Pediatric Neurology 48 (2013) 424e431

Contents lists available at ScienceDirect

Pediatric Neurology
journal homepage: www.elsevier.com/locate/pnu

Review Article

Rehabilitation of Children With Traumatic Brain Injury: A Critical Review

Galit Tal MD, Emanuel Tirosh MD *
The Hannah Khoushy Child Development Center, Bnai Zion Medical Center, Haifa, Israel

article information

abstract

Article history:
Received 27 February 2012
Accepted 26 November 2012

The purpose of this critical review of the English literature published between 1975 and
2009 was to assess the quality of the evidence for the efcacy of rehabilitation intervention
after traumatic brain injury in children. Evidence for intervention studies was used to
classify the research strength of design and report. Only a minority (16/439) of the published
studies has been related to traumatic brain injury in children and has used a scientic or
quasiscientic design. Only one study met the criteria of class I evidence for intervention
studies. However, this study included adults, as well as children. The other 15 studies,
although reporting positive results, had many methodologic deciencies, and consequently
their validity is questionable. Although the methodologic and ethical difculties involved are
acknowleged, a multicenter approach is required to achieve valid conclusions. Use of designs
such as comparative effectiveness research might prove to be a practical solution. Highquality intervention research would facilitate stronger evidence-based counseling for children and families requiring posttraumatic brain injury intervention and to policy makers.
2013 Elsevier Inc. All rights reserved.

Introduction

Acquired brain injury (ABI) is a general term referring to

events such as traumatic brain injury (TBI), central nervous
system infection, stroke, hypoxia, tumor, and radiation
treatments, with marked variability in outcome. Of the
causes of ABI, traumatic brain injury (TBI) has been reported
as the major cause of morbidity and death [1,2]. In the
United States, the Centers for Disease Control and Prevention has estimated that 250 of every 100,000 people will
sustain a brain injury every year [3]. Severe TBI has been
strongly associated with persistent moderate or severe
disability [4,5]. Approximately 25% of the children and
youth with acquired brain injury will require rehabilitation
services [3]. It is assumed that brain plasticity provides the
foundation for recovery that can be enhanced by rehabilitative intervention. The less optimal recovery after severe
early brain injury as compared with later sustained injury
has been recently acknowledged [5,6]. Despite the large
* Communications should be addressed to: Dr. E. Tirosh; The Hannah
Khoushy Child Development Center; Bnai Zion Medical Center; POB
4940; Haifa, Israel.
E-mail address: tirosh-e@b-zion.org.il
0887-8994/$ - see front matter 2013 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.pediatrneurol.2012.11.008

number of patients and massive resources invested in

rehabilitation programs after TBI, very little is known about
the efcacy (and more so effectiveness) of such interventions [7-11]. In view of the natural course after TBI there are
ethical and methodologic difculties in the evaluation of
the treatment impact [11]. The randomized controlled trial
design is widely accepted as providing the best evidence in
scientic terms. In rehabilitation research, however, it is
controversial because of ethical and practical problems,
such as deprivation of possible valuable intervention, the
small number of cases in any one center at a certain point of
time, the diversity of patient disease, the diversity of rehabilitation goals and the difculty in applying a uniform
therapeutic approach [12]. Three review articles on cognitive rehabilitation after ABI in children found a critical lack
of randomized controlled studies addressing treatment
efcacy [8-10]. The purpose of this article is to review the
literature for evidence-based studies related to the effectiveness of rehabilitation in children who have sustained
TBI between 1975 and 2009. Studies pertaining to functional physical and cognitive outcomes were the focus of
the review. Unlike previous reviews focusing specically on
cognitive outcomes [8-10] or ABI [6], we have addressed all
possible functional outcomes.

G. Tal, E. Tirosh / Pediatric Neurology 48 (2013) 424e431

Table 1. Denition of classication of evidence for intervention studies

Denition for Classication

Class I

Class II

Class III

Class IV

Prospective, randomized controlled clinical trials with

blind assessment of outcome in a representative
population. The primary outcome, inclusion and exclusion
criteria clearly dened. Drop outs and crossover to the
other group are minimal. Baseline characteristics should
be similar between groups.
Prospective matched group cohort study in
a representative population with blind assessment of
outcome or a randomized controlled trial with one
signicant ow (e.g., primary outcome not clearly
dened).
All other controlled trials (including well dened natural
history controls or patients serving as own controls) in
a representative population where outcome is blindly
assessed or independently derived by objective outcome
measurement.
Evidence from uncontrolled studies, case series, case
reports, or expert opinion.

Materials and Methods

A computerized search of the literature published in English, during
the years 1975 to 2009, by use of the Pubmed, Cochrane, and MD Consult
databases, was performed. We used the following entries in different
combinations: rehabilitation, traumatic brain injury, children or pediatrics, intervention, outcome, effectiveness or efcacy, and review. The
inclusion criteria were as follow: (1) Children under the age of 18 years
were included in the report; (2) the study pertained to children who
sustained TBI; (3) the intervention took place in either an inpatient or
community framework; (4) the follow-up period was at least 3 months;
and (5) a dened outcome of interest was assessed.
Two pediatricians in the eld of child development and neurology
did the preliminary search independently. The literature search yielded
460 articles, of which 456 had abstracts. All abstracts were studied
independently by a pediatrician involved with child development and
rehabilitation and a certied neurodevelopmental pediatrician. Of the
456 abstracts read, 439 were found to be irrelevant as agreed to by the
two reviewers (adult studies, neonatal studies, and technique descriptions with no outcome assessment, studies addressing emotional/
behavioral aspects or family function only and case reports). Singlesubject research design suggested for the assessment of rehabilitation
outcome [13] has not been included in our review, because only a large
number of replicated studies and properly calculated effect size allow
their results to be generalized. Only 16 studies were found to be relevant
to this review, and their full text was critically reviewed. After a discussion between the reviewers, a consensus was reached with regard to
class assignment. The studies were dened and classied as evidence
for intervention studies on the basis of the system by Edlund et al. [14]
as class I (strongest) -IV (weakest) (Table 1). This classication system
has previously been used in other reviews pertaining to rehabilitation
[9,10]. Thus a comprehensive conclusion stemming from these reviews
can be generated.

Results

Description and classication of the 16 research articles

[15-30] are summarized in Table 2.The reviews are listed
from the weakest (IV) to the strongest (I) evidence and in
chronological order. The following are the included research
reports in order from class IV to I evidence-class and their
main methodologic problems.
The reports assigned class IV evidence for intervention:
the seemingly retrospective descriptive research by Berger
et al. [15] included 38 children, with 44.7% resulting from

425

TBI, with no differential analysis of this subgroup. The

intervention was used during the inpatient period only but
included a follow-up period of 6 months, while being back
in the community. The intervention is reported in general
terms and therefore cannot be replicated from the basis of
the report, and the outcome measures other than the level
of vigilance were descriptive. No control or contrast group
was included, and the results were not statistically
analyzed.
One research report [16] pertained specically to pragmatic skills in six adolescents, of whom ve sustained
a closed head injury. Although the procedure and instrument, as well as the intervention, are well described, and
a wash-out period of few months has been included, no
control or contrast group was used. Therefore the results
cannot be attributed specically to the intervention per se.
The retrospective, descriptive, uncontrolled study by
Pace et al. [17] included 77 participants with ABI, who were
enrolled in a home-based rehabilitation program. Sixtyeight percent of the participants were younger than age
20 years, of whom 73% sustained TBI. Both predetermined
individualized clinical goals and family satisfaction
(assessed in 45 of the families) were used as outcomes of
interest. The intervention is described in general terms only.
However, the follow-up period lasted 12 months.
Another uncontrolled, retrospective descriptive study by
Dumas et al. [18] also included 79 children undergoing
inpatient intervention only and included children in their
acute state after TBI. This study dened the outcome of
interest and used a valid measure of assessment pertaining
to different functional skills.
The report of Chen et al. [19] pertained to a retrospective
analysis of an acute rehabilitation service, using a psychometrically valid and reliable instrument. The data were
from 910 children, of whom 336 suffered TBI. Twelve of 32
institutes agreed to participate. The interval between the
time of diagnosis and rehabilitation onset was signicantly
shorter in children with TBI. These two factors were an
important source of bias. A differential analysis of outcome
in patients with TBI as compared with other causes was
performed. The analysis of the outcomes of interest was of
correlative nature but related to the total amount of rehabilitation time rather than to the specic methods or
disciplines involved.
Conversely the second study by Dumas et al. [20] with an
extended cohort of inpatients, some of whom were also
included in the rst study, focused on the contribution of
physical therapy (PT) to the recovery of children older than
2 years of age who sustained TBI. The study was uncontrolled, and retrospective data extraction was obtained from
a daily-standardized recording system. The valid measurement used was the same as in the rst study. Because the
exact nature of PT administered to each patient could not be
specied, the authors resorted to an analysis of the effect of
PT intensity only.
In 2005, Abdullah et al. [21] reported a descriptive study
of 36 children who suffered mild to severe head injury.
Children with prolonged hypoxia, dilated pupils, or surgical
intervention were not included in the cohort. The follow-up
lasted 12 months. All patients received traditional treatment with no designated rehabilitation program because of
a lack of facilities. The outcome of interest was cognitive

426

G. Tal, E. Tirosh / Pediatric Neurology 48 (2013) 424e431

Table 2. Description and classication of the reviewed articles

Author Year
[No]

Classication Sample Characteristics

Evidence

Study Design

Berger et al.
1997 [15]

IV

Retrospective uncontrolled study.

Evaluates the course of rehabilitation
and the outcome 6 months after the
end of inpatient rehabilitation (GOS
and goals achievements),

Wiseman-Hakes IV
et al. 1998
[16]

Pace et al. 1999

[17]

IV

Dumas et al.
2002 [18]

IV

Chen et al. 2004 IV

[19]

Dumas et al.
2004 [20]

IV

Abdullah et al.
2005 [21]

IV

Dumas et al.
2008 [22]

IV

Tepas et al.
2009 [23]

IV

Melchers 1999
[24]

III

31 patients (average 7.8 years),

during inpatient rehabilitation
following ABI (44.7% TBI) TFI:
3.90 days

Evaluated Outcomes

21% were dened as achieving good

rehabilitation, although nearly 45% of
patients displayed severe impairment.
After 6 months further improvement was
found in 40% of patients, however only
slightly changing the overall GOS values
6 adolescents with pragmatic
Prospective uncontrolled study.
Signicant differences in the
communication decits
Evaluates a method of group training pretreatment-to-posttreatment period
secondary to ABI (83% TBI) TFI: for adolescents with pragmatic decits and were maintained at follow-up.
4 months-9 years.
secondary to ABI, using standard scores
before treatment, immediately after
treatment and6 months post treatment.
77 patients with ABI (73% TBI) Prospective uncontrolled study.
77% of individualized outcome goals
TFI: not reported
Individual goals achievement following were achieved at discharge, 80% at 6
home-base rehabilitation program.
months and 78% at 12 months
Assessment at discharge from the
program, and at 6 and 12 months
follow up.
79 children and adolescents
Retrospective descriptive study. The
At discharge signicant improvement
with TBI, mean age 9.9 years.
patients were assessed using PEDI
was noted within all three domains
TFI: 34-53 days.
functional skills and care given
(the greatest) change was in mobility.
assistance domains of self-care,
mobility and social function, on
admission and discharge
814 pediatric patients who
Retrospective cohort design. Collecting Most children improved after receiving
received inpatient
records of pediatric patients who
rehabilitation. Children > 7 years and
rehabilitation (41% after TBI)
received inpatient rehabilitation and
those with traumatic injury made larger
TFI: 1-1525
their functional assessment. Data were gains. The amount of total treatment,
extracted from patient billing or
combined or discipline-specic,
treatment records or both and the
signicantly contribute to gains in selfWeeFIM data base.
care, mobility, and cognition
The intensity of PT intervention in an
80 children and adolescents
Retrospective cohort study. Describe
inpatient rehabilitation hospital is
with TBI, admitted to inpatient the types and intensity of PT
related to positive changes in
rehabilitation. TFI: 6-78 days
intervention used during inpatient
functional mobility scores and the
rehabilitation. To examine the
achievement of a minimal clinical
relationship between the intensity of
important difference.
PT procedurals intervention and
mobility changes throughout the
intervention using standardized scales
(data were collected from the hospital
medical records).
36 patients with mild to severe Prospective uncontrolled study.
No general improvement when
head trauma (92% with severe Assessment of patients at 3, 6, and 12 compared to the western literature
TBI) TFI: not applicable.
months after the occurring event for
neuropsychological functions.
452 pediatric patients who
Prospective uncontrolled study.
More than 55% of the children achieved
received inpatient
Evaluating the proportion of children MID. Highest proportion in mobility.
rehabilitation (65% following
who achieved minimal important
Greater chance of achieving MID: older
ABI) TFI: not reported
difference during inpatient
age >10 years, longer length of stay,
rehabilitation, using change scores
lower admission PEDI score and
(Average length of stay: 46.3 days).
a diagnosis of ABI.
60 children, mean age
Retrospective, uncontrolled study.
Delay was signicantly correlated with
11.2 years, with TBI and initial Correlation between the delay (days
both outcomes and rehabilitation
GCS score of 8 or lower. TFI:
between release from ICU) and
efciency.
0-24 days
initiation of inpatient rehabilitation),
the outcome and the rehabilitation
efciency (assessed at the end of the
inpatient rehabilitation).
45 children and adolescents
Controlled, prospective randomized
Higher score of intelligence in the
with severe TBI. TFI: 48 hours. study. Assessing the effectiveness of
experimental group. Over the time of
early stimulation in coma (while the
one year the development of nonverbal
patient is on ICU) and
learning potential is better in the
neuropsychological therapy after
experimental group. Less
regaining consciousness, compared
psychopathological alternations in the
with routine treatment. Follow-up: 6, experimental group and better quality
12, and 24 months after trauma.
of life.
(continued on next page)

G. Tal, E. Tirosh / Pediatric Neurology 48 (2013) 424e431

427

Table 2 (continued )
Author Year
[No]

Classication Sample Characteristics

Evidence

Study Design

Swaine et al.
2000[25]

III

Retrospective controlled trial.

Assessing the rehabilitation outcome at
least 2 years after the TBI using
a telephone interview with the parents
after a new, more coordinated,
comprehensive rehabilitation program
vs the old rehabilitation program in the
same institute.

Braga et al.
2005 [26]

II

vant Hooft et al. II

2005 [27]*

vant Hoof et al.

2007 [28]

II

Katz-Leurer
et al. 2009
[29]

II

Zhu et al.
2001 [30]

64 children with head injury

(mild moderate and severe)
TFI: 4 weeks

Evaluated Outcomes

No statistical signicant differences

were found between the two groups in
respect to mean WeeFIM/FIM scores
and mean PARSIII scores (a scale to
assess psychosocial adjustment of
children with chronic physical
illness).The historic group (1993) had
poorer scores on the withdrawal
subscale (in the PARSIII.
87 children with moderate or Randomized controlled trial. The
Both groups beneted from their
severe TBI during the chronic
outcome after 12 months of familyintervention, but only the familypost injury stages TFI: 6-30
supported intervention for children
supported group results achieved
months
with chronic impairment after TBI was statistical signicance.
compared with the outcome after
months of clinician-delivered
intervention. Outcomes were assessed
with formal scales.
38 children with ABI, 9-16
Randomized control study. Testing the Signicant improvements in memory
years of age (55% following
effectiveness of a cognitive training
and attention performance were shown
TBI). TFI: 1-5 years
program, 6 months after the end of the in the treatment group as compared
program.
with control subjects.
38 children with ABI (55% TBI) Randomized control study. The
The treatment group exhibited more
TFI: 1-5 years
treatment group was trained with
persistent improvement with respect
a broad-based cognitive training
to complex tasks of attention and
program. The control group had a daily memory. No difference on simple
activity freely chosen by the parent,
reaction time test.
child or teacher. The two groups have
been assessed prior to the intervention
and 6 months after the training was
completed.
There was a positive signicant change
20 children aged 7-13 years:
Randomized controlled study. To
10 after TBI and10 with CP
evaluate the feasibility and efcacy of in the balance performance in the
experimental group. In all other
TFI:
12 months
a home-based exercise program
outcomes, no signicant differences
(experimental group). The control
were found.
group had regular daily activities.
Measurement tools-valid functional
tests and assessing isometric strength
at the end of the intervention (6 weeks)
and at 6 months follow-up.
36 children and adults after
A randomized controlled assessorAt 2 and 3 months there was a trend of
moderate and severe TBI Mean blind study Assessing the outcome in more patients in the study group
age 33 (12-56/57) years.
the early posttraumatic stage (6 ms
achieving full FIM scores, but the
TFI: 4-99 days
post injury) following intensive
control group appeared to be catching
rehabilitation (4 hr/day, physiotherapy, up toward 6 months.
occupational therapy and speech
therapy) vs 2 hr/day (conventional
group) assessments after 2, 3 and
6 months.

Abbreviations:
ABI
Acquired brain injury
CP
Cerebral palsy
FIM
Functional independence measure
GCS
Glasgow Coma Scale
GOS
Glasgow Outcome Scale
ICU
Intensive care unit
MID
Minimal important difference
PEDI
Pediatric Evaluation of Disability Inventory
PARS
Personal Adjustment and Role Skills Scale
PT
Physical therapy
TBI
Traumatic brain injury
TFI
Time following injury
WeeFIM Functional Independence Measure For Children

428

G. Tal, E. Tirosh / Pediatric Neurology 48 (2013) 424e431

functions, and the measurements were commonly used

intelligence scales. In the discussion, the authors compare
their results, indicating no improvement of cognitive
functions over time, to other studies with interventions and
point to the probable value of rehabilitation to the cognitive
outcome after TBI.
The third study by Dumas et al. [22] pertained to children
with ABI, including TBI, and examined the proportion of
children who achieved minimally important difference
derived from the same measure used in their other two
studies. The analyzed data were retrieved retrospectively
from standardized records. The details of the inpatient
intervention were not reported, and diagnostic groups, as
well as severity of the injury, were not accounted for in the
analysis.
The retrospective chart review reported by Tepas et al.
[23] addressed the possible importance of time since
injury for the start of rehabilitation on outcome. This
method might be regarded as a proxy to the direct
measurement of the effectiveness of intervention. Sixty
children with TBI were assessed with a valid instrument in
correlating the functional outcome of interest with the time
delay before entry to the program. However, the analysis
did not account for the initial functional level of the
participants, but their level of severity of injury at the very
acute stage. Furthermore the bias introduced by the events
possibly contributing to the delay in enrollment to the
program as well as that related to other interventions
possibly occurring prior to the start of the intervention,
were not controlled for.
All of these studies lacked in reporting level of intrarater
and interrater reliabilities and attributed positive effects to
rehabilitation intervention.
Two studies were assigned class III evidence [24,25].
One study [24], however, used a randomized controlled
trial design and reported a 12-month follow-up in less
than 50% of the participants. The neuropsychological
intervention has been described in detail but reported an
unequal distribution of severity of TBI between groups and
no blind assessment of the outcome. Although noted by
the authors, the nal analysis after full recruitment of the
patients and completion of the study has not been reported in the literature sources searched for the purpose of
this review.
The other study by Swaine et al. [25] included 26 children, constituting 30% of the eligible sample who sustained
TBI. A historical contrast group of 38 nonmatched children
exposed to a less-comprehensive service 2 years earlier was
used. The intervention is described in general terms only.
An acceptable valid instrument, administered by telephone
to the parents by blind interviewers, with no evidence for
interrater reliability, was used. A recall bias and no control
for events since the end of the rehabilitation program up to
the time of the interview at 2 and 4 years after the injury
were not accounted for.
Four reports were assigned as class II evidence. The
study by Braga et al. [26] investigated the effectiveness of
trained parents (study group) as compared with professionals - delivered rehabilitation program on motor and
cognitive outcome in 72 of 87 children after TBI. Participants were randomly assigned to study and comparison
groups; however, the same therapists administered both

types of interventions with a predetermined hypothesis

that the study group would derive more benet. Assessments were performed with blind researchers, using
instruments, one of which had no established psychometric properties at the time of the study. This study used
both clinical and statistical analysis. Although referring the
reader to data pertaining to the distribution of time
elapsed from the injury until the time of the intervention
(ranging 6-30 months), no such information was reported.
This potential bias is highly important because the analysis
demonstrated the signicant effect of this interval on the
outcome.
The rst of two reports by vant Hooft et al. [27] included
18 (12 with TBI) and 20 (9 with TBI) children randomly
assigned to the study and comparison groups, respectively,
with the rest of the children sustaining ABI of different
causes, including brain tumors. The studied cognitive
training was administered by a coach who was trained
weekly by a psychologist, whereas the parents of the
comparison group were instructed to play interactively
with their children with no exposure to computers or
television games. They reported on their activities using
weekly mailed diaries and telephone contacts. Standardized
cognitive measures were used to assess outcome before and
at the end of the 17-week program. There is no indication
with regard to the assessors and their reliability, as well as
their blindness to group assignment.
The second report by vant Hooft et al. [28] pertained to
the same study. However, results were evaluated 6 months
after completion of the training. Two drawbacks to this
study are the lack of control over the possible continued
exposure of the children in the study group or control
subjects during the time from intervention (contamination)
and again the lack of information related to the assessors
and their independence.
The report by Katz-Leurer et al. [29] included children
with cerebral palsy (n 10) and children after TBI
(n 10) who were randomly assigned to the study and
contrast groups. The intervention consisting of home
based exercises, as well as the outcome measures, have
been reported in detail. A differential analysis of the
outcomes related to TBI has not been reported, and
neither the assessors nor participants were blind to their
group assignment.
Class I evidence was only assigned to the study by Zhu
et al. [30]. This study evaluated the effectiveness of a 2(n 21) versus 4- (n 15) hour multidisciplinary inpatient
rehabilitation program. Assessments were carried out
independently by therapists not involved in the program
with standardized measures up to 6 months after the
injury. However, it should be noted that, although age
distribution (12-56 years) was equally distributed in the
two groups, no differential analysis related to age was
reported.
Discussion

It is generally accepted that rehabilitation is effective for

patients with TBI [31], although there have been a limited
number of research reports evaluating the effectiveness of
rehabilitation interventions, especially for children and
adolescents. Unfortunately, ethical and methodologic

G. Tal, E. Tirosh / Pediatric Neurology 48 (2013) 424e431

problems consistently confound rehabilitation research,

therefore rendering the RCT often impractical [10].
Some of the major points of concern are as follow: the
outcomes measured, preferably including measures of
quality of life, have often been unrelated to daily functions
and of limited clinical relevance. The use of control subjects
(i.e., no treatment) on the basis of the existing practices is
unacceptable whereas a comparison group (an alternative
mode of treatment) is highly difcult to design and requires
a large, well-matched sample size. Yet a few of the studies
overcame this problem and included a comparison group in
their design. Although some of the studies included in this
review pertained to a well-dened but not necessarily
representative group of patients, although being a more
practical and potentially valid approach, the inclusion of
participants with different underlying mechanisms and
therefore an expected different natural courses and prognosis renders their results less valid. The reliability and
validity of the measurements used were not consistently
ascertained, and the assessments were carried out by an
independent examiner. Finally the statistical analysis, when
performed, was not always related to the initial clinical
severity, and proper analysis accounting for repeated
measures, not found in this review, may in future studies
control for possible type I or II errors.
The integrative critical summary of the literature
addressing the efcacy of rehabilitation in children with TBI
raises concerns related to this list of weaknesses.
With regard to general design, only one of 16 research
reports included in this review used sufcient methodic
rigor required by class I [30], suggesting a signicant relationship between treatment intensity and outcome (dose
response).
However, after 6 months, this difference was not evident.
The cohort in this study, however, included children and
adults aged 12 to 56 years, with a mean age of 33 years.
Although four studies were assigned as evidence class II, the
major aws such as heterogenic groups and lack of independent assessment of outcome renders their results less
valid [26-29].
There were two reports with a large sample size of
814 and 452 patients, respectively [19,22]. Both studies
showed a signicant improvement after hospital rehabilitation, with larger gains for older children aged 7
years after TBI and 10 years of age after acquired brain
injury, respectively. Both articles suffer from a long list of
methodologic aws (well acknowledged by the authors),
including a variety of underlying causes for their brain
injuries. The separate analysis of outcome as related to
intervention in children who sustained TBI within the
group of children with ABI is not feasible; this last
limitation is true for most other studies that pertained to
ABI included in this review.
Ten of the articles described the severity of the traumatic
brain injury, mostly on the basis of the Glasgow coma scale.
Only four of them referred to the correlation between the
severity and outcome after the intervention. Only six
studies described the intervention characteristics in
sufcient detail to enable replication of the study. Four of
the studies did not accurately describe the time elapsed
between the end of the acute phase (discharge from

429

the intensive care unit) and the commencement of

rehabilitation.
Except for one article [25], all of the results supported
a clinically and some statistically signicant improvement
after a rehabilitation program and emphasized the importance of a variety of factors such as a short delay between
release from the intensive care unit and initiation of rehabilitation [15] and the importance of a coordinated
comprehensive rehabilitation program [17,25].
Three articles [16,27,28] pertained specically to cognitive rehabilitation after acquired brain injury. All of them
addressed specic cognitive rehabilitation programs or
techniques. All of them evaluated the cognitive improvement with standardized scales, and all showed signicant
improvement after cognitive rehabilitation. In these three
studies, as found in most of the reviewed articles, there are
signicant methodologic problems, such as absence of
details on the injury severity or a heterogeneous cause (e.g.,
brain tumor and TBI) or lack of independent assessment of
outcome [27,28].
Three articles [17,26,29] examined home rehabilitation
programs, carried out by family members and multidisciplinary professional teams. All three articles support the
idea of home-based rehabilitation and the involvement of
the family. However, different aws in design, for example,
a heterogeneous diagnostic group such as cerebral palsy
and TBI, as well as a lack of independent assessment of
outcome, weakens the validity of their results. Of these
studies the study reported by Braga et al. [26], although
using a randomized design and therefore of a relatively
robust method, included children up to 12 years only and
used two primary outcome measures, of which one
(SARAH scale of motor development) had no established
psychometric properties. Observer reliability and whether
preintervention or postintervention measurements were
performed by independent researchers, detailed information related to the interventions characteristics and
distribution of time interval between groups were not reported. Therefore this study, assigned as class I evidence in
a previous report [10], is qualied as class II in this review.
Interestingly one report [21] due to the lack of modern
rehabilitation resources in a community in Malaysia could
assess the spontaneous psychomotor improvement during
the rst year after TBI in children. The results of this report
showed the poor recovery when children and their families are deprived of rehabilitative services as compared
with results obtained after rehabilitation programs reported from other countries.
A review of the literature seeking evidence-based studies
on rehabilitation after TBI in children revealed three relevant
review articles [8-10]. All reviews referred to rehabilitation
after acquired brain injury and did not focus on TBI. The
conclusion in all of the reviews was that there was a critical
lack of evidence-based studies in the eld of childrens
rehabilitation. In a recent Cochrane review of treatment for
dysarthria after acquired brain injury in children and
adolescents, a critical lack of studies addressing treatment
efcacy was found [7]. Because the inclusion criteria in these
review articles were not necessarily identical to ours, their
reference list might include articles not reviewed by us and
vice versa. However, because both the classication used in

430

G. Tal, E. Tirosh / Pediatric Neurology 48 (2013) 424e431

three of these reviews and their nal conclusions are the

same, a general critical view can be concluded.
Although the number of studies related to the acute
phase treatment in TBI in children is rather extensive,
evidence-based studies concerning the rehabilitation phase
are lacking. This scarcity of research in the face of considerable invested effort and resources by the child himself,
the family and the service providers are surprising. To make
informed decisions that will lead to the best outcomes for
the children and their families, more knowledge about the
efcacy of rehabilitation is needed [2].
Although evidence from randomized clinical trials
remains central to the assessment of efcacy of interventions and for creating clinical guidelines, it appears that the
listed methodologic difculties often renders properly
designed randomized controlled trials impractical. Therefore model-based evaluations are a valuable resource for
health care decision makers [32], and possibly these
procedures would lead to objective exploratory studies of
the disease-modifying effects of rehabilitation. A feasible
approach in studying the effects of different interventions is
the method of comparative effectiveness research [33].
This approach uses analysis of groups of patients for an
association between specic medical interventions and
patient outcomes without randomization and usually
entails a collaborative multicenter study. A collaborative
multicenter approach with such a design would preferably
target a variety of outcomes of interest, measured routinely
before and after the intervention, including a period of
postintervention follow-up to assess the degree of sustainable long-term effects. Because it appears that the main
resources for such studies would be national agencies with
a health and economic interests, daily functional outcomes
and quality of life of both the child and parents seem
reasonable main outcomes of interest. Recently an attempt
to delineate an accepted list of functions for research and
clinical purposes has been reported [34]. The use of a singlesubject research design, if repeatedly implemented could be
valuable. However, it should be reserved for the unusual
case of a unique patient and or an innovative intervention.
Such a design, if properly used, including measurements
of effect size, could be a preliminary step before a more
rigorous design necessary for an acceptable evidence-based
intervention. The heterogeneity in patients characteristics,
intervention methods, and intensity of rehabilitation
should be accounted for in these methodologic approaches.

[4]
[5]

[6]
[7]

[8]
[9]

[10]

[11]

[12]

[13]

[14]
[15]

[16]

[17]

[18]

[19]

[20]

[21]

[22]
We are indebted to Prof. Michael Jaffe for reviewing the manuscript and to Rivka
Abiry for her assistance in the preparation of the manuscript.

[23]

[24]

References
[1] Langlois JA, Rutland-Brown W, Thomas KE. Traumatic brain injury
in the United States: emergency department visits, hospitalizations, and deaths. Atlanta: US Department of Health and Human
Services, CDC; 2006.
[2] Teplicky R, Law M, Rosenbaum P, Stewart D, DeMatteo C,
Rumney P. Effective rehabilitation for children and adolescents
with brain injury: evaluating and disseminating the Evidence. Arch
Phys Med Rehab 2005;86:924e31.
[3] Carney N, du Coudray H, Davis-OReilly C, et al. Rehabilitation for
traumatic brain injury in children and adolescents. Rockville:

[25]

[26]

[27]

Agency for Health Care Policy and Research. Evid Rep Technol
Assess 1999;2:1e5.
Boyer MG, Edwards P. Outcome 1 to 3 years after severe traumatic
brain injury in children and adolescents. Injury 1991;22:315e20.
Anderson V, Catroppa C, Morse S, Haritou F, Rosenfeld J. Functional
plasticity or vulnerability after early injury? Pediatrics 2005;116:
1374e82.
Forsyth RJ. Back to the future; rehabilitation of children after brain
injury. Arch Dis Child 2010;95:554e9.
Morgan AT, Vogel AP. A Cochrane review of treatment for dysarthria following acquired brain injury in children and adolescents.
Eur J Phys Rehabil Med 2009;45:197e204.
Slomine B, Locascio G. Cognitive rehabilitation for children with
acquired brain injury. Dev Disabil Res Rev 2009;15:133e43.
Limond J, Leeke R. Practitioner Review: Cognitive rehabilitation for
children with acquired brain injury. J Child Psychol Psychiatr 2005;
464:339e52.
Laatsch L, Harrington D, Hotz G, et al. An evidence-based review of
cognitive and behavioral rehabilitation treatment studies in children with acquired brain injury. J Head Trauma Rehabil 2007;22:
248e56.
Forsyth RJ, Thuy V, Salorio C, Christensen J, Holford N. Review:
Efcient rehabilitation trial designs using disease progress
modeling: A pediatric traumatic brain injury example. Neurorehabil Neural Repair 2010;24:225e34.
Turner-Stokes L. Introduction to The effectiveness of rehabilitation: a critical review of the evidence Closing Remarks. Clin
Rehabil 1999;13:83.
Logan LR, Hickman RR, Harris SR, Heriza CB. Single-subject
research design: recommendations for levels of evidence and
quality rating. Dev Med Child Neurol 2008;50:99e103.
Edlund W, Gronseth G, So Y, Franklin G. Clinical practice guideline
process manual. St Paul: Am Acad Neurol 2004. p. 5e55.
Berger E, Woergoetter G, Oppolzer A, Kessler J, Vavrik K, Fiala S.
Neurological rehabilitation in children and adolescents. Pediatr
Rehab 1997;1:229e33.
Wiseman-Hakes C, Stewart ML, Wasserman R, Schuller R. Peer
group training of pragmatic skills in adolescents with acquired
brain injury. J Head Trauma Rehabil 1998;13:23e38.
Pace GM, Schlund MW, Hazard-Haupt T, et al. Characteristics and
outcomes of a home and community-based neuro- rehabilitation
program. Brain Injury 1999;13:535e46.
Dumas HM, Haley SM, Ludlow LH, Rabin JP. Functional recovery in
pediatric traumatic brain injury during inpatient rehabilitation.
Am J Phys Med Rehabil 2002;81:661e9.
Chen CC, Heinemann AW, Bode RK, Granger CV, Mallinson T.
Impact of pediatric rehabilitation services on childrens functional
outcome. Am J Occupational Ther 2004;58:44e54.
Dumas HM, Haley SM, Carey TM, Ni PS. The relationship between
functional mobility and the intensity of physical therapy intervention in children with traumatic brain injury. Pediatr Phys Ther
2004;16:157e64.
Abdullah JM, Kumaraswamy N, Awang N, Ghazali MM,
Abdullah MR. Persistence of cognitive decits following pediatric
head injury without professional rehabilitation in rural east coast
Malaysia. Asian J Surg 2005;28:163e7.
Dumas HM, Haley SM, Ludlow LH. Achieving a minimally important difference in physical function during pediatric inpatient
rehabilitation. Int J Rehabil Res 2008;31:257e60.
Tepas JJ III, Leaphart CL, Pieper P, et al. The effect of delay in
rehabilitation on outcome of severe traumatic brain injury.
J Pediatr Surg 2009;44:368e72.
Melchers P, Maluck A, Suhr L, Scholten S, Lehmkuhl G. An early
onset rehabilitation program for children and adolescents after
traumatic brain injury (TBI): Methods and rst results. Restor
Neuro Neurosci 1999;14:153e60.
Swaine BR, Pless IB, Friedman DS, Montes JL. Effectiveness of
a head injury program for children. Am J Phys Med Rehabil 2000;
79:412e20.
Braga LW, Da Paz AC Jr, Ylvisaker M. Direct clinician-delivered
versus indirect family-supported rehabilitation of children with
traumatic brain injury: A randomized controlled trial. Brain Injury
2005;19:819e31.
vant Hooft I, Andersson K, Bergman B, Sejersen T, Von Wendt L,
Bartfai A. Benecial effect from a cognitive training program on

G. Tal, E. Tirosh / Pediatric Neurology 48 (2013) 424e431

children with acquired brain injuries demonstrated in a controlled
study. Brain Injury 2005;19:511e8.
[28] vant Hooft I, Andersson K, Bergman B, Sejersen T, von Wendt L,
Bartfai A. Sustained favorable effect of cognitive training in children with acquired brain injuries. Neuro Rehabil 2007;22:
109e16.
[29] Katz-Leurer M, Rotem H, Keren O, Meyer S. The effect of a homebased task-oriented exercise program on motor and balance
performance in children with spastic cerebral palsy and severe
traumatic brain injury. Clin Rehabil 2009;23:714e24.
[30] Zhu XL, Poon WS, Chan CH, Chan SH. Does intensive rehabilitation
improve the functional outcome of patients with traumatic brain
injury? Interim results of a randomized controlled trial. Br J Neurosurg 2001;15:464e73.

431

[31] NIH consensus development panel on rehabilitation of persons

with traumatic brain injury: rehabilitation of persons with traumatic brain injury. JAMA 1999;282:974e83.
[32] Weinstein MC, OBrien B, Hornberger J, et al. Principles of good
practice for decision analytic modeling in health-care evaluation:
report of the ISPOR task force on good research practice-modeling
studies. Value Health 2003;6:9e17.
[33] Horn SD, Gassaway J. Practice based evidence incorporating clinical heterogeneity and patient-reported outcomes for comparative
effectiveness research. Med Care 2010;48:S17e22.
[34] Dalen EH, Niquist A, Saebu M, Roe C, Bautz-Holter E. Implementation of ICF in goal setting in rehabilitation of children with
chronic disabilities at Beitostolen Healthsports Center. Disab
Rehabil [Epub ahead of print] doi:10.3109/09638288.