BSID-MS ≤69/MQ ≤25 achieved
the best balance between sensitivity and specificity (t0=78%, 43%; t1=78%, 62%;
t 2=74.0%, 62%).
Interpretation: Combining tests improved prediction and the predictive validity
was lowest at 6 months.
90
comparing the predictive validity of two infant motor tests
Introduction
Preterm infants are at risk for a poor motor performance throughout childhood,1
and motor performance in early childhood tend to predict motor and cognitive
outcome at school-age. 2 Evaluation and prediction of motor performance in
preterm infants is important for the timely initiation of interventions. However,
few longitudinal follow-up studies included the predictive validity of motor tests.
To investigate the predictive validity of motor test before two years of life is
challenging, because motor performance is highly variable at this age.3-5 The
Bayley Scales of Infant Development, 2nd edition (BSID-II)6 and the Alberta Infant
Motor Scale (AIMS)7 are frequently used assessment tools. These tests are
reliable, validated, and have discriminative validity, but their predictive validity
was mainly studied on the short-term.8-10
The aim of this study was to compare the predictive validity of two motor tests
used in the first two years of life for prediction of motor performance at 5 years.
We determined the predictive validity of the BSID-II Motor Scale (BSID-II-MS) and
BSID-II Behavior Rating Scale (BSID-II-BRS) and its subscale motor quality at 6,
12 and 24 months corrected age (CA), the AIMS at 6 and 12 months CA, dichotomized
with different cut-off points (-1 SD, -2 SD), and various combinations of these tests
for motor performance at 5 years of age, as measured with the Movement
Assessment Battery for Children-second edition, Dutch version (MABC-2-NL).11
Method
In this prospective longitudinal cohort study, pediatric physical therapists assessed
motor performance at four different time points from 6 months to 5 years of age
as part of a standardized multidisciplinary follow-up program for preterm infants
at the Radboud University Nijmegen Medical Centre, the Netherlands. In
conformation to the declaration of Helsinki, medical ethical committee approval
was not required, since the protocol is part of accepted medical practice.
Participants
The study cohort comprised surviving infants born between March 2003 and
December 2005 at less than 32 weeks of gestation and who had been admitted
to the neonatal intensive care unit. Children with known disabilities, such as
91
5
Chapter 5
cerebral palsy (CP), syndromes, blindness, and deafness, were excluded
because the tests used are not appropriate for the prediction of these children’s
motor performance. All infants were assessed at the at 6 (t 0), 12 (t1), and 24 (t 2)
months CA and at the chronological age of 5 (t3) years. Parents and/or caregivers
were present during test procedures.
Instruments
Motor performance and behavior were assessed at t 0, t1, and t 2 using BSID-II-MS
and BSID-II-BRS.6 Based on content, construct, and concurrent validity, the
BSID-II has been shown to be a comprehensive and appropriate instrument
to asses motor performance.12 American norm references were used because
the Dutch version includes modifications to the test procedure that hamper
comparison with international studies.13
The BSID-II-MS assesses gross and fine motor skills. Raw scores (range 0–111)
are first recalculated into the Psychomotor Developmental Index (PDI; mean of
100 standard deviation [SD] 15) using norm-referenced tables. The PDI scores
are then classified into motor performance categories, with scores ≤69 (–2 SD)
indicating ‘significantly delayed’; 70–84 (–1 SD) indicating ‘mildly delayed’; 85–114
indicating ‘within normal limits’; ≥115 (+1 SD) indicating ‘accelerated’ motor
performance.
The BSID-II-BRS assesses the qualitative aspects of the infant’s behavior during
testing (‘test-taking behavior’) and comprises 30 questions, each scored on a
5-point scale. At the ages assessed, the questions are related to three subscales:
(1) orientation/engagement, (2) emotional regulation, and (3) motor quality
(MQ). The total raw score and subscale raw scores can be calculated and
transformed into percentiles; scores ≤10 th percentile reflect ‘non-optimal’
behavior; 11th –25th percentile reflect ‘questionable’ behavior; ≥26 th percentile
reflect behavior ‘within normal limits’. In this study, the total scores and the
subscale MQ scores were selected because the correlation at t 0, t1, and t 2
between the BSID-II-BRS total percentile score and the subscale percentiles
scores for orientation/engagement (rs= 0.76/0.76/0.77) and emotional regulation
(rs= 0.69/0.74/0.84) was higher than that for the subscale MQ (rs= 0.60/0.46/0.50).
Moreover, in a previous study of our group the subscale MQ showed predictive
validity for delayed motor performance.14
The AIMS measures gross motor performance in infants from birth to independent
walking and was recorded at t 0 and t1.7 The AIMS consists of four positional
92
comparing the predictive validity of two infant motor tests
subscales: prone, supine, sitting, and standing. The total raw score (range 0–58)
can be converted to a percentile rank. The reliability and content validity of the
AIMS are sufficient.15-18 A score of ≤5th percentile classifies motor performance
as ‘suspicious/abnormal’ and we defined a score of ≥6 th and ≤10 th percentile as
‘at risk’.
The MABC-2-NL was designed to identify and describe impairments of motor
function in children aged 3–16 years.11 In this test motor tasks are ordered into
three components: manual dexterity, aiming and catching, and balance. The
raw scores are converted to a standard score of the components and a total test
score (TTS), which can be transformed into percentile scores. Scores ≤5th
percentile reflect a significant movement difficulty and scores between the 6 th
and 16 th percentile reflect being ‘at risk’ of a movement difficulty. Scores of ≥17th
percentile reflect ‘unlikely to have a movement difficulty.’ The inter-rater reliability
and validity of the MABC-2 of age band 1 (3-6 years) are sufficient.19 The manual
of the MABC-2-NL also reports sufficient reliability and validity.11
The Gross Motor Function Classification System (GMFCS) was used to describe
the subgroup of infants with CP at t3. 20 It consists of a 5-level pattern recognition
system in which level I and II indicate that there is potential to walk, and levels III
to V represent an increasing limitation in the self-mobility of the infants.
Statistical analysis
Data were analyzed by using SPSS (version 16) and Simple Interactive Statistical
Analysis. 21 Cohort demographics, number, and percentages were described.
Means, SD, median percentages, and interquartile ranges describe the motor
performance outcome of the MABC-2-NL. Correlations between BSID-II-MS
(normal versus delayed performance PDI -1 SD [≤69] and -2 SD [≤84]), the
BSID-II-BRS total score, and subscale MQ (≤10th and ≤25th percentile) at t0, t1, t2,
AIMS (≤5th and ≤10 th percentile) at t 0, t1, and MABC-2-NL (standard score ≤7;
percentile ≤16 th) at t3 were assessed by using Spearman’s rank correlation
coefficient (r s). In addition, the correlations between a combination of predictor
assessments (BSID-II-MS, BSID-II-BRS total scale and subscale MQ at t 0, t1,
and t2, and AIMS at t0 and t1 were also assessed with Spearman’s rank correlation
coefficient. In all cases, two sided p-values < 0.05 were considered to be
statistically significant.
Sensitivity, specificity, positive and negative predictive values, and the diagnostic
odds ratio22 (DOR) with 95% confidence intervals were calculated for the
93
5
Chapter 5
BSID-II-MS and BSID-II-BRS total score and the subscale MQ at t 0, t1, and t 2 and
for the AIMS at t 0 and t1 for the prediction of the MABC-2-NL outcome at t3
(normal versus delayed; MACB standard score ≤7 [percentile ≤16 th]). To assess
the predictive validity of the combination of predictor assessments (AIMS and
BSID-II-MS and BSID-II-BRS total score or MQ), results were dichotomized as
normal versus one or more abnormal results.
The DOR was used because it depends significantly on the sensitivity and
specificity of a test. 22 Moreover, the balance between sensitivity and specificity
was considered important because false positive identification can have
negative consequences on families and result in unnecessary pediatric physical
therapy interventions.
Results
Figure 1 provides an overview of the children assessed. A total of 306 children
were eligible; of these, 44 (14%) did not participate in the follow-up. The neonatal
characteristics of the 44 non-participants did not differ significantly from those
of the participants. Of the 262 children who did participate in the follow-up, 41
had no outcome assessment at 5 years of age, of whom nine (22%) had already
been diagnosed with a disability. An additional 20 (9%) of the 221 children
assessed at 5 years of age had disabilities. Of the 29 (11%) children with
disabilities 15 (6%) children were diagnosed with CP (5 unilateral CP [GMFCS
level I in all] and 10 bilateral CP [GMFCS level I in 6; level II in 2; level III in 2; level
IV in 1]. Four children were deaf or severely visually impaired; two children had
syndromes; two children had metabolic diseases; three children had a severe
psychomotor retardation; one child had an autistic spectrum disorder; one child
had a neuromuscular disease; one child had an amputation under the knee.
The children with disabilities were excluded and the study cohort therefore
comprised 201 (106 boys, 95 girls) children.
Table 1 separately presents the perinatal characteristics of all 262 (86%) children
and the 201 (66%) children included in the study. The percentage of children
with intraventricular hemorrhage grades III/IV, retinopathy of prematurity, or
necrotizing enterocolitis was lower in the study group. Also there were more
children with a higher birth weight and more mothers with higher education, in
94
comparing the predictive validity of two infant motor tests
Figure 1 Flow chart of children assessed
May
2003
n
90
2004
2005
Total cohort
n
137
n
139
17
13
24
Deaths
73
2
1
124
2
115
1
-
Discharged
Deaths after discharge
Not invited
312
3
3
Eligible
306
n
366
No assessment 14.4%
54
100%
44
Untraceable / living abroad
Reasons for refusal to participate
Too problematic
Unknown
Parents report no problems
Recently checked by GP
Follow-up elsewhere
Did not show up / ill / cancelled
12
2
1
5
4
17
3
Included
85.6%
5
262
No assessment at 5 years 15.7% 41
Disabilities
Untraceable / living abroad
Reasons for refusal to participate
Too problematic
Unknown
Parents report no problems
Follow-up elsewhere
Did not show up / ill / cancelled
Excluded, disabilities 7.6%
9
9
2
8
6
2
5
20
Included at 5 years 65.7%
201
95
Chapter 5
this group. The mean gestational age (GA) of the infants in the study group was
29+6 weeks (SD 1+6, range 25+4 –31+6) and their mean birth weight was 1239
grams (SD 352.5, range 520–2530). Their mean CA was 6mo 1wk (SD 2wk) at
the t 0 assessment, 12mo 0wk (SD 2wk) at the at t1 assessment, and 24mo 3wk
(SD 3wk) at the t 2 assessment; their chronological age at the t 3 assessment was
63mo 1wk (SD 4wk).
Table 1 Characteristics of all children (n=262) and of the children included
in the study at 5 years of age (n=201)
Gender male
Birth weight < 1000 g
Small for gestational age (-2SD)
n=262
n=201
n (%)
n (%)
140 (53)
106 (53)
69 (26)
57 (28)
5 (2)
3 (2)
39 (15)
30 (15)
6 (2)
2 (1)
Cystic PVL (ultrasound)
11 (4)
10 (5)
Neonatal convulsions
15 (6)
10 (5)
IVH Grade I/II
IVH Grade III/IV
Retinopathy of prematurity
Necrotizing enterocolitis
Chronic lung disease
6 (2)
2 (1)
23 (9)
17 (4)
9 (3)
7 (4)
Reported maternal education at t 2:
Low
49 (19)
35 (17)
103 (39)
86 (43)
High
63 (24)
58 (29)
Missing
47 (18)
22 (11)
Average
Abbreviations: IVH = Intraventricular hemorrhage (ultrasound), PVL = periventricular leuko­malacia,
Low = no/elementary school/vocational training, Medium = lower general secondary education/
secondary vocational education, High = upper general secondary education, higher vocational
education, pre-university education/bachelor degree or higher
Table 2 presents the MABC-2-NL components of manual dexterity, aiming and
catching, balance and the total test scores in the study group (n=201). The
mean total standard score was 8.5 (SD 2.7). Girls tended to perform better
(higher mean scores) than boys for the components manual dexterity and
96
comparing the predictive validity of two infant motor tests
Table 2 MABC-2-NL component (MD, A&C, BAL) and total scores (TTS)
at 5 years of age. Standard score, percentile, and classification in
very preterm children without disabilities are given separately for
boys, girls, and the GA
MABC-2-NL Standard
score
Percentile
Mean (SD) Median(IQR)
Total
MD
9.2 (2.4)
37 (25-63)
n=201
A&C
8.9 (2.7)
37 (25-50)
BAL
8.3 (2.5)
37 (16-37)
TTS
8.5 (2.7)
25 (16-50)
Boys
MD
8.9 (2.3)
37 (16-50)
n=106
A&C
9.0 (2.8)
37 (25-63)
BAL
7.9 (2.3)
25 (9-37)
TTS
8.2 (2.6)
25 (9-50)
Girls
MD
9.6 (2.4)
37 (25-63)
n=95
A&C
8.7 (2.7)
25 (16-50)
BAL
8.8 (2.6)
37 (16-50)
TTS
8.8 (2.7)
31 (16-50)
< 28 wk GA
MD
8.1 (2.2)
37 (18-50)
n=36
A&C
9.0 (2.8)
37 (25-63)
BAL
7.9 (2.3)
25 (9-37)
TTS
8.2 (2.6)
25 (9-50)
≥ 28 < 30 wk GA MD
8.8 (2.5)
37 (16-63)
n=59
A&C
8.9 (3.0)
37 (16-63)
BAL
8.0 (2.6)
25 (16-37)
TTS
8.1 (2.9)
25 (5-50)
≥ 30 wk GA
MD
9.6 (2.3)
50 (25-63)
n=106
A&C
8.8 (2.6)
37 (25-50)
BAL
8.6 (2.5)
37 (16-50)
TTS
8.7 (2.5)
37 (16-50)
Classification
≤5th
≥6 th ≤16 th
≥17th
n (%)
n (%)
n (%)
31 (15)
43 (21) 127 (63)
20 (19)
21 (20)
65 (61)
11 (12)
22 (23)
62 (65)
6 (17)
9 (25)
21 (58)
16 (27)
10 (17)
33 (56)
9 (9)
24 (23)
73 (69)
Abbreviations: MABC-2-NL = Movement Assessment Battery for children-second edition-Dutch
version, MD = manual dexterity, A&C = aiming and catching, BAL = balance, TTS = total test score
GA = gestational age, wk = weeks, IQR = interquartile range
97
5
Chapter 5
balance. The standard deviation for the scores ranged from 2.2 to 3.0 (inter­
quartile range: 9–63 percentile).
Seventy-four children (37%) were assigned to the ‘at risk’ or a ‘significant
movement difficulty’ category. In boys this percentage was 39% and in girls
35%. Of the infants born before 28 weeks, between 28 and 30 weeks and above
30 weeks, respectively 42%, 44%, and 31%.
Table 3 shows the correlation of the motor assessments at t 0, t1, and t 2 with the
assessment at t 3. The dichotomized score on the MABC-2-NL ≤16 th percentile
had no or a weak significant association with the t 0 predictor assessments.
Several assessment from t1 and t 2 did show a moderate association (r s >0.30).
The highest association was found at t0 for the BSID-II-BRS ≤10 (rs =0.25), at t1
for the AIMS ≤5/BSID-II-MS ≤69/MQ ≤25 and AIMS ≤10/BSID-II-MS ≤69/MQ ≤25
combinations (r s =0.38), and at t 2 for the BSID-II-MS ≤69/MQ ≤25 combination
(r s =0.34).
Table 3 Correlation between dicotomized motor performance at 6 (t0),
12 (t1), and 24 (t2) months CA and dicotomized motor outcome
at 5 years (t3) based on the MABC-2-NL in very preterm children
without disabilities
MABC Standard score ≤ 7
(≤ 16 percentiel)
Spearman’s r
AIMS ≤ 5 perc
AIMS ≤ 10 perc
BSID-II-MS PDI ≤ 69
BSID-II-MS PDI ≤ 84
98
p value
t0
0.15
.045
t1
0.28
<.001*
t0
0.15
.043*
t1
0.32
<.001*
t0
0.12
.118
t1
0.26
<.001*
t2
0.23
.001*
t0
0.11
.124
t1
0.28
<.001*
t2
0.22
.002*
t0
0.25
<.001*
comparing the predictive validity of two infant motor tests
Table 3 Continued
MABC Standard score ≤ 7
(≤ 16 percentiel)
Spearman’s r
BSID-II-BRS ≤ 10 perc
BSID-II-BRS ≤ 25 perc
BRS subscale MQ ≤ 10 perc
BRS subscale MQ ≤ 25 perc
AIMS ≤ 10 perc/BRS ≤ 25 perc
AIMS ≤ 10 perc/MQ ≤ 25 perc
BSID-II-MS PDI ≤ 69/BRS ≤ 25 perc
BSID-II-MS PDI ≤ 84/BRS ≤ 25 perc
BSID-II-MS PDI ≤ 69/MQ ≤ 25 perc
BSID-II-MS PDI ≤ 84/MQ ≤ 25 perc
AIMS ≤ 5 and BSID-II-MS ≤ 69
AIMS ≤ 5 and BSID-II-MS ≤ 84
p value
t1
0.04
.600
t2
0.12
.088
t0
0.19
t1
< -0.01
t2
0.17
.019*
t0
0.06
.378
t1
0.18
.014*
t2
0.29
<.001*
t0
0.04
.055
t1
0.18
.012*
t2
0.33
<.001*
.007*
.960
t0
0.18
.014*
t1
0.28
<.001*
t0
0.10
.172
t1
0.34
<.001*
t0
0.20
.005*
t1
0.22
.222
t2
0.23
.001*
t0
0.14
.048
t1
0.27
<.001*
t2
0.19
.008*
t0
0.08
.257
t1
0.30
<.001*
t2
0.34
<.001*
t0
0.10
.172
t1
0.25
<.001*
t2
0.30
<.001*
t0
0.20
.006*
t1
0.31
<.001*
t0
0.15
.044*
t1
0.31
<.001*
5
99
Chapter 5
Table 3 Continued
MABC Standard score ≤ 7
(≤ 16 percentiel)
Spearman’s r
AIMS ≤ 10/BSID-II-MS ≤ 84
AIMS ≤ 10/BSID-II-MS ≤ 69
p value
t0
0.13
.082
t1
0.32
<.001*
t0
0.19
.008*
t1
0.35
<.001*
AIMS trajectory ≤ 10 perc
0.28
<.001*
BSID-II trajectory PDI ≤ 69
0.25
.001*
BSID-II trajectory PDI ≤ 84
0.08
.310
AIMS ≤ 10/BSID-II-MS ≤ 69/BRS ≤ 25
AIMS ≤ 10/BSID-II-MS ≤ 69/MQ ≤ 25
AIMS ≤ 10/BSID-II-MS ≤ 84/BRS ≤ 25
AIMS ≤ 10/BSID-II-MS ≤ 84/MQ ≤ 25
AIMS ≤ 5/BSID-II-MS ≤ 69/BRS ≤ 25
AIMS ≤ 5/BSID-II-MS ≤ 69/MQ ≤ 25
AIMS ≤ 5/BSID-II-MS ≤ 84/BRS ≤ 25
AIMS ≤ 5/BSID-II-MS ≤ 84/MQ ≤ 25
t0
0.21
.004*
t1
0.32
<.001*
t0
0.14
.052
t1
0.38
<.001*
t0
0.12
.089
t1
0.33
<.001*
t0
0.09
.233
t1
0.30
<.001*
t0
0.23
.001*
t1
0.27
<.001*
t0
0.13
.082
t1
0.38
<.001*
t0
0.16
.024
t1
0.29
<.001*
t0
0.10
.166
t1
0.29
<.001*
Abbreviations: AIMS = Alberta Infant Motor Scale, BSID-II-MS = Bayley Scales of Infant
Development, 2nd edition, motor scale, PDI = Psychomotor developmental index, BSID-II-BRS
= Bayley Scales of Infant Development, 2nd edition, behavior rating scale total score, MQ =
motor quality.
* p <.05
100
comparing the predictive validity of two infant motor tests
The sensitivity, specificity, positive and negative predictive values, and DOR of
the predictive assessments and combinations are described in Table 4. Overall,
the BSID-II-MS ≤84 had the best sensitivity and the BSID-II-BRS ≤25 had the
best specificity. The balance between sensitivity and specificity was found
important and the AIMS ≤10 had the best combination of sensitivity and
specificity at t1, which was even better than that of the BSID-II-MS ≤84 at t 2.
Moreover, at t 2 the subscale MQ ≤25 demonstrated a better combination of
sensitivity and specificity than the BSID-II-MS ≤84.
The highest DOR at t 0 was that of the BSID-II-BRS ≤25 (DOR=2.7), at t1 that of
the BSID-II-MS ≤84 (DOR=6.0) and at t 2 that of MQ ≤25 (DOR=4.5). Overall, the
best balance between sensitivity and specificity was achieved at t 0 by the AIMS
≤10/BSID-MS ≤69/BRS ≤25 combination (78%, 43%; DOR=2.5), at t1 by the
AIMS ≤10/BSID-MS ≤69/MQ ≤25 combination (78%, 62%; DOR=5.5), and at t 2
by the BSID-MS ≤69/MQ ≤25 combination (74.0%, 62%; DOR=4.5).
5
101
102
BSID-II-MS ≤69/
MQ ≤25
BRS subscale
MQ ≤25
BSID-II-BRS ≤25
BSID-II-MS ≤84
AIMS ≤10
67.8 (57.0-78.6)
64.8 (50.6-79.0)
68.9 (58.3-79.5)
61.5 (50.5-72.5)
74.0 (61.1-86.8)
t2
62.7 (51.9-73.5)
71.6 (58.5-84.7)
t2
63.9 (52.9-74.9)
86.6 (78.7-94.4)
28.2 (14.8-41.5)
t1
62.0 (47.6-76.4)
70.8 (60.5-81.2)
33.3 (19.4-47.2)
t0
44.4 (29.8-59.1)
81.0 (72.2-89.7)
33.8 (20.0-47.5)
t2
t1
89.9 (83.0-96.8)
9.9 ( 1.0-18.7)
t0
86.7 (78.9-94.4)
29.2 (15.8-42.6)
59.3 (48.3-70.4)
63.0 (48.9-77.1)
t1
31.1 (20.5-41.7)
93.0 ( 85.4-100)
t1
t2
t0
31.1 (20.5-41.7)
79.2 (67.2-91.1)
t0
47.5 (36.0-58.9)
69.4 (55.9-83.0)
t1
Specificity
(95% CI)
t0
Sensitivity
(95% CI)
53.5 (41.1-65.9)
54.3 (40.5-68.2)
42.7 (28.4-56.9)
53.0 (40.5-65.5)
55.6 (34.9-76.3)
40.7 (24.7-56.7)
51.0 (33.2-68.9)
36.8 ( 9.2-64.5)
56.8 (36.4-77.1)
47.9 (35.2-60.7)
44.6 (34.4-54.8)
41.0 (30.6-51.4)
54.8 (41.2-68.3)
44.6 (32.9-56.4)
Positive PV
(95% CI)
79.8 (69.4-90.1)
75.2 (64.9-85.6)
65.5 (54.5-76.6)
79.0 (68.8-89.2)
66.9 (57.4-76.4)
63.9 (53.5-74.3)
67.5 (58.0-71.1)
62.6 (53.3-71.8)
67.1 (57.7-76.5)
73.0 (61.9-84.1)
88.1 (75.6-100)
71.2 (55.4-86.9)
76.2 (65.8-86.6)
71.8 (59.1-84.5)
Negative PV
(95% CI)
0-2.2)
4.5 ( 0.9-8.2)
3.6 ( 0.8-6.5)
1.4 ( 0.3-2.5)
4.2 ( 0.9-7.6)
2.5 ( 0.2-4.9)
1.2 ( 0.2-2.2)
2.2 ( 0.4-4.0)
1.0 (
2.7 ( 0.2-5.2)
2.5 ( 0.6-4.4)
6.0 ( 0-13.5)
1.7 ( 0.2-3.2)
3.9 ( 0.8-7.0)
2.1 ( 0.4-3.7)
DOR
(95% CI)
and 24 months’corrected age, respectively) and outcome assessment (cut-off point MABC ≤16 percentile) at t3
(chronological age of 5 years) in very preterm children without disabilities
Table 4 Sensitivity and specificity of AIMS, BSID-II-MS, BSID-II-BRS total and subscale MQ at t0, t1, and t2 (6, 12,
Chapter 5
41.8 (31.1-52.4)
55.0 (42.6-67.4)
34.5 (23.6-45.3)
61.5 (50.3-72.8)
78.9 (66.8-91.0)
77.5 (65.1-89.9)
t1
52.0 (39.5-64.5)
59.3 (48.0-70.6)
73.2 (60.1-86.4)
t0
44.7 (33.5-55.9)
42.9 (31.5-54.2)
77.5 (65.1-89.9)
t1
54.9 (41.9-68.0)
65.3 (54.3-76.2)
70.4 (56.9-84.0)
t1
t0
44.8 (33.3-56.4)
46.2 (34.8-57.6)
73.2 (60.1-86.4)
53.7 (40.9-66.5)
62.7 (51.6-73.8)
71.8 (58.5-85.2)
t1
t0
40.9 (30.2-51.6)
34.5 (23.6-45.3)
75.0 (62.2-87.8)
t0
81.8 (71.5-92.1)
73.2 (58.4-88.0)
78.7 (67.8-89.5)
76.1 (63.1-89.1)
78.6 (68.2-88.9)
74.3 (61.6-87.0)
78.7 (68.2-89.3)
69.5 (54.5-84.5)
5.5 (0.8-10.2)
2.0 ( 0.3-3.7)
4.0 ( 0.7-7.3)
2.5 ( 0.4-4.8)
4.5 ( 0.9-8.1)
2.4 ( 0.4-4.3)
4.2 ( 0.8-7.8)
1.6 ( 0.3-2.9)
Abbreviations: CI = confidence interval, PV = predictive value, DOR = diagnostic odds ratio, AIMS = Alberta Infant Motor Scale, BSID-II-MS = Bayley Scales
of Infant Development, 2nd edition, motor scale, PDI = Psychomotor developmental index, BSID-II-BRS = Bayley Scales of Infant Development, 2nd edition,
behavior rating scale total score, MQ = motor quality.
AIMS ≤10/BSID-IIMS ≤69/MQ ≤25
AIMS ≤10/BSID-IIMS ≤69/BRS ≤25
AIMS ≤10/BSID-IIMS ≤69
AIMS ≤10/MQ ≤25
comparing the predictive validity of two infant motor tests
5
103
Chapter 5
Discussion
This study demonstrated that the mean MABC-2-NL total standard score in
children born very preterm without known disabilities is ½ SD below the mean
of the reference population. The percentage of children scoring below the 16 th
percentile is twofold higher compared to the reference group.
The significant associations between the results of the MABC-2-NL and the
AIMS, BSID-II-MS and BSID-II-BRS and BRS subscale MQ are weak to moderate.
All infant motor tests are characterized by various limitations with regards to
their ability to predict motor performance at 5 years of age. The BSID-II-MS ≤84
achieved the highest sensitivity (63–93%) in the three assessments, but this
resulted in a large number of children with false positive results. The BSID-II-BRS
had the highest specificity (81–90%) over the three assessments, but its
sensitivity for detecting delayed motor performance was low. The AIMS and the
BRS subscale MQ showed the best balance between sensitivity and specificity
at t1 and t 2, respectively. At t1, the DOR of BSID-II-MS was highest, followed by
the AIMS ≤10/BSID-II-MS ≤69/MQ ≤25 combination. The sensitivity improved by
combining tests, i.e., from 65–69% to 78–79%, but the number of children with
false positive results increased. The combination of sensitivity and specificity of
the AIMS ≤10 at t1 was better than that of the BSID-II-MS ≤84 at t 2. Overall, the
highest predictive validity at t 0 was achieved by the AIMS ≤10/BSID-MS ≤69/
BRS ≤25 combination and at t1 by the AIMS ≤10/BSID-MS ≤69/MQ ≤25
combination. The predictive validity of the best combinations at t 0, t1, and t 2 are
comparable, but as expected lowest at t 0.
To the best of our knowledge, the predictive validity of the AIMS and the BSID-II
for motor performance at 5 years has not been previously investigated. Studies
assessed the predictive validity of the AIMS and BSID-II for gross motor
development at 18 months of age.3;8 In the study on the AIMS, the authors used
the 10 th percentile as the cut-off point at 4 months (sensitivity 77.3%, specificity
81.7%) and the 5th percentile (sensitivity 90.9%, specificity 85.9%) at 8 months.8
The percentages were higher than those found in our study, but the study period
was much shorter. The period was also shorter for the BSID-II study, which
showed a better correlation (r=0.48) with motor performance at 18 months than
in our study.3 We choose the cut-off scores around -1 SD that is a standard
score of 7 or the 16 th percentile for the MABC-2-NL. For the AIMS only the 10 th
104
comparing the predictive validity of two infant motor tests
and 25th percentile are officially listed in the manual, so we choose the 10 th
percentile as most comparable to the -1 SD. Moreover, the choice of a lower
cut-off threshold, such as the 5th percentile or -2 SD, would have decreased the
number of false positives and thereby increasing specificity, and in the same
time would have raised the number of false negatives. Therefore, the choice of
including the 10 th percentile or -1 SD increased sensitivity, but it also increased
the number of false positives.
Depending on the included group of children born preterm, the rates of motor
performance problems at 5 years of age, as assessed using the MABC-1, can
vary from 30 to 60%.14;23-25 The rate found in our study (37%) using the MABC-2,
is slightly higher than the 30% found in an earlier study performed by our group
using the MABC-1.14 The MABC-2 included scores on the 16 th percentile in the ‘at
risk’ category, which may explain the higher percentage. Since about 15% of the
children scored on the 16 th percentile, choosing the 15th percentile as the cut-off
point would have lowered the percentage of children to 22%. We therefore
conclude that any comparison between the outcomes of the MABC-1 and
MABC-2 should be made with caution.
The revised MABC-2 includes transformation of the raw scores into normalized
standard scores with a given mean and SD. The manual indicates that the
standard scores are more suitable for research purposes.11 In our study, the SD
of the MABC scores for preterm children was the same as for the reference
group, indicating that the variability in the reference group and preterm group
was similar. However, the mean standard score for the preterm children was
lower than that for the reference group, and the whole curve had shifted to the
left. This difference is also reflected in the percentile scores, which range
predominantly below or at the 50 th percentile.
For the MABC-2 the Dutch/Flemish reference values were used but for the
BSID-II the U.S. reference values. However, a check revealed that prediction
would not be improved by using the Dutch reference values.
Limitations of the study
This was a single-center study. As such, it is characterized by the limitations of
the generalizability of the findings and possible attrition bias. As can be seen in
table 3 the choices for different cut-off scores per test resulted in many
combinations, with a probability of finding significant correlations by change.
105
5
Chapter 5
The motor tests in this study are frequently used by pediatric physical therapists.
Usually therapists combine the motor scale outcomes, movement quality and
behavioral observations and the parents’ supportiveness in clinical decision
making. This study demonstrates that the prediction of motor performance can
indeed be improved by combining different aspects of motor performance.
However, the outcome of one motor test used in our study showed a weak to
moderate correlation with outcome at 5 years of age. From the developmental
perspective, such a weak to moderate correlation can be expected due to the
influence of numerous external factors, such as the environment, and personal
factors on motor development.
Conflict of interest statement
The authors listed have no conflicts of interest.
Funding
No funding was provided for the study and writing of this manuscript.
Acknowledgements
We thank all of the participating infants and their parents, the staff of the
Neonatal Intensive Care Unit and the Department of Pediatric Physical Therapy.
Special thanks are given to the secretarial assistance from both the Department
of Neonatology and Pediatric Physical Therapy.
106
comparing the predictive validity of two infant motor tests
References
(1) de Kieviet JF, Piek JP, arnoudse-Moens CS, Oosterlaan J. Motor development in very preterm and
very low-birth-weight children from birth to adolescence: a meta-analysis. JAMA 2009; 302:
2235-42.
(2) Piek JP, Dawson L, Smith LM, Gasson N. The role of early fine and gross motor development on
later motor and cognitive ability. Hum Mov Sci 2008; 27: 668-81.
(3) Harris SR, Megens AM, Backman CL, Hayes VE. Stability of the Bayley II Scales of Infant
Development in a sample of low-risk and high-risk infants. Dev Med Child Neurol 2005; 47: 820-23.
(4) Darrah J, Hodge M, Magill-Evans J, Kembhavi G. Stability of serial assessments of motor and
communication abilities in typically developing infants-implications for screening. Early Hum Dev
2003; 72: 97-110.
(5) Janssen AJ, Akkermans RP, Steiner K, de Haes OA, Oostendorp RA, Kollee LA et al. Unstable
longitudinal motor performance in preterm infants from 6 to 24 months on the Bayley Scales of
Infant Development-Second edition. Res Dev Disabil 2011; 32: 1902-09.
(6) Bayley N. Bayley Scales of Infant Development-second edition. San Antonio: The Psychological
Corporation; 1993.
(7) Piper MC, Darrah J. Motor assessment of the developing infant. Philadelphia, PA: WB Saunders;
1994.
(8) Darrah J, Piper M, Watt MJ. Assessment of gross motor skills of at-risk infants: predictive validity
of the Alberta Infant Motor Scale. Dev Med Child Neurol 1998; 40: 485-91.
(9) Palisano RJ. Concurrent and predictive validities of the Bayley Motor Scale and the Peabody
Developmental Motor Scales. Phys Ther 1986; 66: 1714-19.
(10) MacCobb S, Greene S, Nugent K, O’Mahony P. Measurement and prediction of motor proficiency
in children using Bayley infant scales and the Bruininks-Oseretsky test. Phys Occup Ther Pediatr
2005; 25: 59-79.
(11) Henderson SE, Sugden DA, Barnett AL, Smits-Engelsman BCM. Movement Assessment Battery
for children- second edition, Dutch translation. Amsterdam: Pearson Assessment and information
B.V.; 2010.
(12) Provost B, Heimerl S, McClain C, Kim NH, Lopez BR, Kodituwakku P. Concurrent validity of the
Bayley Scales of Infant Development II Motor Scale and the Peabody Developmental Motor
Scales-2 in children with developmental delays. Pediatr Phys Ther 2004; 16: 149-56.
(13) Westera JJ, Houtzager BA, Overdiek B, van Wassenaer AG. Applying Dutch and US versions of the
BSID-II in Dutch children born preterm leads to different outcomes. Dev Med Child Neurol 2008;
50: 445-49.
(14) Janssen AJ, Nijhuis-van der Sanden MW, Akkermans RP, Tissingh J, Oostendorp RA, Kollee LA. A
model to predict motor performance in preterm infants at 5 years. Early Hum Dev 2009; 85:
599-604.
(15) Blanchard Y, Neilan E, Busanich J, Garavuso L, Klimas D. Interrater reliability of early intervention
providers scoring the alberta infant motor scale. Pediatr Phys Ther 2004; 16: 13-8.
(16) Snyder P, Eason JM, Philibert D, Ridgway A, McCaughey T. Concurrent validity and reliability of the
Alberta Infant Motor Scale in infants at dual risk for motor delays. Phys Occup Ther Pediatr 2008;
28: 267-82.
(17) Almeida KM, Dutra MV, de Mello RR, Reis AB, Martins PS. Concurrent validity and reliability of the
Alberta Infant Motor Scale in premature infants. J Pediatr (Rio J ) 2008; 84: 442-48.
(18) Pin TW, de VK, Eldridge B, Galea MP. Clinimetric properties of the alberta infant motor scale in
infants born preterm. Pediatr Phys Ther 2010; 22: 278-86.
(19) Ellinoudis T, Evaggelinou C, Kourtessis T, Konstantinidou Z, Venetsanou F, Kambas A. Reliability
and validity of age band 1 of the Movement Assessment Battery for Children˗second edition. Res
Dev Disabil 2011; 32: 1046-51.
(20) Wood E, Rosenbaum P. The gross motor function classification system for cerebral palsy: a study
of reliability and stability over time. Dev Med Child Neurol 2000; 42: 292-96.
107
5
Chapter 5
(21) Uitenbroek D. Diagnostics, Simple Interactive Statistical Analysis (SISA) 1997. Available from:
URL:http://www.quantitativeskills.com/sisa/statistics/diagnos.htm (20 Aug. 2011)
(22) Glas AS, Lijmer JG, Prins MH, Bonsel GJ, Bossuyt PM. The diagnostic odds ratio: a single indicator
of test performance. J Clin Epidemiol 2003; 56: 1129-35.
(23) Jongmans M, Mercuri E, de VL, Dubowitz L, Henderson SE. Minor neurological signs and perceptual-motor difficulties in prematurely born children. Arch Dis Child Fetal Neonatal Ed 1997; 76:
F9-14.
(24) Bos AF, Roze E. Neurodevelopmental outcome in preterm infants. Dev Med Child Neurol 2011; 53
Suppl 4: 35-9.
(25) de Kleine MJ, den Ouden AL, Kollee LA, Nijhuis-van der Sanden MW, Sondaar M, van
Kessel-Feddema BJ et al. Development and evaluation of a follow up assessment of preterm
infants at 5 years of age. Arch Dis Child 2003; 88: 870-75.
108
comparing the predictive validity of two infant motor tests
5
109
Part III
Chapter 6
Development of a movement quality
measurement tool for children
Anjo JWM Janssen MSc, PT1, Eline TW Diekema MSPT2,
Rob van Dolder MSc, PT3, Louis AA Kollée PhD, MD 4,
Rob AB Oostendorp PhD, PT5, Maria WG Nijhuis-van der Sanden PhD, PT 1, 5
adboud University Nijmegen Medical Centre, Department of Pediatric Physical Therapy;
R
Rehabilitation, Nijmegen, The Netherlands
2
Rijnstate Hospital, Department of Physical Therapy, Arnhem, The Netherlands
3
University of Applied Sciences Utrecht, Centre of Movement Studies, Utrecht, The Netherlands
4
R adboud University Nijmegen Medical Centre, Department of Pediatrics, Nijmegen, The
Netherlands
5
Radboud University Nijmegen Medical Centre, Scientific Institute for Quality of Healthcare,
Nijmegen, The Netherlands
1
Phys ther April 2012; 92:
Chapter 6
Abstract
Background: Pediatric physical therapists assess the quantity and quality of
children’s motor skills. Several quantitative motor tests are currently available,
but a concise measurement tool of observable movement quality (OMQ) is
lacking.
Objective: The purpose of this study was to develop an OMQ measurement tool
for children from the perspective of pediatric physical therapists.
Design: A qualitative, 3-phase study involving pediatric physical therapists was
conducted.
Methods: The first phase consisted of 7 semistructured interviews. The second
phase comprised a structured meeting using a Nominal Group Technique, with
the interviewees required to identify the most relevant OMQ aspects. The third
phase comprised a Delphi technique involving 61 pediatric physical therapy
experts with the aim of achieving at least 80% agreement on relevance,
terminology and definitions of OMQ aspects.
Results: Across all three phases, 32 aspects based on different theoretical
constructs were considered. Fifteen aspects were included in the measurement.
The pediatric physical therapy experts achieved at least 80% agreement on the
definition of 14 OMQ aspects: automated movements, asymmetry in movements,
variation in movements, appropriate gross motor movements, fluency of movements,
reduced muscle tone, increased muscle tone, involuntary movements, accuracy,
slow and/or delayed movements, accelerated and/or abrupt movements, tremors,
strength regulation, and stereotyped movements. The definition of appropriate
fine motor movements achieved 75% agreement. This aspect was included
because gross and fine motor movements are complementary. The aspects
were scored using a 5-point Likert scale, with a total score ranging from 15 to 75
and with a higher score indicating a better OMQ.
Conclusion: The OMQ scale, a concise measurement tool with 15 defined aspects
was developed. Content validity was obtained, but before the OMQ scale can be
used in clinical practice, studies on reliability, construct validity and responsiveness
are needed.
114
Development of a movement quality measurement tool for children
Introduction
Pediatric physical therapists assess the quantity and quality of children’s motor
skills for clinical decision making. Currently available discriminative motor tests
specifically assess quantity in comparison with peers, and most tests are reliable
and validated. In contrast, the assessment of movement quality is observational,
an introspective judgment of a complex phenomenon,1;2 and describes the control
and coordination of movement.
Movement quality assessment is important in diagnostics, evaluation and
prediction.1;3-8 However, descriptions are not standardized, they differ among
therapists, and they depend on the theoretical construct used, which precludes
comparability and longitudinal evaluation. Moreover, it is difficult to reach
agreement among pediatric physical therapists on which aspects should be
observed and the definition of these aspects.
Some discriminative motor test contain few aspects of movement quality.9-12
Other discriminative motor tests are predominantly based on aspects of movement
quality13-16; for example, the Alberta Infant Motor Scale is based on posture,
weight bearing and antigravity movements.14 Several discriminative tests, in
addition to a quantitative assessment, include a separate qualitative assessment.
Seven tests that focus on qualitative assessment will be briefly described:17-23
the Movement-Assessment Battery for Children (MABC), the Structured Observation
of Motor Performance (SOMP), the Neuromotor Behavioral Inventory (NBI), the
Maastricht’s Motor Test (MMT), the Zürich Neuromotor Assessment (ZNA), the
Toddler and Infant Motor Evaluation (TIME), and the Infant Motor Profile (IMP).
The MABC can be used for children from 3 to 16 years of age.17 Movement
quality assessment consists of a hundred observations on posture/body control
and adjustments to task requirements. The individual aspects are scored if
present and are not summed to obtain a total score. 24 The SOMP was designed
for infants from 0 to 10 months of age.18;25 Thirty-six aspects such as asymmetry,
hypertonia, rotation, stands on tiptoes, and tremor, are scored on a 3-point
scale as “no deviation,” “suspected deviation,” or “clear deviation” and are
summed to obtain a total score. The NBI was designed for infants aged 0 to 12
months and 3 years.19;26-28 Stabilizing posture and movement control are scored
as smooth and well coordinated or hypo/hyperreactive with poorly graded
115
6
Chapter 6
movement and are classified as “normal,” “suspect,” or “abnormal.” The MMT
can be used for children aged 5 to 6 years. 20 Thirty-six observations on
associated movements, coordination, and stability are scored on a 3-point
scale and a total score is calculated. The ZNA can be used for children aged 5
to 18 years. 21;29 The associated movements of the extremities, face, head and
body are assessed for their frequency (0-10) and their degree (0-3) and summed
to obtain a total score. The TIME was designed for children aged 4 months to
3.5 years. 22 Subscale scores on quality rating, atypical positions and component
analysis are calculated. The IMP is a video-based measure for infants aged 3 to
18 months. 23;30 Movement quality scale scores for variation (size of repertoire),
variability (ability to select), symmetry and fluency can be calculated.
Eight measurement tools focus on qualitative assessment:31-38 the measure of
General Movements (GM), a measure for children with handicaps, the Pre­schooler
Gross Motor Quality scale (PGMQ), the Gross Motor Performance Measure
(GMPM), the Quality of Upper extremity Skills test (QUEST), the neurological
examination of the child with Minor Neurological Dysfunction (MND), the motor
quality subscale of the Behavior Rating Scale of the second edition of the Bayley
Scales of Infant Development (BSID-II-BRS), and the Combined Assessment of
Motor Performance and Behavior (CAMPB). The measure of GMs assesses
spontaneous movements in babies until 3 months of age, scored on complexity,
variation and fluency and classified into four categories. 31;39-41 The measure for
children with handicaps up to 3 years of age is scored on coordination, stability
and effort.32 (3) The PGMQ is used for children 3 to 6 years of age.33 Aspects
such as trunk and extremity control, alignment, rhythm and balance are scored
as present or not and summed to obtain a total score.
Two measurement tools assess movement quality in children with cerebral
palsy:34;35 (4) The GMPM is for children from 5 months to 12 years of age.34;42-44
Aspects such as alignment, coordination, dissociated movement, stability and
weight shift are scored 0 (“severely abnormal”) to 3 (“normal without cues”).
The QUEST includes additional aspects such as grasp and protective
extension.35
116
Development of a movement quality measurement tool for children
Only three measurement tools evaluate movement quality simultaneously with a
motor test, using a generic measurement, described independently of specific
tasks and diseases:36-38 The neurological examination of the child with MND, first
edition includes three aspects on speed, smoothness and adequacy scored on
a 4- to 5-point Likert scale.36 (7) The motor quality subscale of the BSID-II-BRS
includes 8 aspects.37 Fine and gross motor movements, control of movements,
hypotonicity, hypertonicity, slow and delayed movements, frenetic movements
and tremors are scored on a 5-point Likert scale. The calculated total score can
be classified as “within normal limits”, “questionable” or “non-optimal movement
quality.” The CAMPB scores movement quality in children aged 3 years. Nine
extensive descriptions of coordination, including aspects such as precision,
stability, muscle force, associated movements, asymmetry or tremor are classified
as “good coordination,” “minor incoordination,” “pronounced incoordination” or
“severe incoordination.”38;45;46
In summary, the GMPM and QUEST measure movement quality in children with
cerebral palsy. The MABC, SOMP, NBI, MMT, IMP, ZNA, TIME, and PGMQ
provide quantitative and qualitative information on specific tasks and sometimes
the movement quality is described coupled to the specific tested task. Because
children develop over time, different motor tasks have to be evaluated at different
stages, making longitudinal evaluation impossible. The GMs, the IMP and the
measurement tool for children who are handicapped are video based, which is
time-consuming. Moreover, GMs are no longer present after 3 months of age.
Tests include different movement quality observations, and most have a limited
age range. Finally, the BSID-II-BRS is no longer included in the third version of
this test and 8 aspects as well as 9 descriptions of the CAMPB may be too
limited to assess a complex phenomenon such as movement quality.
To allow comparable qualitative assessment between pediatric physical
therapists and longitudinal evaluation simultaneously with discriminative tests,
movement quality in children needs to be assessed independently from a
specific age, motor task and predetermined theoretical constructs. The aim of
this study was to develop a concise observable movement quality (OMQ)
measurement tool from the perspective of the pediatric physical therapists by
identifying which aspects of OMQ are most relevant for assessment by pediatric
physical therapists in children. Following this, the aim was to develop a first draft
117
6
Chapter 6
of the OMQ scale, reach agreement on the relevance, terminology and definition
of the various aspects, and finally develop the OMQ scale.
Method
This was a qualitative 3-phase study (triangulation method). Figure 1 presents
an overview of the phases. The first phase consisted of semistructured interviews
with pediatric physical therapists to identify the important aspects of OMQ. The
second phase comprised a nominal group technique (NGT), involving a highly
structured meeting, to allow discussion between the interviewed pediatric
physical therapists and prioritization of the observable aspects.47;48 Based on
the results of phase 1 and 2, the draft OMQ scale was developed. The aspects
and definitions included were used in the third phase, the Delphi technique. The
Delphi technique is a structured communication process with 3 characteristics:
(1) anonymity, (2) consecutive rounds of questions with controlled feedback
and (3) a statistical group outcome.49;50 The pediatric physical therapy experts
participating in this study answered questionnaires in 3 rounds, with the
objective of reaching agreement on the relevance, terminology, and definition of
aspects of OMQ.49;51;52
The study design strengthens the perspective and the experience of the
pediatric physical therapists as the starting point. We expected the phenomenon
of movement quality to be too complex to explore using an NGT or the Delphi
technique alone and, therefore, began with interviews. Moreover, the NGT and
Delphi technique were modified using the results from the previous phase to
enhance the outcome.50;53-55
Study phases and participants
Semistructured interviews
We used a convenience sample of 8 pediatric physical therapists from the
department. The educational backgrounds of these pediatric physical therapists
were sufficiently divergent to provide an adequate spread of ideas. Eight
therapists consented to be interviewed and for the interview to be audiorecorded. However, we stopped interviewing after the seventh interview because
no further insights were being found (data saturation).
118
Development of a movement quality measurement tool for children
Figure 1 Flowchart of the phases, number of participants, results, and
number of aspects. For more details, see Table 3
Phases
Results
Start: 8 PPTs
1. Semistructured
n=7
interviews,
2. NGT,
n=7
BSID-II -BRS
Motor quality
subscale
Aspects:
8
1. Semistructured
interviews
Aspects:
25
2. NGT
Aspects
prioritized:
13
Integration process
draft OMQ-scale
Aspects:
16
3. Delphi technique
3. Delphi technique
Invited experts, n= 86
(100%)
Agreement
No response
9
No interest
2
Lack of time
13
Foreign country 1
Participating experts
round 1, 2, and 3
n= 61
(71%)
80%
Round 1
Aspects defined:
4
Round 2
Aspects defined:
3
Round 3
Aspects defined:
75% agreement:
7
1
6
Aspects included in
OMQ-scale:
15
PPT=pediatric physical therapist; NGT=nominal group technique; BSID-II-BRS=Behavior Rating
Scale of the Bayley Scales of Infant Development, second edition; OMQ=observable movement
quality.
119
Chapter 6
In-depth interviews lasted about 1 hour and were performed face-to-face with
the first author in a quiet room. After an explanation of the topic, the interviewees
were asked to freely express their experiences. The semistructured guide
consisted of 2 main questions: (1) “Can you elaborate on the role of movement
quality during your assessment in children of all ages?” (2) “Can you tell me how
you assess MQ in children, what do you observe?” To ensure that all interviewees
reflected on the same subquestions, a topic-list was included in the interview
guide with questions and comments such as: “Can you explain that further?”
and “I’m not sure I understand what you are saying.” The topic list ended with
the questions: “Is there anything else?” and “Did you miss anything in the
interview?” and a process question, “How did you perceive the interview?” The
topic list evolved during the interviews. When aspects of OMQ included in
the list were not mentioned during the interview, the interviewer asked the
interviewees to describe their experience of these aspects.
NGT
Seven pediatric physical therapists participated in the NGT phase. One
interviewed therapist was not able to participate in this phase, and a therapist
who was not interviewed was included (Tab. 1). The NGT lasted up to 2 hours,
was audio-recorded, and consisted of 5 stages: introduction and explanation,
silent generation of ideas, sharing of ideas, group discussion, voting and
ranking of ideas.47;48 The modified NGT procedure is shown in Appendix 1.
In the first draft of the OMQ scale, the prioritized aspects in phase 2 were
integrated with the aspects of the motor quality subscale of the BSID-II-BRS
(Figure 1). We used the 5-point Likert scales of the BRS and scored all aspects
similarly. Each aspect was defined based on the quotes from the interviews, the
NGT, the literature on movement quality, and available existing definitions in the
International Classification of Functioning, Disability and Health for Children and
Youth (ICF-CY)56 by three authors (E.D., A.J., M.N.).1;57-59
Delphi technique
In this phase we initially selected 9 pediatric physical therapy experts, dispersed
across the Netherlands, who are established researchers or have contributed to
the education of pediatric physical therapists. These experts were asked to
provide the names of other pediatric physical therapists experts with relevant
120
Development of a movement quality measurement tool for children
knowledge of movement quality. Expert sampling was applied based on the
following criteria 60;61: (1) pediatric physical therapists working in rehabilitation
centers, general or university hospitals and primary practices across the
Netherlands in order to prevent convergence of ideas and cross-contamination
and (2) teachers of master’s degree-level pediatric physical therapy and neurodevelopment trainees, as such individuals are accustomed to formulating
definitions for educational purposes. Their expertise was documented through
self-reporting of sex, age, years of experience, patient contacts, current work
setting and specialization. All participants answered questionnaires anonymously
via a commercial Internet Web site “Formdesk.”62 A letter providing general
information on the aim of the study was mailed to all potential participants. After
giving informed consent to participate, the therapists received a detailed letter
and an access code to the Web site.
Sixty-one out of 86 invited pediatric physical therapists (71%) participated in all
3 rounds of the Delphi technique; of these, 2 also participated in the interviews
and NGT. The participants worked in 10 of the 12 provinces of the Netherlands.
A flow chart of the participants is included in Figure 1, and Table 1 shows the
demographics of the participants of the semistructured interviews, NGT, and
Delphi technique.
Table 2 shows the themes to be answered by the participants with respect to
relevance, terminology and definitions of the aspects within rounds 1 and 2 of
the Delphi technique. The participants were asked to score each question on a
4-point scale. The relevance was scored as follows: “very important,” “important,”
“unimportant,” and “not important at all.” The terminology and definitions were
scored as follows: “strongly agree,” “agree,” “disagree,” “strongly disagree.”63-65
In round 3, the participants were asked to agree or disagree with the questions.
Relevance, terminology and definition had to be agreed upon by at least 80% of
the participants. In all rounds, open text boxes were provided for criticisms and
proposed improvements. Two e-mail reminders were sent to nonresponders to
maximize the response rate.
121
6
Chapter 6
Table 1 Demographics of the Participants of the Semistructured
Interviews, Nominal Group Technique, and Delphi Technique
Variable
Semistructured
Interviews
Nominal
Group
Technique
Delphi
Technique
7
6 (86%)
7
6 (86%)
61
59 (97%)
1 (14.3%)
1 (14.3%)
12 (19.7%)
Mean
SD
Range
Experience, y
43.4
11.3
25–53
43.9
11.3
25–53
50.6
8.9
28–70
Mean
SD
Range
Work setting
19.2
9.8
4–30
18.9
9.8
4–30
26.1
9.2
5–46
No. of participants
Registered pediatric
physical therapists,a n
Sex, male, n
Age, y
Private practice, n
33 (54%)
Hospital, n
University hospital, n
Rehabilitation center, n
7 (11%)
7
7
Other, n
7 (2+5)a (11%)
11 (18%)
3 (5%)
Also working as
Teacher, n
Scientist, n
2
1
2
1
33 (54%)
12 (20%)
a
Two pediatric physical therapists participated in semistructured interviews and nominal group
technique.
Data analysis
Semistructured interviews
The first author transcribed all audio-recorded semistructured interviews, and
the interviewees verified whether their own transcript expressed what they
intended to say. Only minor textual adjustments were made. Data analysis
consisted of substantive (open) coding of the complete transcripts to generate
all the aspects pediatric physical therapists use to observe movement quality.66;67
The first author: (1) read each complete transcript, (2) selected information on
122
Development of a movement quality measurement tool for children
Table 2 Delphi technique: Themes Included in Rounds 1, 2, and 3
Round 1
Round 2
Round 3
Relevance (%)
Terminology (%)
Agreementb (%)
Definition a (%)
Definition a (%)
Textual feedback by
disagreement
Proposition of textual
improvement
Combine aspects (%)
Add aspectsb (%)
Overlap between aspects
Remove aspects (%)
Vote on terminology of two
related aspects (%)
a
a
Proposition of textual
improvement
Maximum of 3 new aspects
added
New aspects (relevance,
add [%] and textual
improvement)
Overlap between new
added aspects
a
b
Answers on a 4-point scale
Agree or disagree. All other questions open textboxes
relevance, (3) divided the relevant information into parts, (4) coded the parts, (5)
compared the codes within a single interview, (6) compared the codes between
interviews, and (7) defined the main codes. To ensure that the transcripts were
consistently coded, one complete transcript also was analyzed by a pediatric
physical therapist not involved in the interviews and another transcript by a
physical therapy sciences student (data triangulation). In the rare cases of
6
inconsistent coding, consensus was reached.
NGT
The NGT was audio-recorded, transcribed and analyzed by the first author to
identify the motivations for combining aspects, the definitions, and the
modifications. The analytical process followed similar steps as for the interviews.
We used the group discussion to agree upon organizing and defining the core
aspects of movement quality. Next, each pediatric physical therapist voted and
ranked aspects on a scale of 0 to 10, with 10 representing the most important
aspect. For each aspect the scores were summed to provide a total score and
percentages.68
123
Chapter 6
After each round, the percentage of participant agreement on the Likert scales
was calculated for each aspect separately.69 The textual feedback of the
participants was combined per aspect and analyzed separately by 2 of the
authors (E.D. and A.J.) following similar steps as for the interviews. The aspects
and definitions were redefined using the feedback from the participants, until
consensus was reached. The new definitions formed the input for the next
round, together with the percentage of agreement. A summary of the textual
information and an explanation of the choices were e-mailed to the participants
as controlled feedback.61;63
Results
Semistructured interviews
Movement quality was considered important by all participants and, occasionally,
even more relevant than quantitative assessment. The pediatric physical therapists
considered that the following factors could influence movement quality and should
be taken into account during assessment: cognition, social-emotional behavior,
visual or auditory problems, learning style, and parent-child interaction.
Twenty-five aspects of movement quality were identified and numbered, as
listed in Table 3, column A: anticipation, associated reactions, automated motor
programs, stable motor programs, asymmetry, coordination, complexity,
dissociation, movement variation, efficiency, eye-hand coordination, fluency,
timing, involuntary movements, isolation, pathological (spastic) movements,
quantity/amount of movement, range of movements, skills level, speed,
spectrum motor programs, strength regulation, stereotyped movements, trunk
rotation and postural stability. The aspects originated from different theoretical
constructs including neurodevelopmental treatment, motor learning theories,
general movements, and biomechanical, cognitive, and neuromuscular processes.
Movement variation and stereotyped movements are defined differently within
theoretical constructs. Appendix 3 shows for each aspect, the representative
quotations from the semistructured interviews, NGT and Delphi technique. The
Figure presents an overview of the number of aspects per study phase.
NGT
Twenty-five aspects were included in the NGT; of these, 18 were found to be the
most important (Tab. 3, column A, indicated by superscript “a”). Subsequent
124
Development of a movement quality measurement tool for children
discussion during the NGT resulted in the addition of 2 aspects: placing and
degrees of freedom (20, 29). Several aspects were found to be irrelevant for
movement quality: Although anticipation (1) was considered to be observable, it
was agreed that a child with poor movement quality can still be a good anticipator
and not relevant for OMQ. Efficiency (10) was experienced to be unobservable
because energy consumption would have to be measured. Spectrum motor
programs (26) was considered too dependent on one theory and unobservable.
Quantity/amount of movements (21) was experienced to be too dependent on
the time of testing or state of the infant. Range of movements (22) was thought
to be inadequate terminology because when learning a new task a child uses
a small range of motion. Pathological (spastic) movements (19), stereotyped
movements (30), trunk rotation (31) and postural stability (32) were experienced
as overlapping and as such should not be included. Table 3, column B presents
the results of the NGT. During the group discussion, 3 core aspects of movement
quality were recognized: skills level, coordination and variation. Each pediatric
physical therapist prioritized the 13 other aspects by ranking them from 1 to 10,
with the most relevant aspect being given 10 points. Table 3, column B, shows
the relevant percentage for each aspect, with automated motor programs (3)
receiving the highest ranking (12%) and involuntary movements (17) receiving
the lowest ranking (2%). The core aspect skills level (12% of total score) covered
only the aspect automated motor programs. Coordination (54% of total score)
covered the following aspects: complexity, fluency, timing, placing, speed,
and strength regulation. Variation (34% of total score) covered: associated
reactions, asymmetry, dissociation, involuntary movements, isolation, and degrees
of freedom.
Table 3, column C, shows the eight aspects of the motor quality subscale of the
BSID-II-BRS. These were integrated with the 13 aspects prioritized during the
NGT (Table 3, column B) into the first draft of the OMQ scale, giving a total of 16
aspects (Table 3, column D). Choices were made for overlapping aspects,
discussed in detail below.
Delphi technique
Sixteen aspects were included in round 1 (Table 3, column E). There was at
least 80% agreement on the relevance of all 16 aspects, with 11 aspects being
found relevant by at least 90% of the participants. Agreement reached at least
80% for four definitions (Table 3, column F). The aspects fine motor movements
125
6
Chapter 6
Table 3 Names of Aspects in Semistructured interviews (A), Nominal Group
Technique (NGT [B]), Motor Quality Subscale Behavior Rating Scale
(BRS [C]), draft Observable Movement Quality Scale
No.
A
B
C
D
Semistructured
Interviews
NGT
Motor quality
subscale BRS
Draft OMQ scale
Aspect
Relevance (%)
Items
1
Anticipationa
-
-
-
2
Associated reactionsa
-
-
3
Automated motor
programsa
Associated reactions
(3%)
Automated motor
programs (12%)
-
Automated movements
4
Stable motor programa
-
-
5
Asymmetrya
Asymmetry (8%)
-
6
Coordinationa
-
Complexitya
-Coordination
Complexity (10%)
-
7
-
Complexity
8
Dissociation
Dissociation (8%)
-
Dissociation of movements
9
Movement variationa
-
Efficiencya
-Variation
-
-
10
-
-
11
Eye–hand
coordinationa
-
Fine motor movements
required by tasks
Fine motor movements
required by tasks
12
-
-
Gross motor movements
required by tasks
13
Fluencya
Fluency (11%)
Gross motor
movements required
by tasks
Control of movements
14
Timing
Timing (9%)
15
-
-
Hypotonicity
Hypotonicity
16
-
-
Hypertonicity
Hypertonicity
17
Involuntary
movementsa
Involuntary
movements (2%)
-
Involuntary movements
18
Isolation
Isolation (5%)
-
-
126
Asymmetry in quality of
movements
Control of movements
Development of a movement quality measurement tool for children
(Draft OMQ scale [D]), Input round 1, 2 and 3 of Delphi Technique,
and Consensus Reached on Relevance (E, G, and J) Terminology (H)
and Definitions (F, I, and K)
E
F
G
H
I
J
K
Definition
(%)
Input
Definition
(%)
Delphi technique
Round 1
Input
(relevance %)
Round 2
Definition
Input
(%)
(relevance %)
Terminology
(%)
Round 3
-
-
-
-
-
-
-
-
-
-
Automated
movements
-
Automated
movementsb
-
90
Asymmetry in
movementsb
84
Automated
movements (93%)
82
87
-
-
-
-
-
Asymmetry
in quality of
movements (97%)
-
56
Asymmetry in
movements
93
71
-
-
-
-
-
-
Complexity of
movements (97%)
Dissociation of
movements (97%)
-
74
Variation in
movementsb
(Combine 98%)
92
92
-
-
87
-
-
-
-
-
-
-
-
-
Appropriateness
of fine motor
movements (97%)
Appropriateness
of gross motor
movements (97%)
Control of
movements (98%)
74
Appropriate
fine motor
movements
Appropriate
large motor
movements
Fluency of
movements
77
61
Appropriate fine
motor movementsc
75
85
74
Appropriate gross
motor movementsb
90
90
82
Fluency of
movementsb
92
80
75
Reduced muscle
tone (Active tone)
(87%)
46
Reduced
muscle
activation
74
64
Reduced muscle
activation/ reduced
muscle toneb
92
Increased muscle
tone (active tone)
(90%)
54
Increased
muscle
activation
74
72
92
Involuntary
movements (90%)
66
-
90
-
-
Involuntary
movementsb
-
Increased muscle
activation/ increased
muscle toneb
-
-
-
-
-
6
-
127
Chapter 6
Table 3 Continued
A
B
C
D
Semistructured
Interviews
NGT
Motor quality
subscale BRS
Draft OMQ scale
No.
Aspect
Relevance (%)
Items
19
Pathological (spastic)
movements
-
-
-
Pathological patterns
Placing (9%)
-
Placing
-
-
-
22
Quantity/amount of
movements
Range of movementsa
-
-
-
23
Skills levela
-
-
24
Speed
-skills level
Speed (6%)
Slow and delayed
movements
Slow and delayed
movements
Frenetic movements
Frenetic, fast and abrupt
movements
-
20
21
25
26
-
-
27
Spectrum motor
programa
-
-
Tremulousness
Tremors
28
Strength regulationa
Strength regulation (8%)
-
Strength regulation
29
-
-
-
30
Stereotyped
movementsa
Degrees of freedom
(8%)
-
-
-
31
Trunk rotationa
-
-
-
32
Postural stability
-
-
-
25
13
8
16
he therapists found these aspects most important in the third step of the NGT. Dash (-) indicates core
T
aspects: coordination, variation, and skills level.
b
Consensus definition 80% or more.
c
Consensus definition 75% or more.
d
Removed because of overlap with 3, 4, 5 and 6.
e
New aspects added.
a
128
Development of a movement quality measurement tool for children
E
F
G
H
I
J
K
Terminology
(%)
Definition
(%)
Input
Definition
(%)
62
43
Pathological
movement patterns
56
89
79
92
-
-
Accuracy
(Well-aimed)b
-
-
Delphi technique
Round 1
Input
(relevance %)
Round 2
Definition
Input
(%)
(relevance %)
Pathological
patterns (84%)
Placing (84%)
49
-
-
-
-
-
-
-
-
-
-
-
-
77
Slow and/
or delayed
movementsb
92
-
Slow movements
(89%)
98
-
-
Frenetic, fast
movements (89%)
61
Accelerated
and/or abrupt
movementsb
98
86
-
-
-
-
-
-
-
-
-
Tremors (94%)
84
Strength
regulationd (95%)
77
Tremorsb
-
-
-
-
-
-
-
-
Stereotyped
movementse
(90%)
-
61
Add (80%)
-
-
-
57
Add (59%)
-
-
Trunk
movementse
(79%)
Trunk stabilitye
(90%)
-
71
Add
(85%)
16
71
Pattern
movements
Accuracy
(Well-aimed)
Round 3
-
17
95
Remove Strength regulationb
(64%)
(Add 75%)
-
-
80
-
Stereotyped
movementsb
(Add 98%)
-
92
-
-
6
-
11
129
Chapter 6
and gross motor movements (11, 12) were experienced to give an overall
impression of OMQ. The majority of the participants added variation (9) and
found that complexity (7) and dissociation (8) should be combined as variation
in movements (7/8/9). This combined aspect was included. Three unselected
aspects in the NGT were frequently added and included again: stereotyped
movements, trunk movements and trunk stability (30, 31, 32). Strength regulation
(28) was removed because of overlap with the aspects control of movements
and reduced muscle tone and increased muscle tone (13/14, 15,16).
Seventeen aspects were included in round 2 (Tab. 3, column G), of which ten
were renamed or had their name modified. In this round there was at least 80%
agreement on the terminology and definitions of 7 aspects (Table 3, columns H
and I). However, 2 aspects (automated movements and fluency of movements)
required textual modifications and, therefore, were reintroduced in round 3.
Agreement was reached on the terminology but not the definition of three
aspects (asymmetry in movements, appropriate large motor movements and
fluency of movements), and these aspects were redefined. No agreement on
terminology and definition was reached for appropriate fine motor movements,
reduced muscle tone, increased muscle tone, and pathological movement
patterns, and these were reformulated and redefined. Of the 3 added aspects in
round 2, trunk rotation (31) was found to be irrelevant and, therefore, was
excluded. Because no agreement on the definition of stereotyped movements
(30) was reached, this aspect was redefined. Moreover, despite 85% agreement
that trunk stability (32) should be added it was found to have too much overlap
with appropriate gross motor movements (12). Therefore, it was combined as
appropriate gross motor movements.
Eleven aspects were included in round 3 (Tab. 3, column J). Strength regulation
(28) was reintroduced because 75% of the participants found that this aspect no
longer overlapped with fluency of movements (13/14) due to altered terminology
for the latter and agreement on the definition (Tab. 3, column K). Stereotyped
movements (30) received agreement on terminology and definition. Participants
voted for the terminology of muscle tone (15) or muscle activation (16), and 74%
choose muscle tone. Therefore, this terminology was adopted, and there was
92% agreement on its definition. Appropriate fine motor movements (11)
received 75% agreement and because gross and fine motor movements are
complementary, this aspect was included. Of all the relevant aspects, no
130
Development of a movement quality measurement tool for children
agreement on terminology and definition was reached for pathological patterns
(19) and this aspect was excluded.
In summary, across all 3 phases, 32 aspects based on different theoretical
constructs were considered. From these, 15 aspects agreed on for their relevance,
terminology and definitions were included in the OMQ scale: Appropriate fine
motor movements and appropriate gross motor movements, giving an overall
impression as to whether the child is able to adapt movements to a task and the
environment conditions; automated movements, reflecting whether the child is
able to perform a task with a low degree of attention; asymmetry in movements,
indicating asymmetry that does not fit the task or age of the child; variation in
movements, indicating an extensive repertoire of movements, independent
movement of body parts and the necessary degrees of freedom; fluency of
movements, referring to the control of movement, without faltering; reduced
muscle tone and increased muscle tone separately, giving the impression of
movements being slack or stiff; involuntary movements, reflecting unconscious
movements disrupting proper execution and inappropriate for the age; accuracy
(well-aimed), showing that the target is reached immediately; slow/delayed
movements and accelerated /abrupt movements, indicating movement with a
lower or higher speed than suitable for the task; tremors, describing an
involuntary rhythmic trembling during movement; strength regulation, showing
force recruitment suitable for the task; and stereotyped movements, defined as
repetitive, nonpurposive movements. The appendix 2 presents the OMQ scale,
including definitions and scoring using a 5-point Likert scale. The total score
ranges from 15 to 75, with a higher score indicating a better OMQ. The actual
Dutch version was translated into English by 3 of the authors (A.J., E.D., R.N.). A
native speaker corrected the translation.
Discussion
This study involved the use of 3 consecutive qualitative research methods to
develop a measurement tool of OMQ performed by pediatric physical therapists
in children. The therapists reached agreement on relevance, terminology, and
definitions of 15 OMQ aspects, included in the OMQ scale (Appendix 2). The 4
areas discussed are: (1) the research process, (2) the included aspects, (3) the
relation to the movement quality model, and (4) the OMQ scale.
131
6
Chapter 6
Research process
The study started with semistructured interviews to ensure equal contribution of
all individual ideas and explore the pediatric physical therapists’ own perspectives
on this complex phenomenon. Three consecutive phases were performed, the
results from the previous phase were used in the subsequent phase. Ultimately,
we reached high agreement, which is important for future implementation.
The influence of the researchers was most explicit in the first draft of the OMQ
scale. The results from the interviews and NGT had to be integrated with the
motor quality subscale of the BSID-II-BRS. The NGT aspects fluency (13) and
timing (14) were combined with control of movements from the subscale. The
NGT aspect speed was divided into the subscale aspects of slow/delayed
movements (24) and frenetic, fast and abrupt movements (25). Gross motor
movements required by tasks (12) and tremors (27) were not identified in the
interviews, but were included from the subscale. The majority of the participants
considered muscle tone unobservable; therefore, it was not identified as an
aspect in the interviews and NGT. However, as hypotonicity (15), hypertonicity
(16) are part of the subscale, they were included.
From the overlapping aspects, complexity (7), isolation (18), and degrees of
freedom (29) only the first was included because complexity is defined as “the
spatial aspects of movement variation.”70
We chose to include involuntary movements (17) rather than associated
reactions/movements (2) because the latter can also be part of involuntary
movements.57;71 Automated motor programs (3) was combined with stable motor
programs (4) as automated movements. Automated movements (3) was
considered as an aspect of motor learning, but preferred over stable motor
programs because these were considered unobservable. Eye-hand coordination
was renamed into fine motor movements required by tasks (11). Finally,
pathological (spastic) movement (19) was included by the authors, although the
interviewees considered it irrelevant.
During the Delphi rounds, the researchers gave feedback on level of agreement,
all criticisms and the proposed improvements of all participants. Additionally,
the researchers gave their motivations and considerations for the new proposals.72
132
Development of a movement quality measurement tool for children
The reformulated terminology and definitions produced independently by the 2
researchers did not differ significantly. The final definitions were extensively
discussed and the participants then had the opportunity to reflect on the
researchers’ choices. The researchers realized that their personal experience
and background influenced the interpretation of the data.73
The geographic and workplace settings of the participating pediatric physical
therapists were well dispersed, and they had many years of experience. In this
study, the percentage of male pediatric physical therapists (20%) seems high,
but reliable data for the Netherlands are lacking. Sixty-one experts, about 5% of
the pediatric physical therapists in the Netherlands, participated in this study.
The literature is inconclusive regarding the appropriate number of participants
in the Delphi technique. Some authors state that including more than 30 experts
does not improve the quality of the technique.60;64 However, higher numbers
have been reported to improve transferability.49;60
Included aspects
In the first Delphi round, 84% of the participants found pathological (spastic)
movements a relevant aspect. Unexpectedly, no agreement was reached on its
definition. Spasticity and cerebral palsy are defined in the literature 56;57;74; we
based the definition of pathological (spastic) movements on relevant definitions
described in neurology books.71;75 However, the definition used may have been
described too specifically. Moreover, the aspect overlapped with the aspects
variation and increased muscle tone and, therefore, was not included.
The definition of fine motor movements was a subject of debate. Some
participants stated that all tasks should be named explicitly to avoid misinterpretation. However, we did not want to include all detailed descriptions in the
definition and, therefore, obtained only 75% agreement. Although an agreement
of at least 80% is the usual cutoff,69 75% is occasionally accepted.76 Because
gross and fine motor aspects are complementary, we decided to include the
fine motor aspect in the OMQ scale. In general, the different theoretical
constructs used by the pediatric physical therapists remained an important
determinant in the disagreement on terminology and definitions.
133
6
Chapter 6
The final OMQ scale holds 3 core aspects and one aspect indicating the
repertoire of movement on a total of 15 aspects. The core aspect of skills level
included as automated movements was an aspect not present in existing qualitative
measurement tools. Two aspects, appropriate gross motor movements and fine
motor movements, also are closely related to skills level. The appropriateness
of these aspects is also called “variability.”70 The core aspect skills level
originates from more recent motor learning theories
59
and is important for
evaluation of interventions.
The core aspect coordination of movement is included in many measurement
tools.3;8;17;19;23;32;33;35;38;44 Coordination covers 6 aspects. The aspect fluency of
movements is included in 3 measurement tools17;23;33; it is referred to as
“rhythmically” in 1 measurement tool 33 and as ‘smooth’ in 2 others. 26;36 The
aspects of slow/delayed movements and accelerated/abrupt movements (i.e.
speed) are included in 4 measurement tools.17;26;36 Accuracy indicating the
end-point of the movement is present in 2 measurement tools. 26;38 Strength
regulation is used in 2 measurement tools.17;38 Muscle tone is present in 3
measurement tools.18;19;22 Both strength regulation and muscle tone need to be
included in the OMQ scale because “a balance between active and passive
muscle power (muscle tone)77 is needed to create stable posture and fluent
motility78 and can be used to predict outcome”79
In the second phase of this study, variation was defined a core aspect covering
several aspects, but in the Delphi technique it was reintroduced as a single
aspect. Moreover, the definition of variation differs among theoretical constructs.23;31
It is important, therefore, that it be defined as clearly as possible.80 The final
definition of variation, included an extensive repertoire of movements that had
to match the demands of the tasks and the environment. This last part is more
expressed in variability. Variability is defined as the capacity to select from the
repertoire the motor strategy that fits the situation best,70 but also refers to
different solutions to the same challenge and intra-individual variations in
repetitions of the same task.81 In the description, variation and variability were
combined with degrees of freedom82 and independent movement. The feasibility
of this combined option has yet to be proven.
134
Development of a movement quality measurement tool for children
The aspect of asymmetry is included in most measurement tools.17;18;23;26;38
Involuntary movements are considered important to assess OMQ, 20;26;29;36 and
used one-dimensionally as an outcome measure of OMQ.3;29 However, in the
OMQ scale it is necessary to judge what is normal at a certain age and task.
Correct interpretation may require experienced pediatric physical therapists.
Tremors and stereotyped movements were not related to a core aspect because
they were included after the NGT phase. Tremors are described in five
measurement tools18;23;26;36;38 and were defined in the ICF-CY.56 Tremors can be
added to the core aspect coordination of movement. Stereotyped movements
cannot be added to one of the core aspects because they refer to the repertoire
of the observable movements. In this study stereotyped movements were
defined similarly as in the ICF-CY.56 Only one measurement tool includes a
stereotyped movement. 23 The above described diversity of aspects shows the
multidimensionality of the OMQ scale.
Relation to the Movement Quality Model
Skjaerven et. al. 2 published a qualitative study searching for understanding of
movement quality by interviewing physical therapists. This study resulted in the
movement quality model, which defines 4 themes: biomechanical, psycho-social-cultural, physiological and existential. All aspects of the OMQ scale are
related to the biomechanical theme: space, postural stability, path and form.
The factors influencing OMQ of our study are related to the psycho-social-cultural theme: energy, awareness, intention, emotion, sociocultural. However,
these factors were considered unobservable and not included in the OMQ
scale. Within the physiological theme: breathing and centering are not directly
recognizable in our aspects, whereas flow, elasticity and rhythm do relate to our
aspects fluency, variation, strength regulation, accelerated/abrupt movements,
and slow/delayed movements. We did not find aspects directly related to the
existential theme: the person or self-awareness. Skjaerven’s and colleagues’
study objective was to identify as many features and characteristics of movement
quality as possible on intuitive associations and descriptions on pictures of
sculptures and drawings. Our objective was uniform observation of movement
quality by pediatric physical therapists which might explain why breathing,
centering and the existential theme were not found.
135
6
Chapter 6
The OMQ-scale
The aspects of OMQ scale were defined independently from age, tasks, and
theoretical constructs to allow longitudinal comparison and to avoid outdated
terminology and definitions in case of new theories.73 Moreover, the aspects
needed to be observable and reported in the OMQ scale, while interpretation
should be part of the clinical decision making. The introspective judgment of the
pediatric physical therapist is required to decide what OMQ is “adequate” at that
age. Therefore, experience, cultural background, theoretical knowledge of the
pediatric physical therapist and the interaction between the therapist and the
child will influence the scoring. Future studies are needed to establish the
weight of these factors.
We used consistent terminology for the definitions of the aspects within the
OMQ scale. We used movements and postures when relevant, although in the
discussion also changing and maintaining body position was proposed. 56
Changing and maintaining body position is used in ICF-CY for defining mobility,
we focused on the quality of movements and postures. The term body parts in
this OMQ scale includes the extremities, head and trunk. Furthermore, we used
the tasks and the environment and not in the environment because the
environment itself can also influence movement quality.57 The child as a whole,
the context in which the child is moving, and the tasks that have to be performed
should be taken into account when the movements are observed.57 The 5-point Likert scale, choosing from a range of 5 declarative statements, was
preferred as outcome scale, since the visual analog scale has shown low
inter-rater reliability on overall movement quality in adults.83 The 5-point Likert
scale enhances comparability and was used in BSID-II-BRS, in which motor
quality subscale was found to be significantly predictive for delayed motor
performance in preterm infants.4 Future research on the multidimensionality of
the OMQ scale may show the need for subscales or weighing of scores.
To our knowledge, concise measurement tools to evaluate generic OMQ are
limited to the BSID-II-BRS motor quality subscale and the CAMPB. The feasibility
of the multidimensionality of the OMQ scale compared to one-dimensional
measurement tools such as the ZNA on associated reactions, GMPM on
extended description of fundamental movement patters, or GM on gestalt
136
Development of a movement quality measurement tool for children
perception has to be proven. The OMQ scale requires gestalt perception and an
introspective judgment on what is normal at what age, taking into account the
task performed and environmental constraints.
Limitations
The transferability of the study to the national level is guaranteed because we
used method and data triangulation, the participants were representative for the
country, and good agreement was obtained in the Delphi technique. However,
we cannot exclude that the researchers’ experiences and backgrounds influenced
the outcome. It is unlikely that movement quality differs among countries,
although language differences may influence the international transferability of
the results. The content validity in the OMQ scale for assessment in children
was obtained. To establish whether generic OMQ assessment simultaneously
with discriminative tests is possible, studies on reliability, construct validity, and
responsiveness are needed.
Conclusion
The OMQ scale, a concise measurement tool with 15 defined aspects, was
developed from the perspective of pediatric physical therapists. All aspects are
scored on a 5-point Likert scale and the total score can be used for qualitative
assessment and longitudinal evaluation. The content validity of the OMQ scale
was obtained. Before the scale can be used in clinical practice, the international
validation, reliability, construct validity and responsiveness must be investigated,
6
for which we will initiate future studies.
Acknowledgments
The authors thank all participants for their valuable contribution to this study,
sharing their knowledge and their time to respond to the extensive questionnaires,
interviews, and discussions.
137
Chapter 6
References
(1) Knudson DV, Morrison CS. Qualitative Analysis of Human Movement, second edition. Champaign,
IL: Human Kinetics Inc; 2002.
(2) Skjaerven LH, Kristoffersen K, Gard G. An eye for movement quality: a phenomenological study of
movement quality reflecting a group of physiotherapists’ understanding of the phenomenon.
Physiother Theory Pract 2008; 24: 13-27.
(3) Kroes M, Kessels AG, Kalff AC, Feron FJ, Vissers YL, Jolles J et al. Quality of movement as
predictor of ADHD: results from a prospective population study in 5- and 6-year-old children. Dev
Med Child Neurol 2002; 44: 753-760.
(4) Janssen AJ, Nijhuis-van der Sanden MW, Akkermans RP, Tissingh J, Oostendorp RA, Kollée LA. A
model to predict motor performance in preterm infants at 5 years. Early Hum Dev 2009; 85: 599-604.
(5) Harris SR, Heriza CB. Measuring infant movement: clinical and technological assessment
techniques. Phys Ther 1987; 67: 1877-1880.
(6) Heineman KR, Hadders-Algra M. Evaluation of neuromotor function in infancy: a systematic review
of available methods. J Dev Behav Pediatr 2008; 29: 315-323.
(7) Wainwright SF, Shepard KF, Harman LB, Stephens J. Factors that influence the clinical decision
making of novice and experienced physical therapists. Phys Ther 2011; 91: 87-101.
(8) Prechtl HF, Einspieler C, Cioni G, Bos AF, Ferrari F, Sontheimer D. An early marker for neurological
deficits after perinatal brain lesions. Lancet 1997; 349: 1361-1363.
(9) Chandler LS, Andrews MS, Swanson MW, Larson A. Movement Assessment of Infants: a manual.
Rolling Bay, WA: Infant Movement Research; 1980.
(10) Folio MR, Fewell RR. Peabody Developmental Motor Scales: examiner’s manual, second edition.
Austin, TX: Pro-Ed; 2000.
(11) Bruininks RH, Bruininks BD. Bruininks–Oseretsky Test of Motor Proficiency, second edition.
Minneapolis, MN: NCS Pearson Inc; 2005.
(12) Bayley N. Bayley Scales of Infant and Toddler Development, third edition. San Antonio, TX:
Harcourt Asessment Inc; 2006.
(13) Burns YR, Ensbey RM, Norrie MA. The Neuro-Sensory Motor Developmental Assessment part 1:
development and administration of the test. The Aust J Physioth 1989; 35: 141-149.
(14) Piper MC, Pinnell LE, Darrah J, Maguire T, Byrne PJ. Construction and validation of the Alberta
Infant Motor Scale (AIMS). Can J Public Health 1992; 83: S46-S50.
(15) Campbell SK, Kolobe TH, Osten ET, Lenke M, Girolami GL. Construct validity of the Test of Infant
Motor Performance. Phys Ther 1995; 75: 585-596.
(16) Ulrich DA. Test of Gross Motor Development, second edition. Austin, TX: Pro-Ed; 2000.
(17) Henderson SE, Sugden DA, Bernett AL. Movement Assessment Battery for Children, second
edition (Movement ABC-2). London: United Kingdom: Pearson Education Inc; 2007.
(18) Persson K, Stromberg B. Structured Observation of Motor Performance (SOMP-I) applied to
neonatally healthy fullterm infants at the ages of 0-10 months. Early Hum Dev 1995; 40: 127-143.
(19) Gorga D, Stern FM. The Neuromotor Behavioural Inventory. New York, NY: New York Hospital
Cornell Medical Centre; 1979.
(20) Kroes M, Vissers YL, Sleijpen FA, Feron FJ, Kessels AG, Bakker E et al. Reliability and validity of a
qualitative and quantitative motor test for 5- to 6-year-old children. Eur J Paediatr Neurol 2004; 8:
135-143.
(21) Largo RH, Caflisch JA, Hug F, Muggli K, Molnar AA, Molinari L et al. Neuromotor development from
5 to 18 years, part 1: timed performance. Dev Med Child Neurol 2001; 43: 436-443.
(22) Miller LJ, Roid GH. The T.I.M.E., Toddler and Infant Motor Evaluation, a standardized assessment.
San Antonio, TX: Therapy skills builders; 1994.
(23) Heineman KR, Bos AF, Hadders-Algra M. The Infant Motor Profile: a standardized and qualitative
method to assess motor behaviour in infancy. Dev Med Child Neurol 2008; 50: 275-282.
(24) Gard L, Rösblad B. The qualitative motor observations in Movement ABC: aspects of reliability and
validity. Adv in Physioth 2009; 11: 51-57.
138
Development of a movement quality measurement tool for children
(25) Persson K, Stromberg B. Structured Observation of Motor Performance (SOMP-I) applied to
preterm and full term infants who needed neonatal intensive care: a cross-sectional analysis of
progress and quality of motor performance at ages 0-10 months. Early Hum Dev 1995; 43: 205-224.
(26) Gorga D, Stern FM, Ross G. Trends in neuromotor behavior of preterm and fullterm infants in the
first year of life: a preliminary report. Dev Med Child Neurol 1985; 27: 756-766.
(27) Gorga D, Stern FM, Ross G, Nagler W. The neuromotor behavior of preterm and full-term children
by three years of age: quality of movement and variability. J Dev Behav Pediatr 1991; 12: 102-107.
(28) Gorga D, Stern FM, Ross G, Nagler W. Neuromotor development of preterm and full-term infants.
Early Hum Dev 1988; 18: 137-149.
(29) Largo RH, Caflisch JA, Hug F, Muggli K, Molnar AA, Molinari L. Neuromotor development from 5 to
18 years, part 2: associated movements. Dev Med Child Neurol 2001; 43: 444-453.
(30) Heineman KR, La Bastide-Van Gemert S, Fidler V, Middelburg KJ, Bos AF, Hadders-Algra M.
Construct validity of the Infant Motor Profile: relation with prenatal, perinatal, and neonatal risk
factors. Dev Med Child Neurol 2010; 52: e209-e215.
(31) Prechtl HF. State of the art of a new functional assessment of the young nervous system: an early
predictor of cerebral palsy. Early Hum Dev 1997; 50: 1-11.
(32) Kraus de Camargo O, Storck M, Bode H. Video-based documentation and rating system of the
motor behaviour of handicapped children treated with physiotherapy: a new outcome measure.
Pediatr Rehabil 1998; 2: 21-26.
(33) Sun SH, Zhu YC, Shih CL, Lin CH, Wu SK. Development and initial validation of the Preschooler
Gross Motor Quality Scale. Res Dev Disabil 2010; 31: 1187-1196.
(34) Boyce WF, Gowland C, Hardy S, Rosenbaum PL, Lane M, Plews N et al. Development of a qualityof-movement measure for children with cerebral palsy. Phys Ther 1991; 71: 820-828.
(35) De Matteo C, Law M, Russel D, Pollock N, Rosenbaum PL, Walter S. QUEST: Quality of upper
extremity skills test. Hamilton, Ontario, Canada: Mc Masters University, Neurodevelopmental
Clinical Research Unit; 1992.
(36) Touwen BCL. Examination of the Child with Minor Neurological Dysfunction, second edition.
London, United Kingdom: Spastics international medical publications; 1979.
(37) Bayley N. Bayley Scales of Infant Development, second edition. San Antonio, TX: The Psychological
Corporation; 1993.
(38) Hemgren E, Persson K. A model for combined assessment of motor performance and behaviour
in 3-year-old children. Ups J Med Sci 1999; 104: 49-85.
(39) Cioni G, Bos AF, Einspieler C, Ferrari F, Martijn A, Paolicelli PB et al. Early neurological signs in
preterm infants with unilateral intraparenchymal echodensity. Neuropediatrics 2000; 31: 240-251.
(40) Bos AF, Martijn A, Okken A, Prechtl HF. Quality of general movements in preterm infants with
transient periventricular echodensities. Acta Paediatr 1998; 87: 328-335.
(41) Butcher PR, van BK, Bouma A, Einspieler C, Stremmelaar EF, Bos AF. The quality of preterm
infants’ spontaneous movements: an early indicator of intelligence and behaviour at school age. J
Child Psychol Psychiatry 2009; 50: 920-930.
(42) Boyce WF, Gowland C, Rosenbaum PL, Lane M, Plews N, Goldsmith C et al. Gross motor
performance measure for children with cerebral palsy: study design and preliminary findings. Can
J Public Health 1992; 83: S34-S40.
(43) Boyce WF, Gowland C, Rosenbaum PL, Lane M, Plews N, Goldsmith CH et al. The Gross Motor
Performance Measure: validity and responsiveness of a measure of quality of movement. Phys
Ther 1995; 75: 603-613.
(44) Boyce WF, Gowland C, Rosenbaum PL, Hardy S, Lane C, Plews N et al. Gross Motor Performance
Measure Manual. Kingston, Ontario, Canada: Queen’s University, School of Rehabilitation
Therapy; 1998.
(45) Hemgren E, Persson K. Quality of motor performance in preterm and full-term 3-year-old children.
Child Care Health Dev 2004; 30: 515-527.
139
6
Chapter 6
(46) Hemgren E, Persson K. Deficits in motor co-ordination and attention at 3 years of age predict
motor deviations in 6.5-year-old children who needed neonatal intensive care. Child Care Health
Dev 2009; 35: 120-129.
(47) Potter M, Gordon S, Hamer P. The difficult patient in private practice physiotherapy: a qualitative
study. Aust J Physiother 2003; 49: 53-61.
(48) Potter M., Gordon S., Hamer P. The nominal group technique: a useful consensus methodology in
physiotherapy research. NZ Journal of Physiotherapy 2004; 32: 126-130.
(49) Dalkey N, Helmer O. An experimental application of the Delphi method to the use of experts.
Management Science 1963; 9: 458-467.
(50) Rowe G, Wright g, Bolder F. Delphi a reevaluation of research and theory. Technical forecasting
and social change 1991; 39: 235-251.
(51) Brown AK, O’connor PJ, Roberts TE, Wakefield RJ, Karim Z, Emery P. Recommendations for musculoskeletal ultrasonography by rheumatologists: setting global standards for best practice by
expert consensus. Arthritis Rheum 2005; 53: 83-92.
(52) Powell C. The Delphi technique: myths and realities. J Adv Nurs 2003; 41: 376-382.
(53) Ryan M, Scott DA, Reeves C, Bate A, van Teijlingen ER, Russell EM et al. Eliciting public preferences
for healthcare: a systematic review of techniques. Health Technol Assess 2001; 5: 1-186.
(54) Dobbie A, Rhodes M, Tysinger JW, Freeman J. Using a modified nominal group technique as a
curriculum evaluation tool. Fam Med 2004; 36: 402-406.
(55) Clayton R, Perera R, Burge S. Defining the dermatological content of the undergraduate medical
curriculum: a modified Delphi study. Br J Dermatol 2006; 155: 137-144.
(56) International Classification of Functioning, Disability and Health: Children and Youth Version
(ICF-CY). Geneva, Switserland: World Health Organization; 2007.
(57) Shumway-Cook A, Woollacott M. Motor Control; Translating Research into Clinical Practice, third
edition. Philadelphia, PA: Lippincott Williams & Wilkins; 2007.
(58) Cech D, Martin S. Functional Movement Development Across the Life Span, second edition.
Elsevier Health Sciences; 2002.
(59) Schmidt R, Wrisberg CA. Motor Learning and Performance, fourth revised edition. Champaign, IL:
Human Kinetics Inc; 2010.
(60) Fink A, Kosecoff J, Chassin M, Brook RH. Consensus methods: characteristics and guidelines for
use. Am J Public Health 1984; 74: 979-983.
(61) Keeney S, Hasson F, McKenna H. Consulting the oracle: ten lessons from using the Delphi
technique in nursing research. J Adv Nurs 2006; 53: 205-212.
(62) Formdesk. [Sept. 2009] Available at: http://www.formdesk.com
(63) Goodman CM. The Delphi technique: a critique. J Adv Nurs 1987; 12: 729-734.
(64) de Villiers MR, de Villiers PJ, Kent AP. The Delphi technique in health sciences education research.
Med Teach 2005; 27: 639-643.
(65) Garland R. The mid-point on a rating scale: is it desirable? Marketing Bulletin 1991; 2:66-70.
(66) Glaser BG. Theoretical sensitivity. Mill Valley, CA: Sociology Press; 1978.
(67) Glaser BG. Emergence versus Forcing: Basics of Grounded Theory Analysis. Mill Valley, CA:
Sociology Press; 1992.
(68) Delbecq AL, Ven van de AH, Gustafson DH. Group Techniques for Program Planning: a Guide to
Nominal and Delphi Processes. Glenview, IL: Scott, Foresman and Company: 1975.
(69) Weigl M, Cieza A, Andersen C, Kollerits B, Amann E, Stucki G. Identification of relevant ICF categories
in patients with chronic health conditions: a Delphi exercise. J Rehabil Med 2004; 44: 12-21.
(70) Hadders-Algra M. Variation and variability: key words in human motor development. Phys Ther
2010; 90: 1823-1837.
(71) Campbell SK. Pediatric Neurologic Physical therapy, second edition. New York, NY: Churchill
Livingstone; 1991.
(72) Hasson F, Keeney S, McKenna H. Research guidelines for the Delphi survey technique. J Adv Nurs
2000; 32: 1008-1015.
(73) Malterud K. Qualitative research: standards, challenges, and guidelines. Lancet 2001; 358: 483-488.
140
Development of a movement quality measurement tool for children
(74) Rosenbaum PL, Paneth N, Leviton A, Goldstein M, Bax M, Damiano D et al. A report: the definition
and classification of cerebral palsy, April 2006. Dev Med Child Neurol Suppl 2007; 109: 8-14.
(75) McGarthy, Gillian T. Physical Disability in Childhood: An Interdisciplinairy Approach to
Management. New York, NY: Churchill Livingstone; 1992.
(76) Dole RL, Arvidson K, Byrne E, Robbins J, Schasberger B. Consensus among experts in pediatric
occupational and physical therapy on elements of individualized education programs. Pediatr
Phys Ther 2003; 15: 159-166.
(77) de Groot L, Hopkins B, Touwen BC. A method to assess the development of muscle power in
preterms after term age. Neuropediatrics 1992; 23: 172-179.
(78) Samsom JF, de GL, Hopkins B. Muscle power and medical history in high risk preterm infants at 3
months of corrected age. Neuropediatrics 1998; 29:127-132.
(79) Samsom JF, de GL, Bezemer PD, Lafeber HN, Fetter WP. Muscle power development during the
first year of life predicts neuromotor behaviour at 7 years in preterm born high-risk infants. Early
Hum Dev 2002; 68: 103-118.
(80) Piek JP. The role of variability in early motor development. Infant Behav Dev 2002; 25: 452-465.
(81) Vereijken B. The complexity of childhood development: variability in perspective. Phys Ther 2010;
90: 1850-1859.
(82) Newell KM, Broderick MP, Deutsch KM, Slifkin AB. Task goals and change in dynamical degrees of
freedom with motor learning. J Exp Psychol Hum Percept Perform 2003; 29: 379-387.
(83) Pomeroy VM, Pramanik A, Sykes L, Richards J, Hill E. Agreement between physiotherapists on
quality of movement rated via videotape. Clin Rehabil 2003; 17: 264-272.
6
141
Chapter 6
Appendix 1
Nominal Group Technique procedure
1: Introduction and explanation
Welcome to the participants and explanation of the purpose and procedure of the
meeting. PowerPoint presentation of observable movement quality (OMQ) and
last slide with the results of the interviews: 25 aspects of OMQ (Tab. 3, column A).
Approximately 15 minutes.
2: Silent generation of ideas
Each physical therapist is provided with a sheet of paper with the question: Which
aspects of OMQ, do you think, are most important to assess OMQ? During this
period participants are asked not to consult or discuss their ideas with others.
Approximately 10 minutes.
3: Sharing ideas
Participants are invited to share their most important aspects and motivation.
The facilitator records each idea on the flip chart using the actual words of the
participant. At this stage, there is no debate on the aspects and participants are
encouraged to write down any new idea that may arise from what others share.
This procedure ensures that all participants have the opportunity to make an
equal contribution and provides written records of all ideas generated by the
group. About 15 minutes
4: Group discussion
Participants are invited to seek verbal explanation or further details about any
of the ideas that colleagues have produced that may not be clear to them. The
facilitator’s task is to ensure that each person is allowed to contribute and that
discussion of all ideas is thorough, without excessive time being spent on any one
idea. It is important to ensure that the procedure is as neutral as possible and that
judgment and criticism are avoided. The group combines aspects into categories
by combining aspects and eliminating aspects. About 45-60 minutes.
5: Voting and ranking
The participants individually prioritize the aspects in relation to the original
question. Ranking takes place by assessing the aspects on a scale of 1 to 10 (10
is most important and 1 the least important).
Sheets of paper listing the aspects were distributed to allow prioritizing; these
were individually scored and added up.
142
Development of a movement quality measurement tool for children
Appendix 2
Observable Movement Quality scale: Terminology, Definitions, and 5-point Likert
scale with description of Aspects
Automated movements (aspect 3)
The child has mastered the skills appropriate for his or her age in such a
way that these are consistent and can be executed without much attention,
if necessary in combination with another task or tasks.
1 Movements consistently not automated
2 Typically movements not automated; 1 or 2 instances of automated movements
3 No automated movements half of the time; automated movements half of the time
4 Typically automated movements; 1 or 2 instances of no automated movements
5 Consistently automated movements
Asymmetry in movements (aspect 5)
In movements and/or maintenance of posture, a body half or part of a body
half inadequately participates in the task. The difference in the use of body
parts does not fit with the age of the child and the demands of the tasks and
the environment.
1 Consistently asymmetric
2 Typically asymmetric; 1 or 2 instances of symmetry
3 Asymmetric half of the time; symmetric half of the time
4 Typically symmetric; 1 or 2 instances of asymmetry
5 Consistently symmetric
Variation in movements (aspects 7/8)
The child has an extensive repertoire of movements and can move body
parts relatively independently from each other in different directions, so
the necessary degrees of freedom are used to match the demands of the
tasks and the environment.
1 Consistently no variation
2 Typically no variation; 1 or 2 instances of variation
3 No variation half of the time; variation half of the time
4 Typically variation; 1 or 2 instances of no variation
5 Consistent variation
143
6
Chapter 6
Appropriate fine motor movements (aspect 11)
The child adapts his or her postures and movements to the demands of the
fine motor tasks and the environment.
1 Consistently inappropriate
2 Typically inappropriate; 1 or 2 instances of appropriate fine motor movements
3 Inappropriate half of the time; appropriate half of the time
4 Typically appropriate; 1 or 2 instances of inappropriate fine motor movements
5 Consistently appropriate
Appropriate gross motor movements (aspect 12)
The child adapts his or her postures and movements to the demands of the
gross motor tasks and the environment.
1 Consistently inappropriate
2 Typically inappropriate; 1 or 2 instances of appropriate gross motor movements
3 Inappropriate half of the time; appropriate half of the time
4 Typically appropriate; 1 or 2 instances of inappropriate gross motor movements
5 Consistently appropriate
Fluency of movements (aspects 13/14)
The movements of the child are controlled in such a manner that they are
adapted to the tasks and the environment in a fluent manner, without
faltering or stumbling.
1 Consistently not fluent
2 Typically not fluent; 1 or 2 instances of fluent movements
3 Not fluent half of the time; fluent half of the time
4 Typically fluent; 1 or 2 instances of not fluent movements
5 Consistently fluent
Reduced muscle tone (aspect 15)
The movements and/or maintenance of posture give the impression of
being slack and not adapted to the tasks and the environment.
1 Consistently low muscle tone; like a rag doll
2 Typically low muscle tone; 1 or 2 instances of normal muscle tone
3 Low muscle tone half of the time; normal muscle tone half of the time
4 Typically normal muscle tone; 1 or 2 instances of low muscle tone
5 Absence of low muscle tone
144
Development of a movement quality measurement tool for children
Increased muscle tone (aspect 16)
The movements and/or maintenance of posture give the impression of
being stiff/rigid and not adapted to the tasks and the environment.
1 Consistently high muscle tone; muscles are rigid and tight
2 Typically high muscle tone; 1 or 2 instances of normal muscle tone
3 High muscle tone half of the time; normal muscle tone half of the time
4 Typically normal muscle tone; 1 or 2 instances of high muscle tone
5 Absence of high muscle tone
Involuntary movements (aspect 17)
In moving, parts of the child’s body and/or those parts not directly involved
in the movements show unconscious movements that disrupt the proper
execution of the task and are not appropriate to the child’s age.
1 Consistently involuntary movements
2 Typically involuntary movements; 1 or 2 instances of normal movements
3 Involuntary movements half of the time; normal movements half of the time
4 Typically normal movements; 1 or 2 instances of involuntary movements
5 Absence of involuntary movements
Accuracy (well-aimed) (aspect 20)
The child moves his or her body parts so that the target is reached accurately
and immediately.
1 Target consistently not reached
2 Target typically not reached; 1 or 2 instances of target being reached
3 Target not reached half of the time; target reached half of the time
4 Target typically reached; 1 or 2 instances of target not being reached
5 Target consistently reached
6
Slow/delayed movements (aspect 24)
The child moves his or her body or a part of it at a lower speed than is suitable
for the task and can, despite instruction, not accelerate sufficiently.
1 Consistently slow and delayed movements
2 Typically slow and delayed; 1 or 2 instances of appropriate timing and pacing
3 Slow and delayed half of the time; appropriately timed and paced half of the time
4 Typically appropriate timing and pacing; 1 or 2 instances of slow and delayed
movements
5 Absence of slow and delayed movements
145
Chapter 6
Accelerated/abrupt movements (aspect 25)
The child moves his or her body or a part of it at a higher speed, or abruptly
at a higher speed than is suitable for the task and can, despite instruction,
not slow down sufficiently.
1 Consistently accelerated and abrupt movements
2 Typically accelerated and/or abrupt movements; 1 or 2 instances of appropriate
timing and pacing
3 Accelerated and/or abrupt movements half of the time; appropriately timed
and paced half of the time
4 Typically appropriate timing and pacing; 1 or 2 instances of accelerated and/
or abrupt movements
5 Absence of accelerated and abrupt movements
Tremors (aspect 27)
There is an involuntary, rhythmic, periodical, uncontrollable trembling of a
body part or body parts during the child’s movements, which can vary in
amplitude from barely observable to clearly visible, or in frequency from
low to high.
1 Constantly
2 Frequently
3 Occasionally
4 Infrequently
5 None
Strength regulation (aspect 28)
The movements of the child are in terms of force/strength well suited to the
task and the environment.
1 Strength consistently not adapted to task
2 S trength typically not adapted to task; 1 or 2 instances of strength adapted to
task
3 S trength not adapted to task half of the time; strength adapted to task half of
the time
4 S trength typically adapted to task; 1 or 2 instances of strength not adapted to
task
5 Strength consistently adapted to task
146
Development of a movement quality measurement tool for children
Stereotyped movements (aspect 29)
The child shows spontaneous, repetitive, non-purposeful movements (examples
include repeatedly turning and shaking of the head and/or wobble or flutter
of body parts).
1 Consistent stereotyped movements
2 Typically stereotyped movements; 1 or 2 instances of normal movements
3 Stereotyped movements half of the time; normal movements half of the time
4 Typically normal movements; 1 or 2 instances of stereotyped movements
5 Absence of stereotyped movements
Appendix 3
Representative Quotations per Aspect from the Semistructured Interviews,
Nominal Group Technique and Delphi Technique, Recognizable per Study Phasea.
1. Anticipation
Sometimes I deliberately put a lot of toys on the floor to see whether the child
can anticipate on the toys on the floor? For example, does the child actually look
at the toys or does he detect the toys in his peripheral vision? Does the child
adjust his movements or does the muscle tone increase? If the latter, the child is
already aware that something will go wrong [that the child will stumble over the
toys?]. So can the child judge the situation? (6b)
This is related to judgment and cognition. Even a child who has some
problems with the control of movement can still be a good anticipator –
within the framework of his own abilities. (8)
2. Associated reactions
What you are continuously searching for? Are those associated reactions what
you expect, based on the age of the child and the difficulty of the task? ….. this
also refers to experience. I am searching for the boundaries of what is normal
and what just exceeds normal to become a diskinetic movement. (5) This
approach provides me with information on the skills level of the child as well as
on possible impairments that are already present. (6)
3. Automated motor programs
Automated means that the child pays a small amount of attention to the movement
itself. It should be possible to perform a double task, but that is a bit too limited.
Attention is a cognitive process; consistency is something you can see, especially
because it concerns the consistency required to achieve the motor skill.
147
6
Chapter 6
4. Stable motor programs
If a child crawls and his attention is diverted from the motor task, for example, by
having a ball rolled alongside of him, how stable is the child in terms of the motor
program, namely, crawling? Therefore, how does the child cope with controlling
environmental issues? Or do you see the child stiffen, the crawling stops and/or
the movement quality deteriorates, or does the child not know how to deal with
it? How stable is the motor program under different circumstances? (6)
5. Asymmetry
Asymmetry in movement refers to the entire left side or the entire right side being
consistently retarded compared to the other side. Asymmetry in movement is
also associated with a specific age group, and in many tasks asymmetry is
functional. (2)
6. Coordination
Accuracy, strength regulation, and control of movements are all about coordination.
I would define the control of movements as movements that are temporally and
spatially aligned so that they continue fluently. Control is an important aspect but
overlaps with many other aspects. Both; coordination and control of movement
are terms which we want to avoid because they exist of multiple variables and
therefore are not useful as a solo aspect. Perchance these aspects are useful as
umbrella aspects.
7. Complexity
Complexity defined from the General Movements is the use of three different
axes; therefore, you are closer to dissociation and isolation. I find it a problem
that dissociation and isolation are not affiliated with complexity. (4) What do
you consider as variation? Is variation the situation in which one movement is
different from the other movement, while complexity is that situation in which you
use all three axes – or do you use different definitions? If you define complexity
from motor control theories, then you do this with degrees of freedom. (8)
8. Dissociation
Not pure extension in both joints but, for example, when one joint can flex and
the other can extend in the same movement. (5)
9. Variation
Are all of the movements you expect to be present in the performance, or is a
task performed in different ways, or can it be performed in different ways, or
perhaps it is not performed in different ways because the skill has already been
mastered. (1) Variability: one and the same skill may be performed in different
ways, in different positions, or under different environmental variables. (6)
Variation is the first thing to observe. If the movements are varied, there are
no pathological movements. (8)
148
Development of a movement quality measurement tool for children
Variation/complexity/fluency
These are actually new terminologies. Thirty years ago we talked about
isolation, variation, fluency and dissociation. I would combine complexity with
dissociation, with isolation and dissociation being more or less the same, and
combine variation.(4)
10. Efficiency
Efficient to me is being accurate and quick, with little effort. (6) Efficient
compared to what? Normal development would be, for example, grasping with
a pincer grasp, but a child with CP will not use a pincer grasp, but will rather
develop a tripod grasp. For this child, given his abilities, this is an efficient
movement, while efficient for a normally developing child would be a pincer
grasp. More energy is required for a child with CP to use a pincer grasp. (3)
Efficient means that you have the most adequate solution, but how should
you measure this? You must do exercise testing, which is something I cannot
observe. (7)
11. Eye–hand coordination
You do not actually have to look at an object to be able to grasp it. If the
child has seen the object in his peripheral field of vision, then he has already
prepared himself. If the child looks the other way at the moment he grasps
the object, in my role as a therapist this makes no difference as this is eye–
hand coordination in another dimension – and not the classical definition that
you actually should be looking at. It can also say something about advanced
control. For example, if you place the pins in a simple serial task and you look
three times at where the pins are, but the fourth time, because you already
know where the pins are, you can look ahead to where each pin should be
placed; i.e., your eyes are a step ahead of what your hands are doing. (6)
Fine motor movement is about skills with your hands. It says nothing about the
refinement of movement. Reaching, grasping, and manipulating are fine motor
tasks and should be named in the definition. Because manus (Latin) means hand,
talking about manipulation excludes movement with arms, eyes, and trunk, and
these should also be observed. Fine motor movement is indeed clear enough.
12. Gross motor movement
Gross motor movement stands for everything that is not manipulative, reaching
or grasping (with the hands). Gross gives a qualitative value to the movement,
gross means that the movement is not adjusted, but it can be very refined. The
name should be used consistently: gross is related to fine and large is related to
small. A large motor movement is also not a good description, so I choose gross
motor skills. The literature mostly uses gross.
149
6
Chapter 6
13. Fluency
With fluency I mean that the path is gradual without a beginning or end, without
faltering, stumbling or shocks, smoothly, dosed and matched to each other. (1)
Flexible also means pliable. ‘Without faltering’ is a better term than flexible or
smooth.
14 Timing
Timing means that you have adequate control for that skill, if you are on time
with your reactions, if the reactions are correct, if the reactions are not too much
in terms of strength, speed and therefore a sort of adequate regulation. I think
that there is overlap between timing and strength regulation, which are very
close to each other. (1)
15/16. Muscle tone
Do we look at muscle tone? That is a very difficult question. Do you see tone?
Secretly, you formulate it for yourself, what I observe has probably to do with
muscle tone. But if you look at a child with cerebral palsy, where we can see
the hypertonicity, it is clear you can actually see it. However, in children whose
movements are on the boundaries of normal, it is related to fluency, strength
regulation, and hypermobility. (6)
From a physiological point of view, muscle tone is not observable. Slack is what
you see, and the interpretation is a low muscle tone. Muscle activation pinpoints
a conscious process and can also be interpreted as the direction of muscle
power.
17. Involuntary movements
Whether the movements are performed selectively or dissociated or that all body
parts show involuntary movements. Involuntary movements do not always affect
the skill negatively, think of tightly pressing the lips together while threading beads.
That says something of movement quality, but the task is not disrupted. So in the
definition it is important to accent the disruptiveness of the involuntary movements
to the task performance.
18. Isolation
Whether the child can move a joint freely, therefore it is actually the complexity of
movement.(5)
19. Pathological (spastic) movements
Has much overlap with the aspect variation, can it be merged? The aspect
pathological (spastic) movement does not belong to the other aspects on the
list. What to do with a child who walks on his toes without spasticity? Pathological
patterns can also be choreatic or athetotic. Movement patterns originate from a
previous time period when we were thinking in terms of hierarchical theories of
motor development.
150
Development of a movement quality measurement tool for children
20. Placing/accuracy
Placing causes confusion with the term used in neurodevelopmental treatment.
I find the term ‘accurate’ to be more precise. It concerns target reaching and the
way the target is reached. It is important that it concerns the final position of the
body; therefore, tasks included are pointing, grasping, and jumping on a line, but
not those such as aiming at a target. When a target is achieved accurately but not
directly, you can still talk about accuracy.
21. Quantity, amount of movement
This is a difficult issue; the amount of movement heavily depends on the
time of testing, whether a child is very anxious, or whether the child is
almost asleep. So, this aspect is not indicative of movement quality. (4)
22. Range of movement
If someone learns a new task, such as skiing, then the person limits his
degrees of freedom, but he does not really have a small range of motion; in
contrast, he uses a small range of movement. (7)
23. Skills level
The skill level can be interpreted as the level of one task, but then you have
narrowed it down. You can also see it within a broader spectrum: can the
child perform all the skills taken to be normal for his age? Anticipation is
something that could fall within the skill level. (7)
24/25. Speed
I always report whether a child works very slowly or very quickly, as it can
certainly affect accuracy. Speed can also hide several things. A child who has
difficulties in balancing and, for example, has to walk on a line, performs this
task very quickly because he is unable to perform the task slowly. Or, on the
other hand, the child may work very slowly because associated reactions occur
or there is increased muscle tone when the child speeds up. (3)
Slow can still be normal, while delayed movement is abnormal. The instruction
that the child cannot correct should be added to the definition in order to avoid
misinterpretation if the child was in a hurry or tired/bored.
26. Spectrum motor program
Another cluster is spectrum motor programs, stable motor programs, and
the skills level of the task. (5) Motor programs are not observable. (7) I think
you can see motor programs: a child throwing overhand and one throwing
underhand are using two different motor programs. (6) We should realize
that this is an interpretation within the one theory. (7)
151
6
Chapter 6
27. Tremors
Tremors are described in neurology books and the ICF-CY. A tremor is not always
uncontrollable; children sometimes suppress a tremor through fixation on the
surface. A tremor can be divided into various subforms.
28. Strength regulation
Whether the child pushes on a pen sufficiently or uses too little pressure. Can
catch a ball and does not squeeze the ball completely but catches the ball just
between the hands. (5)
The more I think about it, if you look at some of those children with
neuromuscular diseases, even with the little power they have, the movement
can show complexity, fluency, and timing. (6)
We can observe strength and regulation as an essential part of movement quality.
29. Degrees of freedom/movement in all three axes
How to control the many joints and muscles of the body. The terminology
‘degrees of freedom’ could be used instead of ‘range of movement’ to avoid
confusion with a truly limited range of motion.
30. Stereotype movements
I tend to think of it as isolation, which is more a description of stereotype
movements. If you have a few degrees of freedom because you have muscle
spasm or decreased muscle strength, then you only have a few options
for performing the task in that context and so movement will always be the
same. (7)
Stereotyped behavior is always the same act and not just movement. Repeated
shaking and untargeted movement without a purpose can also be a tic. Thus,
it is more a behavioral characteristic, expressed in movement, but is says less
about movement quality. You observe it and you should describe it. Stereotype
movements can be disruptive in terms of posture and movement. The aspect
‘stereotype movements’ is of added value, and there is a difference between
involuntary movements and stereotype movements.
31. Trunk rotation
In both younger and older children, rotations are part of movement quality, the
more you move without trunk rotation, the less fluent the movement. (1) Can a
child rotate his shoulders and pelvis independently? This can be seen in the
walking movement of a 5-year-old child: the shoulders are forward and there is
no smooth movement of the trunk. Trunk rotations are quite an important base.
(3) Trunk rotation is dependent on tasks, posture, and environmental factors
and how a child controls his motor skills. This is one aspect I do not value highly.
I am considering trunk rotation, it is an important base. A baby sliding on
his buttocks (3) …. has symmetry (7) ... and too little variation. (4)
Trunk movements need not be mentioned separately; they relate to the whole
body and may, in my opinion, be combined with variation.
152
Development of a movement quality measurement tool for children
32. Postural stability
Is the child capable of adjusting his postural control to different surfaces?
I think of balance; that the child learns to adjust to disturbances that he
experiences every day. Sometimes, when observing young children walking
you see clearly that they slow down in advance and almost dribble because
they are hardly able to stand still. So they really practice balance strategies
and implement these in their action planning. (6) Can the child reach outside
the boundaries of the supporting surface? What happens in a balance game
when the child is suddenly let go of, or experiences similar actions? (7)
Postural stability is not an isolated aspect, and is included in other definitions
because of the sentences-parts; “maintenance of posture..” or “adapts his or her
posture and movements.”
a
The
quotations from the semistructured interviews are in normal font, those from the nominal group
technique are in bold and those from the Delphi technique in italics. Numbers in parentheses (1-8)
represent pediatric physical therapists 1 to 8 who participated in the semistructured interviews
and nominal group technique. “He” means “he or she” and is ‘used for purposes of simplicity.
CP=cerebral palsy, ICF-CY=the International Classification of Functioning, Disability and Health
for Children and Youth.
6
153
Part IV
Chapter 7
Summary
Summary
Summary
The aim of our studies was to determine longitudinal motor performance in very
preterm infants in the context of the follow-up. We studied the influence of the
child’s behavior during the test (‘test-taking’ behavior) and risk factors on motor
performance. We hypothesized that the efficiency of follow-up could be
improved by a prediction model, predicting delayed motor performance taking
risk factors into account. We also determined the predictive validity of two infant
motor tests at 6, 12 and 24 months and their combinations for outcome at 5
years of age. In addition, we developed a concise measurement tool for
longitudinal evaluation of movement quality by pediatric physical therapists.
The studies are the result of the neonatal follow-up outpatient clinic of infants
admitted to the neonatal intensive care of the Radboud University Nijmegen
Medical Centre, Nijmegen, The Netherlands.
We included preterm infants, born before 32 weeks of gestation. The first
cohorts consisted of surviving infants born between January 1996 and
December 2001 assessed at 24 months corrected age (chronological age
reduced by the number of days born before 40 weeks of gestation) and the
chronological age of 5 years. The second cohort consisted of infants born
between March 2003 and December 2006, assessed at the of 6, 12, 24 months
corrected age and the chronological age of 5 years.
The pediatric physical therapist assessed motor performance, test-taking
behavior and movement quality. At 6, 12 and 24 months corrected age motor
performance was assessed with the Bayley Scales of Infant development, 2nd
edition Motor Scale (BSID-II-MS) and Behavior Rating Scale (BSID-II-BRS) total
and motor quality subscale score (MQ).1 At the corrected age of 6 and 12 months
the Alberta Infant Motor Scale (AIMS)2 was used as well. At the chronological
age of 5 years the Movement Assessment Battery for Children (M-ABC),3 and in
the second cohort the 2nd edition, Dutch version (MABC-2-NL) were recorded.4;5
Chapter 1 introduces the subject with background information and the aims
and outline of the thesis. This chapter focuses on the definition and terminology
of preterm birth, the history of follow-up in Nijmegen and the protocol used. We
described what is known of follow-up studies, motor performance outcome,
159
7
Chapter 7
movement quality and risk factors for a delayed motor performance.
In Part I, Chapter 2 cross-sectional motor performance in 437 preterm infants
in the cohort born between January 1996 and December 2001 was described.
Children with a severe disability were excluded. The influence of ‘test-taking’
behavior and 21 risk factors for delayed motor performance at 2 years of age
was determined in binary logistic regression analysis. Motor performance in
terms of the mean BSID-II-MS, Psychomotor Developmental Index (PDI) at 2
years of age (mean 29 months [SD3.3]) was 86, almost one standard deviation
below the reference population. ‘Delayed’ motor performance was observed in
about 47% of the infants, and ‘test-taking’ behavior was ‘not-optimal’ in about
31%. Motor performance and ‘test-taking’ behavior scores were significantly
correlated, indicating that a better motor performance was present if behavior
was more normal. Four of the 21 risk factors contributed to a delayed motor
performance at 2 years of age: neonatal convulsions (odds ratio [OR] 4.5; 95%
confidence interval [CI] 1.6–12.9), low maternal educational level (OR 3.3; 95%
CI 1.7–6.5), male gender (OR 2.8; 95% CI 1.8–4.3), and chronic lung disease (OR
2.1; CI 1.1– 4.1).
Part I, Chapter 3 describes longitudinal motor performance in 348 preterm
infants in the cohort born between March 2003 and December 2006. In contrast
to the study in chapter 1, the aim was to detect the influence of risk factors on
motor performance trajectories at 6, 12 and 24 months. The BSID-II-MS raw
score was the dependent variable in random coefficient analysis for risk factors
in the cohort if infants with cerebral damage were in- and excluded in the
analysis. The BSID-II-MS raw score increased, showed the highest correlation
over 3 assessments (rp= 0.48–0.67) and was more stable than the PDI and its
classification. Fifteen percent of the infants had a stable classification, while
45% changed one class. Male gender and intra-ventricular hemorrhage (IVH)
were related to lower raw motor scores. Higher motor quality scores and height
were related to higher raw motor scores. Low maternal education was related to
lower raw scores at 6 and 24 months and higher raw scores at 12 months.
Removal of infants with cerebral damage from the analysis and thereby
excluding IVH, did not change the other four risk factors. The results showed
that the BSID-II-MS raw score trajectories were relatively stable, but after
corrections for norm data, the measurements became highly unstable. This is
160
Summary
clinically important when reporting results to parents, guiding intervention and
in randomized trials. The risk factors predominantly influenced the level of motor
performance raw scores. Maternal education additionally influenced the
trajectory and should be included in randomisation procedures.
In Part II, Chapter 4 an algorithm was developed to decide if children should be
reassessed in the follow-up at 5 years of age in the cohort born between January
1996 and December 2001. Children with severe disabilities were excluded.
Twenty neonatal risk factors, maternal education level, BSID-II-MS and
BSID-II-BRS total and motor quality subscale scores at 2 years were used for
the prediction in binary logistic regression analysis (n=345). We defined a
‘delayed’ motor performance as below the 15th percentile on the Movement
Assessment Battery for Children (M-ABC) at 5 years of age. Five factors
contributed significantly (p < 0.05) to the model: the BSID-II-MS, PDI < 90 and
a BSID-II-BRS, ‘motor quality’ subscale < 26 percentile at 2 years, and the
risk factors gestational age < 30 weeks, male gender and intra-ventricular
hemorrhage. The prediction model improved the efficiency of motor assessment
by 37%, with a sensitivity of 94% and a specificity of 50%. Applying this model,
we would not have assessed 129 children and would have missed six children
with a delayed motor performance.
In Part II, Chapter 5 we compared the predictive validity of two infant motor
tests used at 6 (t 0), 12 (t1),
and 24 (t 2) months corrected age for motor
performance at 5 years (t3), in the cohort born between March 2003 and
December 2005. Children with known disabilities were excluded.
Children were assessed at t 0, t1, and t 2 with the BSID-II-MS, BSID-II-BRS and its
motor quality (MQ) subscale, at t 0 and t1 with the AIMS, and at t3 with the
MABC-2-NL. The dichotomized performance normal versus delayed/abnormal,
per predictive test and per combination of tests were examined.
At 5 years of age the MABC-2-NL (n=201), mean total standard score was ½
standard deviation below the reference population. The MABC-2-NL categorized
74 (37%) children as delayed/abnormal. The AIMS ≤10 percentile showed the
best combination of sensitivity and specificity (t1=65%, 68%). The BSID-II-MS
≤84 had the best sensitivity (t 0=79%, t1=93%, t 2=63%) and the BSID-II-BRS ≤25
percentile had the best specificity (t 0=87%, t1=90%, t 2=81%). At t 0 the
combination AIMS ≤10/BSID-MS ≤69/BRS ≤25, at t1 the AIMS ≤10/BSID-MS ≤69/
161
7
Chapter 7
MQ ≤25 and at t 2 the BSID-MS ≤69/MQ ≤25 achieved the best balance between
sensitivity and specificity (t 0=78%, 43%; t1=78%, 62%; t 2=74.0%, 62%).
Combining tests improved the predictive validity and the predictive validity was
lowest at 6 months.
Part III, Chapter 6 describes the development of a movement quality
measurement tool for children. Movement quality in preterm infants differs from
that of term infants. Pediatric physical therapists describe movement quality
because a concise measurement tool of observable movement quality (OMQ) is
lacking. The purpose of this study was to develop an OMQ measurement tool
for children from the perspective of pediatric physical therapists. A qualitative,
3-phase study involving pediatric physical therapists was conducted. The first
phase consisted of 7 semistructured interviews. The second phase comprised
a structured meeting using a Nominal Group Technique, with the interviewees
required to identify the most relevant OMQ aspects. The third phase comprised
a Delphi technique involving 61 pediatric physical therapy experts with the aim
of achieving at least 80% agreement on relevance, terminology and definitions
of OMQ aspects.
Across all three phases, 32 aspects based on different theoretical constructs
were considered. Fifteen aspects were included in the measurement. The
pediatric physical therapy experts achieved at least 80% agreement on the
definition of 14 OMQ aspects. The definition of appropriate fine motor movements
achieved 75% agreement. This aspect was included because gross and fine
motor movements are complementary. The aspects were scored using a 5-point
Likert scale, with a total score ranging from 15 to 75 and with a higher score
indicating a better OMQ. The content validity was obtained, but before the OMQ
scale can be used in clinical practice, studies on reliability, construct validity
and responsiveness are needed.
Conclusions
- At 2 years of age a delayed motor performance was observed in 47% of the
children, and ‘test-taking’ behavior was not-optimal in about 31%, in the first
cohort.
- A delayed motor performance was observed at 6 months in ±75%, at 12
162
Summary
months in ±78%, at 24 months in ±50% of the infants, in the second cohort.
- At 5 years of age a delayed motor performance was observed in 30% of the
children in the first cohort and in 37% of the children in the second cohort,
both excluding children with disabilities.
- After corrections for norm data, the BSID-II-MS trajectories at 6, 12 and 24
months are unstable.
- Risk factors for a delayed motor performance at 2 of age are neonatal
convulsions, low maternal educational level, male gender and chronic lung
disease.
- Motor trajectories in the first two years of life were influenced by male gender
and intra-ventricular hemorrhage (IVH) which were related to lower BSID-II-MS
raw scores. Higher motor quality scores and height were related to higher raw
motor scores, while maternal education was related to lower raw scores at 6
and 24 months and higher raw scores at 12 months.
- A delayed motor performance at 5 years of age was predicted by a BSID-II-MS,
PDI < 90 and a BSID-II-BRS, ‘motor quality’ subscale < 26 percentile at 2
years, and the risk factors gestational age < 30 weeks, male gender and
intra-­ventricular hemorrhage.
- The correlation of assessments in preterm infants from 6, 12, 24 months to 5
years of age is low to moderate.
- Combining tests in the first two years of life improved the predictive value. At
6 months the combination AIMS ≤10/BSID-MS ≤69/BRS ≤25, at 12 months the
AIMS ≤10/BSID-MS ≤69/MQ ≤25 and at 24 months the BSID-MS ≤69/MQ ≤25
achieved the best balance between sensitivity and specificity in predicting
motor outcome at 5 years of age.
- The assessment at 6, 12 and 24 months showed similar sensitivity and
specificity with the lowest predictive validity at 6 months.
- The Observable Movement Quality scale (OMQ scale) including 15 defined
aspects of movement quality from the pediatric physical therapists’ perspective
was developed as a concise measurement of observable movement quality
in children.
7
163
Chapter 7
References
(1) Bayley N. Bayley Scales of Infant Development, second edition. San Antonio: The Psychological
Corporation; 1993.
(2) Piper MC, Pinnell LE, Darrah J, Maguire T, Byrne PJ. Construction and validation of the Alberta
Infant Motor Scale (AIMS). Can J Public Health 1992; 83 Suppl 2: S46-S50.
(3) Henderson SE, Sugden DA. Movement Assessment Battery for Children. London: Psychological
Corporation, Harcourt Brace Jonanovich; 1992.
(4) Henderson SE, Sugden DA, Bernett AL. Movement Assessment Battery for Children−second
edition (Movement ABC-2). London: Pearson Education, Inc.; 2007.
(5) Henderson SE, Sugden DA, Barnett AL, Smits-Engelsman BCM. Movement Assessment Battery
for children−second edition, Dutch version. Amsterdam: Pearson Assessment and information
B.V.; 2010.
164
Summary
7
165
Chapter 8
General discussion
General discussion
General discussion
Insight in longitudinal motor performance in very preterm infants is important to
prevent both over- or under-referral for pediatric physical therapy inter­ventions
and to be able to measure the effect in future intervention studies. Therefore, we
studied the influence of neonatal risk factors, the influence of qualitative aspects
of the child’s behavior during the test (‘test-taking’ behavior), environmental and
personal factors, as well as the role of movement quality on longitudinal motor
performance and the prediction of motor performance outcome.
We performed follow-up studies in two cohorts, which included motor tests, a
questionnaire filled in by the PPT of ‘test-taking’ behavior including movement
quality, a series of neonatal risk factors and the maternal educational level at 6, 12,
24 months corrected age and a motor test at the chronological age of 5 years.
Risk factors influencing motor performance
In a cross-sectional study four of the 21 risk factors that we studied contributed
to a delayed motor performance at 2 years of age: neonatal convulsions, low
maternal educational level, male gender, and chronic lung disease (CLD). In a
longitudinal study at 6, 12 and 24 months male gender and intra-ventricular
hemorrhage (IVH) were the factors that were related to lower raw motor scores.
Higher motor quality scores and height were related to higher raw motor scores,
while maternal education was related to both lower and higher raw scores at
different time points.
From these studies we conclude that when motor performance at that time is
still normal, the presence of risk factors can be an indication to start PPT intervention. In addition, intervention studies should be stratified for CLD, gender,
IVH, motor quality and neonatal convulsions as well as maternal education.
Two risk factors, male gender and a low maternal educational level, influenced
motor performance in both the cross-sectional and in the longitudinal analysis.
The question is why male gender is a specific risk factor for delayed motor
performance in preterm infants. So far, motor tests do not include separate
reference values for boys and girls. We found a larger standard deviation of
169
8
Chapter 8
scores for motor performance in boys than in girls. It might be that male gender
is not only a risk factor for preterm infants but for the general population as well.
The reference group for the Dutch version of the MABC-2 included the children
stratified to the paternal educational level.1 However, we included maternal
education and although related to paternal education, we cannot compare our
cohort to the reference group of the MABC-2-NL at this point.
The growth measurements weight, height and head circumference were not
included in the cross-sectional study. However, based on the increasing
evidence for the relation between growth and motor performance we included
these three factors in the longitudinal study. 2;3 In our cohort height and not head
circumference was related to the motor performance score. In the Cooke study
height was found related only in preterm boys.3 The fact that we excluded infants
suspected of brain injury might explain why we did not find head circumference
to be a significant risk factor.
Since we used the raw motor scores, we expect that the risk factors influencing
longitudinal motor performance will remain the same if other motor tests are
used.
Stability of longitudinal motor performance
assessments
In the first 2 years of life the norm corrected longitudinal motor performance in
preterm infants on the BSID-II-MS was unstable. Longitudinal instability in motor
tests was also found in typically (‘normal’) developing children.4;5 The instability
in typically developing children can be explained by the non-linear motor
development conform the dynamic systems theory.6 The discriminative motor
tests use reference values based on cross-sectional norm data. So, to get more
insight into the variation of motor performance trajectories norm data should be
gathered longitudinally. Furthermore, such longitudinal data are clinically more
relevant, because children during PPT interventions are always followed
longitudinally, with typically developing children as a reference.
In randomized controlled intervention studies the BSID-II often fails to reveal
intervention effects.7 Although it might be that the interventions in itself are not
effective also the instability of repeated measurements during longitudinal
evaluation could explain its negative findings.
170
General discussion
Evaluative tests such as the Gross Motor Function Measure for children with
cerebral palsy (CP) are sensitive to detect small changes in motor performance.
This test is specifically designed to establish intervention effects in children with
CP and therefore, by this test the measured skills shows a ceiling effect for
children without CP. It would be interesting to study whether a discriminative
test with more evaluative characteristics is also more sensitive in children with
motor performance problems but without CP.
PPTs need better motor tests combining both discriminative and evaluative
characteristics. It might help if the steps between the successive items of a
motor test are smaller for infants and children with motor performance difficulties.
The predictive value of motor performance
assessments for outcome at 5 years of age
We developed a prediction model that improved the efficiency of the follow-up
for motor assessments by 37% with a high sensitivity of 94% accepting a
specificity of only 50%. Using this prediction model in our cohort we would not
have assessed 129 out of 345 children and would have missed six children with
a delayed motor performance. However, prediction needs to be further
investigated since the BSID-II we used in this study was revised. This third
version of the Bayley Scales of Infant and Toddler Development is used in the
follow-up now. The impact of prediction models becomes questionable if tests
are replaced by newer editions. However, our method for development of
predictive models can still be used in future research.
Since human movement is complex and influenced by the interaction of the
individual with the task and the environment,8 prediction of motor development
will remain difficult. In our studies we included many personal factors but only
one external factor, maternal education level. We think that in future research
other external factors, for example the expectancy of the parents, the intervention
received and the cultural background should be included.
Prediction is influenced by many factors such as the stability in the trajectories,
the influence of neonatal risk factors and the motor tests and cut-off scores
used. We made a first step in comparing different cut-off scores for motor tests
171
8
Chapter 8
and combining them for prediction. Combining the AIMS, BSID-II-MS and
BSID-II-BRS total score and the subscale score MQ improved the balance
between sensitivity and specificity for prediction of motor performance. We are
the first to show that combinations of these factors, which PPTs use in clinical
reasoning, indeed improved prediction.
The role of movement quality
In the study described in chapter 4, movement quality evaluation based on only
eight questions at 2 years of age was, to our surprise, already predictive in
children without severe disabilities for motor performance at 5 years. We
therefore performed a qualitative study to develop a generic tool for measuring
movement quality. In future studies this tool for observable movement quality
evaluation and prediction of motor performance needs to be explored.
Implications for clinical practice and policy
Due the increasing possibilities of medical care and the lowering of the limit for
active intervention to 24 weeks gestation in the Netherlands, more children will
survive with longer NICU treatment. Many of these will need pediatric physical
therapy intervention to improve motor performance. Prediction of motor
performance will remain an important issue to limit the number of children who
need pediatric physical therapy interventions.
Assessments at 6 and 12 months make the evaluation at 24 months more
representative for the actual outcome. More children responded if the first
evaluation was at 6 months. Especially children with impairments were detected
earlier and parents were more motivated to return for follow-up visits. Based on
our findings we suggest not to delay the first follow-up until 24 months of age,
which sometimes is proposed to reduce costs.
For international comparison test procedures should not be changed during the
translation of tests into another language, as happened during the translation of
BSID-II in Dutch.
In the decision process for intervention PPTs should continue to use the outcome
of motor tests combined with observable movement quality and the risk factors
for delayed motor performance.
172
General discussion
In order to establish more uniform indications for pediatric physical therapy
interventions and treatment frequency an evidence based guideline needs to be
developed.
Limitations
The follow-up studies in this thesis were not standardized for already applied
interventions. Many children received all kinds of interventions and at different
treatment frequencies, which of course influenced our findings.
The studies were performed in a single university hospital limiting the generalizability of our findings. Attrition bias in longitudinal studies is always present
which results in underestimation of delayed motor performance.
In the studies of this thesis the BSID-II mental scale was not included. We
explored the predictive value of the motor and the mental scale at 2 years of age
for outcome at 5 years in the first cohort (unpublished data). However, the
mental scale included different test versions during the years of follow-up. The
BSID-II-MS at 2 years of age predicted both delayed cognitive and motor
outcome at 5 years. In contrast, the BSID-II mental scale at 2 years did not
predict all children with delayed motor performance, so motor performance was
a stronger predictor for both cognitive and motor development. This needs to
be investigated in the future.
We included in the studies the ‘test-taking’ behavior but the behavior of the child
at home or in school situations may differ from the test situation.
Recommendations for future research
- We recommend to develop multidisciplinary prediction models for preterm
infants. This might reduce the number of children that have to be included in
a follow-up program.
- To improve the efficiency of the follow-up program, studies to establish the
value of the Ages and Stages Questionnaire (ASQ) need to be performed.
The ASQ is a parent-completed child-monitoring system to evaluate development.
- The role of movement quality in prediction and evaluation of intervention
effects needs to be further investigated.
- Intervention studies to improve motor performance in children should not
be started until better motor tests combining discriminative with evaluative
173
8
Chapter 8
characteristics are available. To be more sensitive the tests needs to evaluate
motor performance in smaller steps between successive items. More
knowledge about longitudinal variation of typically developing infants is a
prerequisite.
- The difference between cross-sectional and longitudinal norm data for motor
tests, separately for boys and girls should be investigated to determine the
rate and the variation of longitudinal motor performance in typically developing
children.
174
General discussion
References
(1) Henderson SE, Sugden DA, Bernett AL. Movement Assessment Battery for Children−second
edition (Movement ABC-2). London: Pearson Education, Inc.; 2007.
(2) Badr LK, Bookheimer S, Purdy I, Deeb M. Predictors of neurodevelopmental outcome for preterm
infants with brain injury: MRI, medical and environmental factors. Early Hum Dev 2009; 85:
279-284.
(3) Cooke RW, Foulder-Hughes L. Growth impairment in the very preterm and cognitive and motor
performance at 7 years. Arch Dis Child 2003; 88: 482-487.
(4) Darrah J, Hodge M, Magill-Evans J, Kembhavi G. Stability of serial assessments of motor and
communication abilities in typically developing infants−implications for screening. Early Hum Dev
2003; 72: 97-110.
(5) Roze E, Meijer L, VAN Braeckel KN, Ruiter SA, Bruggink JL, Bos AF. Developmental trajectories
from birth to school age in healthy term-born children. Pediatrics 2010; 126: e1134-e1142.
(6) Thelen E, Ulrich BD. Hidden skills: a dynamic systems analysis of treadmill stepping during the first
year. Monogr Soc Res Child Dev 1991; 56: 1-98.
(7) Orton J, Spittle A, Doyle L, Anderson P, Boyd R. Do early intervention programmes improve
cognitive and motor outcomes for preterm infants after discharge? A systematic review. Dev Med
Child Neurol 2009; 51: 851-859.
(8) Shumway-Cook A, Woollacott MH. Motor control: issues and theories. Motor Control; Translating
research into clinical practice, third edition. Philadelphia, Pennsylvania: Lippincott Williams &
Wilkins; 2007.
175
8
Nederlandse samenvatting (Summary in Dutch)
Dankwoord (Acknowledgements)
Curriculum Vitae
List of publications
Samenvatting in het Nederlands (Summery in Dutch)
Nederlandse samenvatting
Summary in Dutch
Het doel van de studies was het motorisch niveau van vroeg prematuur geboren
kinderen te bepalen gedurende het follow-up programma vroeggeborenen. We
onderzochten de invloed op het motorische niveau van zowel het gedrag van
het kind tijdens de motorische test alsmede de invloed van een groot aantal
risicofactoren. Onze veronderstelling was dat de efficiëntie van het follow-up
programma zou kunnen worden verbeterd met een voorspellend model, waarin
naast de motorische test ook relevante risicofactoren zijn opgenomen. De
voorspellende waarde van diverse motorische testen op 6, 12 and 24 maanden,
en combinaties van die testen werden bepaald voor de motorische uitkomst op
de leeftijd van 5 jaar. Daarnaast ontwikkelden we een meetinstrument voor kinderfysiotherapeuten om de kwaliteit van bewegen zo objectief mogelijk vast te
leggen.
In de studies zijn de vroeg prematuur geboren kinderen die werden opgenomen
op de neonatale intensive care unit van het Universitair Medisch Centrum St
Radboud Nijmegen vervolgd. Vroeg prematuur geboren is gedefinieerd als een
zwangerschapsduur van minder dan 32 weken. Het eerste cohort bestond uit
kinderen geboren tussen januari 1996 en december 2001. Dit cohort werd
onderzocht op de gecorrigeerde leeftijd (kalenderleeftijd verminderd met het
aantal dagen te vroeg geboren) van 2 jaar en op de kalenderleeftijd van 5 jaar.
Het tweede cohort bestond uit kinderen geboren tussen maart 2003 en
december 2006. Dit cohort werd vaker onderzocht, op de gecorrigeerde leeftijd
van 6, 12 en 24 maanden en op de kalenderleeftijd van 5 jaar.
Een kinderfysiotherapeut bepaalde het motorisch niveau, het gedrag van het
kind tijdens de test en de kwaliteit van bewegen. We gebruikten zo mogelijk
betrouwbare en valide motorische testen om vast te leggen hoe het kind scoorde
ten opzichte van een op tijd geboren groep kinderen. Op de gecorrigeerde
leeftijd van 6, 12 en 24 maanden werd gebruik gemaakt van twee onderdelen
van de Bayley Scales of Infant Development, tweede versie: de Motor Scale
(BSID-II-MS) en de Behavior Rating Scale (BSID-II-BRS). Op de gecorrigeerde
leeftijd van 6 en 12 maanden werd ook de Alberta Infant Motor Scale (AIMS)
afgenomen. Op de kalenderleeftijd van 5 jaar werd in het eerste cohort de
179
Samenvatting in het Nederlands (Summery in Dutch)
Movement Assessment Battery for Children (MABC), eerste versie, gebruikt, in
het tweede cohort de tweede versie met Nederlandse normwaarden (MABC-2-NL).
In hoofdstuk 1 wordt de algemene achtergrondinformatie, het doel en de opzet
van dit proefschrift beschreven. Er wordt ingegaan op de definities en terminologie
bij vroeggeboorte, de geschiedenis van het neonatale follow-up programma in
Nijmegen, en het toegepaste protocol. We beschrijven wat er vanuit de literatuur
over follow-up studies bekend is over het motorisch niveau en de kwaliteit van
bewegen van vroeg prematuur geboren kinderen.
De Wereldgezondheidsorganisatie definieert vroeggeboorte als een geboorte
vóór de 37ste complete zwangerschapsweek (259 dagen zwangerschapsduur),
gemeten vanaf de eerste dag van de laatste normale menstruatie. In Nederland
wordt 7-8% van alle kinderen te vroeg geboren. Hoe korter de zwangerschap,
hoe minder de organen van het kind in staat zijn hun taak buiten de baarmoeder
te vervullen. Dit leidt tot (ernstige) problemen met de ademhaling, de bloeds­omloop, de zuurstofvoorziening van de hersenen en andere vitale organen,
infecties door onvoldoende ontwikkeling van het afweersysteem, voedings­
problemen door een maag-darmstelsel dat voeding nog slecht verdraagt en
onderkoeling omdat de lichaamstemperatuur niet op peil kan worden gehouden.
Voor kinderen geboren na een zwangerschapsduur van minder dan 32 weken
(vroeg prematuur) is vaak zeer intensieve zorg nodig.
Na de opkomst van de neonatale intensieve zorg in de jaren zeventig van de
vorige eeuw overleven er nu veel meer vroeg prematuur geboren kinderen.
Aanvankelijk lag de nadruk van studies vooral op overleving, maar met de afname
van de sterfte verschoof de aandacht naar de gevolgen van de vroeggeboorte
op de latere ontwikkeling van het kind.
In opdracht van het Ministerie van Volksgezondheid zijn de neonatale intensive
care units sinds 1993 verplicht deze kinderen te volgen. De Landelijke Neonatale
Follow-up werkgroep adviseert kinderen geboren na een zwangerschapsduur
van minder dan 32 weken te volgen op de gecorrigeerde leeftijd van 6, 12 en 24
maanden en de kalenderleeftijd van 5 en 8 jaar. Een multidisciplinaire follow-up
polikliniek werd in Nijmegen in 1995 gestart, waarbij aanvankelijk de kinderen
werden teruggezien op hun 2e en 5e jaar. Vanaf 2003 is besloten de follow-up
uit te breiden met onderzoek op de leeftijd van 6 en 12 maanden.
Het is bekend dat het motorisch niveau van vroeg prematuur geboren kinderen
180
Samenvatting in het Nederlands (Summery in Dutch)
vertraagd is en dat de kwaliteit van bewegen anders is dan die van op tijd
geboren kinderen, zelfs na correctie van de leeftijd. Ondanks een toegenomen
overleving door betere behandelingsmogelijkheden blijven de ontwikkelingsproblemen bij vroeg prematuur geboren kinderen relatief constant. Ongeveer
de helft van de kinderen heeft motorische, cognitieve of gedragsstoornissen.
Bij 5 tot 15% is er sprake van een spastische cerebrale parese. Het percentage
kinderen dat bijzonder onderwijs volgt loopt op, afhankelijk van de zwangerschapsduur en de leeftijd tot wel 64%.
In Deel I, Hoofdstuk 2 wordt het motorisch niveau van 437 vroeg prematuur
geboren kinderen uit het eerste cohort beschreven. Kinderen met beperkingen,
zoals blinde kinderen, kinderen met chromosomale afwijkingen, neuromusculaire aandoeningen of cerebral palsy werden geëxcludeerd. De invloed van het
gedrag tijdens de test en 21 geselecteerde risicofactoren voor een vertraagd
motorisch niveau bij vroeg prematuur geboren kinderen op de leeftijd van 2 tot
3 jaar werden onderzocht, gebruik makend van binaire logistische regressie
analyse. We vonden een vertraagde motorische ontwikkeling op de BSID-II-MS
bij 47% van de kinderen. Het gedrag tijdens de test op de BSID-II-BRS was bij
31% van de kinderen niet optimaal. Het motorisch niveau en het gedrag tijdens
de test bleken te correleren, dus het motorisch niveau tendeert naar normaal als
het gedrag tijdens de test meer normaal was. Vier van de 21 risicofactoren
droegen significant bij aan een vertraagd motorisch niveau: neonatale
convulsies (odds ratio [OR] 4.5; 95% betrouwbaarheidsinterval [BI] 1.6-12.9),
laag opleidingsniveau van moeder (OR 3.3; 95% BI 1.7-6.5), mannelijk geslacht
(OR 2.8; 95% BI 1.8-4.3)en de aanwezigheid van een chronische longziekte (OR
2.1; 95% BI 1.1-4.1).
Deel I, Hoofdstuk 3 beschrijft de motorische ontwikkelingstrajecten van 348
vroeg prematuur geboren kinderen uit het tweede cohort. In tegenstelling tot de
studie in hoofdstuk 1 was het doel de invloed van risicofactoren op het motorisch
traject te onderzoeken, met metingen op de leeftijden van 6 en 12 en 24
maanden. De ruwe score van de BSID-II-MS werd gebruikt in multilevel analyse
en meer specifiek in random coëfficiënt analyse. Dit werd gedaan voor het
gehele cohort maar ook na uitsluiting van de kinderen met (een groot risico op)
cerebrale schade.
De resultaten lieten zien dat de ruwe score van de BSID-II-MS bij alle kinderen
181
Samenvatting in het Nederlands (Summery in Dutch)
toenam in de tijd, de hoogste correlatie liet zien over de 3 metingen (rp=0.48-0.67)
en het meest stabiel was. De motorische ontwikkelingsindex (PDI score),
bepaald vanuit de ruwe score met behulp van de normtabellen uit de handleiding
van de test bleek zeer instabiel. De classificatie van de PDI: sterk vertraagd,
vertraagd en een normaal ontwikkelingsniveau over de drie meetmomenten
was stabiel in 15% van de kinderen. Bij 45% van de kinderen verschoof de
classificatie over de drie meetmomenten één klasse. De overige 40% van de
kinderen was de classificatie nog instabieler, verschoof over meerdere klassen.
De factoren mannelijk geslacht en het doorgemaakt hebben van een hersen­
bloeding rondom de geboorte gaven lagere ruwe motorische scores. Een
hogere score voor de kwaliteit van de motoriek en een grotere lengte in cm
gaven juist hogere ruwe motorische scores. Een laag opleidingsniveau van de
moeder gaf lagere scores op 6 en 24 maanden en hogere scores op 12 maanden.
Na uitsluiten van kinderen met (een groot risico op) cerebrale schade, bleven de
andere vier risicofactoren en hun invloed gelijk. De resultaten lieten zien dat de
ruwe scores van de BSID-II-MS relatief stabiel zijn, maar dat na correctie met
normdata de ontwikkelingsindex en de bijbehorende classificatie instabiel was.
Dit is klinisch relevant voor het rapporteren van uitslagen van de BSID-II-MS
naar ouders, voor de evaluatie van motorische interventies in het kader van de
kinderfysiotherapeutische behandeling en bij evaluatie van het effect van die
behandelingen in gerandomiseerd onderzoek. De risicofactoren beïnvloeden
voornamelijk het niveau van de motorische ruwe scores. Het opleidingsniveau
van moeder beïnvloedt het traject en zou meegenomen moeten worden in
­randomisatie-procedures voor interventieonderzoek.
In Deel II, Hoofdstuk 4 wordt een algoritme ontwikkeld om te bepalen of vroeg
prematuur geboren kinderen zouden moeten worden opgeroepen voor follow-up
onderzoek op de leeftijd van 5 jaar. Dit werd gedaan met data uit het eerste
cohort. Kinderen met beperkingen werden geëxcludeerd. Twintig neonatale
risicofactoren, het opleidingsniveau van moeder, de BSID-II-MS en de BSID-II-BRS
op 2 jaar werden gebruikt voor de voorspelling in binaire logistische regressie
analyse (n=345). We definieerde het motorisch niveau als vertraagd als het kind
op 5 jaar onder de 15 percentiel op de Movement Assessement Battery for
children (MABC) scoorde. We wilden kinderen die vertraagd scoorde zoveel
mogelijk terugzien en accepteerden daarom dat we daarbij dan nog steeds veel
kinderen onnodig testen.
182
Samenvatting in het Nederlands (Summery in Dutch)
Vijf factoren droegen significant bij aan het model: op 2 jaar een motorische
ontwikkelingindex (PDI score) < 90 op de BSID-II-MS, en van de BSID-II-BRS
de score voor de subschaal kwaliteit van de motoriek < 26 percentiel en de
risicofactoren: zwangerschapsduur, mannelijk geslacht en (intraventriculaire)
hersenbloeding. Toepassing van het model verbetert de efficiëntie van de
motorische onderzoeken in het follow-up programma met 37% en voorspelt met
een sensitiviteit van 94% en een specificiteit van 50%. Dit betekent dat we na
toepassing van het model 129 van de 345 kinderen niet zouden hebben op­geroepen voor een motorische test en 6 kinderen met een vertraagd motorische
niveau zouden hebben gemist.
In Deel II, Hoofdstuk 5 wordt de voorspellende waarde van twee motorische
testen op 6 (t 0), 12 (t1) en 24 (t 2) maanden voor het motorisch niveau op 5 jaar
(t3) vergeleken. Kinderen uit cohort 2, geboren vanaf maart 2003 tot en met
december 2005, met uitsluiting van de kinderen met bekende beperkingen
werden geïncludeerd. Op t 0, t1 en t 2 werd gebruik gemaakt van de BSID-II-MS,
BSID-II-BRS totale score en subschaal kwaliteit van bewegen (MQ) en op t 0, en
t1 van de AIMS en op t3 van de MABC-2-NL. De uitkomsten werden gedichotomiseerd in normaal versus een vertraagde/abnormale motorische ontwikkeling.
Onderzocht werden de voorspellende waarde per test met verschillende
afkapwaarden per test en combinaties van testen.
Op 5-jarige leeftijd was de MABC-2-NL (n=201), totale standaardscore
gemiddeld een ½ standaarddeviatie onder dat van de normpopulatie. De
MABC-2-NL categoriseerde 74 (37%) kinderen als vertraagd/abnormaal. De
AIMS ≤10 percentiel gaf de beste combinatie van sensitiviteit en specificiteit om
de motorische uitkomst te voorspellen, met een sensitiviteit op t1 van 65%, en
specificiteit van 68%. De BSID-II-MS ≤84 gaf de beste sensitiviteit (t 0=79%,
t1=93%, t 2=63%) en de BSID-II-BRS ≤25 percentiel de grootste specificiteit
(t 0=87%, t1=90%, t 2=81%). Op t 0 had de combinatie van de AIMS ≤10/BSID-MS
≤69/BRS ≤25, op t1 de combinatie AIMS ≤10/BSID-MS ≤69/MQ ≤25 en op t 2 de
combinatie BSID-MS ≤69/MQ ≤25 de beste balans tussen sensitiviteit en
specificiteit (t 0=78%, 43%; t1=78%, 62%; t 2=74.0%, 62%). Door een combinatie
van motorische niveau, gedrag tijdens de test of de kwaliteit van bewegen
konden we de voorspellende waarde voor motorische problemen op 5 jaar
verbeteren. De follow-up op 6 maanden was minder voorspellend.
183
Samenvatting in het Nederlands (Summery in Dutch)
Deel III, Hoofdstuk 6 beschrijft de ontwikkeling van een meetinstrument voor
het beoordelen van de kwaliteit van bewegen. De kwaliteit van bewegen bij
vroeg prematuur geboren kinderen is anders dan bij op tijd geboren kinderen.
Kinderfysiotherapeuten beschrijven deze kwaliteit van het bewegen omdat een
compacte test ontbreekt. Het doel van deze studie was een meetinstrument te
ontwikkelen voor de observeerbare kwaliteit van bewegen vanuit het perspectief
van de kinderfysiotherapeut. Voor het ontwikkelen van dit meetinstrument
gebruikten we kwalitatief onderzoek. In de eerste fase werden 7 semigestructureerde interviews bij ervaren kinderfysiotherapeuten afgenomen. De tweede
fase bevatte een gestructureerde vergadering volgens een Nominale Groeps­
Techniek (NGT) bij de geïnterviewden om de meest relevante observeerbare
aspecten van kwaliteit van bewegen te identificeren. De derde fase bestond uit
een onderzoek met de Delphi techniek met 61 expert kinder­fysiotherapeuten
met als doel 80% overeenstemming te bereiken over relevantie, terminologie en
definities van de aspecten.
In totaal werden 32 aspecten gebaseerd op verschillende theoretische
constructen overwogen. Vijftien aspecten werden in het meetinstrument, de
Observable Movement Quality Scale (OMQ scale), opgenomen. Voor de definitie
van 14 aspecten kon 80% overeenstemming worden bereikt, voor de definitie
van adequate fijne motoriek 75%. Dit laatste aspect was complementair aan de
grove motoriek en werd daarom wel opgenomen in de schaal. De aspecten
kunnen worden gescoord op een 5-punts Likert schaal en de score loopt van 15
tot 75, waarbij een hogere score een betere kwaliteit van de motoriek aangeeft.
De inhoudsvaliditeit van deze schaal is geborgd door de manier waarop deze
schaal werd ontwikkeld. Voor klinische toepassing van de OMQ scale dient
deze nog te worden onderzocht op betrouwbaarheid, construct validiteit en
responsiviteit.
Deel IV, Hoofdstuk 7 bevat de samenvatting van dit proefschrift in het Engels.
In Deel IV, hoofdstuk 8 worden de resultaten van dit proefschrift bediscussieerd.
Daarnaast worden aanbevelingen gedaan voor beleid en toekomstig onderzoek.
Inzicht in het longitudinale beloop van het motorische niveau, dus het volgen in
de tijd, is belangrijk voor de indicatiestelling van kinderfysiotherapeutische
behandeling. Zowel te weinig als teveel verwijzingen moeten worden voorkomen.
184
Samenvatting in het Nederlands (Summery in Dutch)
Maar nog belangrijker is om in de toekomst het effect van de kinderfysiotherapeutische behandeling te kunnen evalueren.
Bij vroeg prematuur geboren kinderen is de aanwezigheid van risicofactoren
voor een vertraagde motorische ontwikkeling een indicatie voor het opstarten
van kinderfysiotherapeutische begeleiding, ook als de motoriek dan nog
normaal is.
Bij evaluatie van motorische interventies in de eerste twee levensjaren in gerandomi­
seerd onderzoek dient waar mogelijk gestratificeerd te worden voor de risico­
factoren chronische longziekten, geslacht, neonatale convulsies, kwaliteit van
bewegen en (meest belangrijk) het opleidingsniveau van de moeder.
De vraag blijft waarom het geslacht een grote invloed heeft op het motorische
niveau. Een aannemelijke verklaring is dat de normwaarden voor testen niet
apart zijn bepaald voor jongens en meisjes. Het is mogelijk dat het geslacht ook
een risicofactor is voor op tijd geboren kinderen. Dit wordt ondersteund door
het feit dat de standaarddeviatie van de testscores bij vroeg prematuur geboren
kinderen niet groter dan bij op tijd geboren kinderen.
Kinderfysiotherapeuten gebruiken discriminatieve testen, om te bepalen of
kinderen vertraagd zijn in hun motorische ontwikkeling ten opzichtte van
“normale” kinderen. De testen worden ook gebruikt om het beloop van de
ontwikkeling te volgen. Deze testen blijken echter niet gevoelig genoeg om
kleine verschillen in de motoriek te meten. Daarnaast wil men de effecten van
kinderfysiotherapeutische interventies kunnen vaststellen, waarvoor evaluatieve
testen noodzakelijk zijn. Er zijn ziekte-specifieke evaluatieve testen ontwikkeld
maar deze zijn niet bruikbaar voor kinderen met minder grote motorische
problemen. Er moeten dus motorische testen worden ontwikkeld die discriminatieve en evaluatieve eigenschappen combineren met kleine stapjes tussen de
opeenvolgende de items van de test.
In toekomstige follow-up studies moeten meer omgevingsfactoren worden
meegenomen, zoals bijvoorbeeld de verwachtingen van de ouders, de toepaste
interventies en de culturele achtergrond.
Het combineren van uitkomsten van motorische testen, het gedrag van het kind
tijdens de test en de kwaliteit van de motoriek laten zien dat de voorspelling kan
worden verbeterd. Het is dus van groot belang dat kinderfysiotherapeuten deze
informatie blijven integreren in hun klinisch denken.
185
Samenvatting in het Nederlands (Summery in Dutch)
De studies hebben een aantal beperkingen.
1. Ze zijn niet gestandaardiseerd voor de bij de individuele kinderen toegepaste
interventies.
2.De studies zijn uitgevoerd in één (universitair) ziekenhuis, wat mogelijk
invloed heeft op de generaliseerbaarheid van de resultaten.
3.In de groep die niet voor follow-up verscheen is het percentage kinderen met
motorische problemen hoger, wat dus een onderschatting is van de motorische
problemen bij de vroeg prematuur geboren kinderen.
4.De cognitieve ontwikkeling van de kinderen niet is meegenomen in deze studies
en het gedrag van de kind werd alleen geobserveerd tijdens de test.
Aanbevelingen voor onderzoek
1.Het ontwikkelen van multidisciplinair voorspellende modellen omdat deze
modellen de follow-up efficiënter kunnen maken.
2.De Ages and Stages Questionnaire, een vragenlijst voor ouders om de
ontwikkeling van het kind te evalueren, zou moeten worden onderzocht
op zijn bruikbaarheid in het follow-up programma voor vroeg prematuur
geboren kinderen.
3.De rol van kwaliteit van bewegen in de voorspelling van het motorische
niveau van het kind zou verder moeten worden onderzocht.
4.Interventie studies die het effect van de kinderfysiotherapeutische behandeling
evalueren kunnen pas worden opgestart als motorische testen zijn ontwikkeld,
die gevoeliger zijn voor verandering.
5.De longitudinale variatie van zich “normaal” ontwikkelende kinderen en de
verschillen tussen jongens en meisjes zouden moeten worden onderzocht.
186
Dankwoord (Acknowledgements)
Dankwoord
Aan het einde van dit proefschrift wil ik graag iedereen hartelijk bedanken die
een bijdrage, op welke wijze dan ook, heeft geleverd aan mijn promotie. Hieronder
wil ik graag een aantal mensen in het bijzonder noemen, waarbij ik uiteraard
hoop niemand te vergeten.
Allereerst wil ik alle kinderen en ouders, die voor alle follow-up evaluaties kwamen
bedanken. Het gebeurt niet zo vaak dat een kinderfysiotherapeut, langdurig zoveel
kinderen kan volgen. Hierdoor was feedback mogelijk op de uitkomsten van de
motorische observaties op 6, 12 en 24 maanden. Klopte de conclusie van de
vorige testen en was dat op 5 jaar nog relevant? Ik heb er heel, hééél veel van
geleerd.
Mijn bijzondere gaat dank uit naar mijn promotoren prof. dr. Ria Nijhuis-van der
Sanden, prof. dr. Louis Kollée en prof. dr. Rob Oostendorp. Beste Ria, na de
beslissing dat jij van ons tweeën als eerste zou promoveren zijn we gezamenlijk
op weg gegaan. Het was een leertraject voor ons beiden om mijn promotie te
realiseren. Je hebt je samen met Louis Kollée sterk gemaakt om de follow-up
poli te realiseren waardoor dit onderzoek mogelijk was. Altijd was je bereid te
overleggen en razendsnel met feedback op alle stukken. Aan het vergaderen bij
jou thuis en de goede zorgen heb ik goede herinneringen. Beste Rob, het waren
mijn eerste twee artikelen over spierkrachtmetingen, na de cursus “scholing in
wetenschap 3”, die mij op het onderzoekspad brachten. Dank dat je daar toen
zoveel tijd in hebt gestoken. Beste Louis de interpretaties van de medische data
en de correcties van het Engels, ze waren onmisbaar. Ik wil jullie bedanken voor
de goede begeleiding en prettige samenwerking in de afgelopen jaren. Altijd
waren jullie bereid tot ’s avonds laat te vergaderen. Ik heb ontzettend veel van
jullie geleerd.
Mijn onmisbare ondersteuning voor de statische analyse, Ir. Reinier Akkermans.
Beste Reinier, je hebt mij op weg geholpen en de ingewikkelde analyses voor
mij gedaan, dank voor je geduld.
187
Dankwoord (Acknowledgements)
Mijn collega’s van de afdeling kinderfysiotherapie van het UMC St Radboud:
Marlou Essink, Anke Langelaan, Gera Peters, Perijn Verheij, dr. Leo van Vlimmeren,
Ineke Wouters en de ex-collega’s Jasper Köiter en drs. Joke Davio-Tissingh.
Vele kinderen zijn door jullie getest en vele klinische taken overgenomen.
Daardoor was het mogelijk tijd vrij te maken om dit proefschrift te realiseren. Ik
ben jullie veel dank verschuldigd en hoop dat bij een volgende promotie van de
afdeling, ik de patiëntenzorg kan doen. Mede promovendi, drs. Linda Reus en
drs. Marjolein Jongbloed veel succes met de afronding van jullie proefschriften.
Dank aan de vele tijdelijke medewerkers van de afgelopen jaren die betrokken
waren bij de follow-up prematuren, Barbera Kölzer, Saskia Haring, Vivian van
Mierlo, Jennifer Denisson, drs. Isabelle Bonthond-Durein, Hans Bijzet en Ellie
Verheij. Wat zijn er door jullie veel te vroeg geboren kinderen getest.
Eline Diekema en de 61 kinderfysiotherapeuten die veel tijd hebben gestoken
in de Delphi studie, komen veel dank toe. Beste Eline, je hebt een geweldige
prestatie geleverd om het onderzoek rond te krijgen en daar veel tijd in gestoken.
Daarom zie ik voor jou ook een mooie toekomst in de kinderfysiotherapie.
De secretariaten van de afdeling neonatologie en kinderfysiotherapie wil ik ook
bedanken. Het secretariaat neonatologie Corrie van Wolferen, Frederique
Smeets, Heidi verbeet en Marjolein Visser die verantwoordelijk waren voor de
planning. Altijd hielden jullie strak de inclusie criteria en de leeftijd waarop de
kinderen werden opgeroepen in de gaten. Door jullie stiptheid en inzet hebben
zoveel kinderen en ouders mee gedaan in de follow-up. Het secretariaat van de
afdeling kinderfysiotherapie, Trees Plamont al weer een tijdje met pensioen,
Christine Teunissen, Daniëlla Giesbers-Vrins en Marieke Peters. Het was en is
een hele klus al die statussen,de juiste testen, het klaarzetten en het controleren
van de brieven.
Ann Jenks met je vele jaren ervaring van copy-editing, dank voor het corrigeren
van de artikelen. Je zuchtte af en toe door mijn matige kennis van het Engels.
Zonder jouw kritische correcties waren de artikelen niet leesbaar geworden. Het
was fijn dat ik op donderdagmorgen, tijdens onze gezamenlijke fietstocht naar
de trimhockey altijd kon vragen wat de juiste Engelse vertaling was.
188
Dankwoord (Acknowledgements)
Als laatste de belangrijkste personen in mijn leven, mijn ouders, Reinier, Saskia
en Léon. Lieve Reinier, zonder jouw stimulans om te blijven werken en door te
gaan was het misschien nooit zover gekomen. Gelukkig heb jij veel ervaring in
het schrijven van artikelen, jouw kennis en de goede Engelse schrijfvaardig­
heden, ze waren onmisbaar. Lieve Saskia en Léon, jarenlang zat ik op zaterdag
en zondagmorgen achter de computer, om vervolgens nog naar het hockeyveld
te snellen om toch de wedstrijden te kunnen zien. En ma, vaak verwachtte je dat
ik zou komen, maar dan lukte het toch net weer niet. Het huis is de laatste
maanden wat stoffig geworden, maar dat wordt in de komende tijd wel weer
beter. Dank voor jullie begrip en vertrouwen. Het is jammer dat papa dit niet
meer kan meemaken, hij zou zeker trots zijn geweest.
Anjo
189
Curriculum Vitae
Curriculum vitae
Anjo Janssen was born in Nijmegen, the Netherlands, on March 31th 1960. She
graduated in 1976 from the lower and in 1978 from the higher secondary school
at the “Scholengemeenschap Nijmegen Oost”. In 1984 she completed Physical
Therapy at the HAN university of Applied Sciences in Nijmegen. She started
working as physical therapist at the Radboud University Medical Centre at the
Department of Physical Therapy in 1984 (Head: G Worm) and from 1986 she has
worked at the Pediatric Physical Therapy unit (Head: G Worm/ Prof. dr. MWG
Nijhuis-van der Sanden/M Essink). From 1991 to 1994 she studied Pediatric
Physical Therapy at the HU university of Applied Sciences in Utrecht, followed
by scientific education at the Dutch Allied Health Sciences Institute from 1995 to
1996.
The Master of Science degree was acquired in Physical Therapy Science at the
Utrecht University from 2004 to 2006 (Head: Dr. N van Meeteren). Her PHD
trajectory started in 2004 on motor performance in preterm infants.
Anjo Janssen is married in 1991 with Reinier Raymakers. Together the have two
children Saskia (1991) and Léon (1994).
190
List of publications
List of publications
1. Janssen AJ, Elvers JW, Oostendorp RA. Hand held dynamometrie bij kinderen met neuromusculaire aandoeningen (NMA): een literatuurstudie. Ned T Fysiotherapie 2000; 110: 137-143
2. Janssen AJ, Elvers JW, Oostendorp RA. Hand-held dynamometrie: intrabeoordelaarsbetrouwbaarheid bij kinderen met neuromusculaire aandoeningen. Ned T Fysiotherapie 2001; 111: 159-165
3. Van den Beld WA, van der Sanden GA, Feuth T, Janssen AJ, Sengers RC, Verbeek AL, Gabreels
FJ. New motor performance test in a prospective study on children with suspected myopathy. Dev
Med Child Neurol. 2006; 48: 739-43.
4. Nijhuis-van der Sanden MW, Kotte EM, Janssen AJ, Langelaan AL, Oostendorp RA. Bepaling van
het motorisch niveau van extreem vroeggeboren kinderen met de Nederlandstalige BSID-II-NL.
Ned T Fysiotherapie 2007; 117: 42-48.
5. Janssen AJ, Nijhuis-van der Sanden MW, Akkermans RP, Oostendorp RA, Kollée LA. Influence of
behaviour and risk factors on motor performance in preterm infants at age 2 to 3 years. Dev Med
Child Neurol. 2008; 50: 926-931.
6. Nijhuis-van der Sanden MW, van der Cammen-van Zijp MH, Janssen AJ, Reuser JJ, Mazer P, van
Heijst AF, Gischler SJ, Tibboel D, Kollée LA. Motor performance in five-year-old extracorporeal
membrane oxygenation survivors: a population-based study. Crit Care. 2009; 13: R47.
7. Janssen AJ, Nijhuis-van der Sanden MW, Akkermans RP, Tissingh J, Oostendorp RA, Kollée LA. A
model to predict motor performance in preterm infants at 5 years. Early Hum Dev. 2009; 85:
599-604.
8. Van Dijk M, Groen W, Moors S, Bekkering P, Hegeman A, Janssen A, Takken T, van der Net J,
Helders P. The Dutch translation of the revised Childhood Health Assessment Questionnaire: a
preliminary study of score distribution. Clin Exp Rheumatol. 2010; 28: 275-80.
9. Woldringh GH, Horvers M, Janssen AJ, Reuser JJ, de Groot SA, Steiner K, D’Hauwers KW, Wetzels
AM, Kremer JA. Follow-up of children born after ICSI with epididymal spermatozoa. Hum Reprod.
2011; 26: 1759-67.
10. Janssen AJ, Akkermans RP , Steiner K, Haes O de, Oostendorp RA, Kollée LA, Nijhuis-van der
Sanden MW. Unstable longitudinal motor performance in preterm infants from 6 to 24 months on
the Bayley Scales of Infant Development—Second edition. Res Dev Disabil. 2011; 32: 1902–1909.
11. Van den Beld WA, Van der Sanden GA, Janssen AJ, Sengers RC, Verbeek AL, Gabreëls FJ.
Comparison of 3 instruments to measure muscle strength in children: A prospective study. Eur J
Paediatr Neurol. 2011; 15: 512-518.
12. Jongbloed-Pereboom M, Janssen AJ, Steenbergen B, Nijhuis-van der Sanden MW. Motor Learning
and working memory in children born preterm: A systematic review. Submitted
13. Janssen AJ, Diekema ET, van Dolder R, Kollée LA, Oostendorp RA, Nijhuis-van der Sanden MW.
Development of a movement quality measurement tool for children. Physical therapy, April 2012.
191