DEVELOPMENTAL MEDICINE & CHILD NEUROLOGY
REVIEW
A systematic review of psychological interventions to alleviate cognitive and psychosocial problems in children with acquired brain injury KIMBERLEY A ROSS 1 | LIAM DORRIS 1 , 2 | TOM MCMILLAN 1 1 Academic Unit for Mental Health and Wellbeing, University of Glasgow. 2 Fraser of Allander Neurosciences Unit, Royal Hospital for Sick Children, Glasgow, UK. Correspondence to Dr Liam Dorris at Fraser of Allander Neurosciences Unit, Royal Hospital for Sick Children, Glasgow G3 8SJ, UK. E-mail:
[email protected]
PUBLICATION DATA
AIM It is now generally accepted that paediatric acquired brain injury (ABI) can have an impact on
Accepted for publication 21st February 2011. Published online 19th April 2011.
a child’s cognitive, social, and behavioural functioning. However, the lack of guidelines on effective interventions for the affected children and their families, particularly beyond the acute recovery phase, can limit access to effective support. We provide a systematic review of the literature on the effectiveness of psychological interventions aimed at alleviating cognitive and psychosocial outcomes after paediatric ABI. METHOD The search used the Ovid MEDLINE, Embase, Web of Knowledge, and EBSCO databases and hand searches of key journals. Nine studies met inclusion criteria: five intervention studies of cognitive outcome and four of psychosocial outcome. Effect sizes and methodological quality ratings were calculated using specific criteria. RESULTS Only two of the nine studies were rated as high quality. There was limited evidence for effective interventions for cognitive outcomes (attention, memory, and learning difficulties). For psychosocial outcomes, there was evidence that interventions can alleviate internalizing symptoms. INTERPRETATION Although there are some encouraging findings, there is a need for further, more rigorously designed, and better controlled research in this important area. We discuss how future research may consider issues such as age-appropriate interventions, the delivery format, and optimum post-injury timing of interventions, as well as multicentre collaborations.
ABBREVIATIONS
ABI Acquired brain injury AMAT-C Amsterdam Memory and Attention Training for Children RCT Randomized controlled trial TBI Traumatic brain injury
Acquired brain injury (ABI) includes closed and open traumatic brain injury (TBI), infections of the central nervous system, such as meningitis, tumours, radiation, hypoxia, and stroke.1 A common feature of these pathologies is the potential for alteration of brain function and consequent changes in behaviour. It is now generally accepted that paediatric ABI can have an impact on a child’s cognitive, behavioural, and social functioning.2 Despite this, there is wide variability in the type and intensity of treatment received by children who have an ABI. It has been suggested that, after the acute recovery phase, service provision is poor and this may be because visible indicators of injury often do not persist.3 The literature on interventions for adults with ABI is better developed than that on paediatric ABI.4 This may, in part, be related to the ongoing debate about plasticity versus early vulnerability in relation to childhood brain insult. The plasticity theory suggests that the earlier the brain insult, the better the functional outcome. There is evidence, for example, of cortical reorganization and greater sparing of cognitive functioning if injuries occur at a younger age.5 However, Anderson et al.6
did not find evidence in support of this theory in their study of cognitive functioning in children who had sustained a TBI. They found that brain injury in infancy was more detrimental than in later life. In support of the early vulnerability theory, the rehabilitation literature broadly suggests that even if children appear relatively functionally intact immediately after insult, they may fail to keep up with their non-injured peers in terms of developmental gains as they get older. This has led to the clinical impression of children with ABI often ‘growing into’ their deficits.7 This highlights the need for effective interventions for children with ABI, and the need for the interventions to be available throughout the lifespan, as cognitive and psychosocial difficulties may not become apparent until late adolescence or early adulthood. This article reviews the literature on the efficacy of interventions aimed at alleviating cognitive and psychosocial difficulties in children with ABI. It focuses on interventions administered after the acute recovery phase, in view of the aforementioned evidence suggesting that there is a lack of longer-term community-based support. Before presenting the
692 DOI: 10.1111/j.1469-8749.2011.03976.x
ª The Authors. Developmental Medicine & Child Neurology ª 2011 Mac Keith Press
evidence for interventions, the literature on cognitive and psychosocial outcomes after paediatric ABI is summarized, highlighting specific areas of concern.
Cognitive outcomes Deficits in attention, memory, and executive functioning after paediatric ABI can have a negative impact on a child’s school, home, and community life.8 Babikian and Asarnow9 reviewed cognitive outcomes after paediatric TBI and found a dose– response relation with injury severity. In the severe TBI group, they found moderate to large effects in terms of Performance IQ, executive functioning, processing speed, attention, verbal immediate, and delayed memory. Despite some recovery during the first 2 years after injury, the severe TBI group did not catch up with their non-injured peers and, in fact, fell further behind in most cognitive domains. They highlighted the importance of focused and specific cognitive interventions in children with severe TBI. Psychosocial outcomes Psychosocial difficulties are reported even after mild paediatric brain injury.10 McKinlay et al.11 found that mild TBI in preschool children was associated with persisting negative effects on psychosocial development in adolescence. During adolescence these children were more likely to show symptoms of attention-deficit–hyperactivity disorder, conduct disorder, substance abuse, and mood disorders. Andrews et al.12 compared children with mild, moderate, or severe TBI with a comparison group matched for sex, age, and socio-economic status. Children with TBI had lower self-esteem, were lonelier, and had higher rates of maladaptive and antisocial behaviour, and described themselves as less socially competent than children without brain injury.12 Those with more severe injuries reported having fewer friends.13 Janusz et al.14 compared children with moderate and severe TBI with those with orthopaedic injuries. They found that children with severe brain injury had deficits in social problem solving. Fletcher et al.2 reported that children with severe brain injuries had more problems at school and engaged in fewer social activities than less severely injured groups. After paediatric ABI, the incidence of newly emerging behavioural disorders has been estimated to range between 35% and 70%.15–17 These social and behavioural difficulties may not only have an impact on the children and their peer relationships but also on family relationships: there is evidence of family strain, emotional difficulties, and burden in families of children with ABI.18 Wade et al.19 argue for family-centred approaches when intervening with children who have an ABI. They suggest that an ABI affects the whole family; parents experience high levels of distress when facing physical and personality changes in their child, and children develop psychosocial problems that are often not addressed. Given the evidence for the occurrence of psychosocial and cognitive difficulties in children after ABI, there is a need for effective interventions. The present review considers the evidence for interventions in the post-acute period after childhood ABI. Previous reviews differ in important ways. Some
What this paper adds • It demonstrates the current paucity of high-quality intervention studies for cognitive and psychosocial problems in paediatric ABI. • It provides evidence and detailed descriptions of effective interventions for cognitive and psychosocial problems. • It outlines suggestions for future research.
focus on the acute recovery phase.8,20 Some are practitioner reviews and, therefore, do not detail search strategies or calculate effect sizes, making it difficult to replicate these studies.8,21–23 Others do not include promising recent research, particularly around family-based interventions.3 One highquality systematic review15 focused on behavioural outcomes drawing primarily on single-case, multiple baseline designs, and, therefore, had a different focus from the current review.
Research questions We asked the following questions: Are there effective interventions for 1) cognitive effects, and 2) psychosocial effects, of paediatric ABI? METHOD Search strategy Studies were sought from Ovid MEDLINE, Embase, Web of Knowledge, and EBSCO databases (PsycINFO, CINAHL, Health Source: Nursing ⁄ Academic Section, International Bibliography of the Social Sciences, Professional Development Collection, PsycARTICLES, Psychology & Behavioural Sciences Collection, SocINDEX) by searching the following text words: (1) (acquired brain injur* or acquired head injur* or traumatic brain injur* or traumatic head injur*); (2) (intervention* or rehabilitat* or train*); (3) (psycho?social or social* or behavio?r* or emotion* or anx* or cognitive or peer or distress* or memor* or attention or executive); (4) (school?age or pre?school or child or children or adolesc* or teen* or youth or paediatric or pediatric or childhood). The symbols * and ? denote database operators, which include truncations or possible extra letters in the term to be included within the search. These four searches were then combined using ‘AND’. Figure 1 provides a flow diagram of the search strategy. The initial search yielded 536 papers published in or before the third week of March 2010. Inclusion and exclusion criteria were then applied. Inclusion criteria were (1) intervention studies addressing cognitive or psychosocial outcomes, (2) participants aged up to 18 years, and (3) peer-reviewed journal articles. Exclusion criteria were (1) articles not addressing intervention, (2) theoretical articles or descriptions of rehabilitation programmes, with no specific intervention, (3) review articles, (4) articles without adequate specification of interventions, (5) articles that did not include participants with a primary diagnosis of ABI or TBI, (6) single case reports or series of multiple baseline experiments, (7) articles describing surgical or pharmacological interventions, (8) articles not written in English, (9) studies focusing on outcomes after childhood cancers only, as this population may have distinct psychosocial difficulties not present within the general ABI population, and (10) studies where participants were less than 3 months after injury, as post-concussive sympReview 693
536 papers obtained from the computerized database searches.
394 were excluded on title alone, leaving 142 articles.
A further 112 excluded based on abstract, leaving 30 articles.
On reading the full original papers, a further 19 were excluded, for the following reasons: •
seven did not describe the intervention well (often rehabilitation studies);
•
two did not measure specific cognitive or psychosocial outcomes;
•
two focused on parental measures only;
•
two were multiple baseline designs;
•
two included adult participants;
• •
three included participants < 4mo after injury; one study had no standardized outcome measures.
Each of the 11 papers’ methodological quality was rated using a specific set of criteria.
Cognitive outcomes: six articles, two of which were of the same study and therefore counted as one, leaving a total of five studies. Psychosocial outcomes: five articles, two of which were the same study published separately, therefore counted as one, leaving a total of four studies.
Figure 1: Flow diagram of papers excluded at each search stage.
toms are thought to resolve around this time and, therefore, this is after the acute recovery time.7 Using these exclusions, 394 articles were removed on the basis of title, mainly because they were not intervention studies or surgical or pharmacological intervention studies. Of the remaining 142, a further 112 articles were excluded based on their abstract as they did not fulfil the specified inclusion and exclusion criteria. Again, this was mainly because the studies did not relate to a direct intervention. A full article search was performed on the remaining 30 studies, of which 19 (refs 24–42) were excluded for reasons specified in Figure 1. A hand search of the journals Neuropsychology and the Journal of Head Trauma Rehabilitation over the past 10 years was performed. These journals were chosen as they yielded the highest quality papers already included in the review or included two studies that were cited frequently in the literature. No additional article that met inclusion criteria was identified by hand searches. Eleven articles remained for review. Of these, six studies assessed cognitive outcomes; however, two of these reported the same primary data and were counted as one article. Two 694 Developmental Medicine & Child Neurology 2011, 53: 692–701
were pre- and post-design single group studies and three were randomized controlled trials (RCTs) or group-controlled studies. There were five articles assessing psychosocial outcomes, again, two reported the same primary data and were considered as one. Of these four studies, two were RCTs, and two were single group pre- and post-designed studies.
Calculation of effect size Effect sizes were calculated by the authors for three articles.19,43,44 Two other papers calculated effect sizes based on partial g2. Wade et al.19 used a variant of Hedge’s g for pre– post design studies. This method appears to be less susceptible to potential sources of bias,45 was used in a similar review of adult literature,46 and was used here as the preferred procedure for calculating effect sizes. Equation 1 is the formula used for calculating effect sizes in single group pretest–posttest designs; equation 2 is the formula used to calculate effect sizes in independent group pretest–posttest designs. Mpost;Exp Mpre;Exp SDpre;Exp
ð1Þ
Mpost;Exp Mpre;Exp Mpost;Com Mpre;Com SDpre;Exp SDpre;Com
ð2Þ
where M is the mean, Exp is the experimental group, Com is the comparison group, Post is the posttest, Pre is the pretest, and SD is the standard deviation. When there were multiple dependent variables, effect sizes were calculated as mean effect sizes for all dependent variables within the study, or summary scores were used only to calculate effect sizes. There was insufficient information in one article to calculate effect sizes.47,48 A meta-analysis was not felt to be appropriate owing to the small number of studies, the differences in methodological designs, and measurement of different cognitive domains. Related to this, we have not addressed the ‘file drawer’ problem: bias attributable to the tendency for studies with larger effect sizes to be published. Statistical methods have been developed for dealing with this issue within meta-analysis; however, this issue only becomes a problem when there have been many studies conducted and where a significant proportion remains unpublished. This is very unlikely within the emerging evidence base we currently describe.
Methodological appraisal of included studies Criteria for the appraisal of articles were based on Consolidated Standards of Reporting Trials (CONSORT) guidelines, with items added that are specific to ABI (See Appendix 1). Each of the 26 items was awarded a score of 1 (if the criterion was met) and 0 if not met or it was not possible to determine from information given in the article. Therefore, each paper was rated out of 26. Papers that met 75% of the methodological criteria specified were considered of ‘high’ quality. Papers that were rated between 50% and 75% were deemed to have ‘moderate’ quality, and those achieving less than 50% were ‘lower’ quality. To assess the reliability of this tool, a second reviewer using the same tool rated the final nine papers. Overall percentage agreement was high (98%). Individual disagreements were resolved by discussion with the independent reviewer. RESULTS Tables I and II show the methodological quality ratings for each of the cognitive and psychosocial outcome studies respectively. Cognitive outcomes In the cognitive outcomes domain, one study was of ‘high’ methodological quality, three were of ‘moderate’ quality, and the remaining one was of low quality. High quality Galbiati et al.49 compared a treatment group that received attention-specific neuropsychological training with a comparison group that had opted out of the training programme. The training programme was relatively intense, with 45-minute sessions with a therapist offered four times a week for 6 months. Large effect sizes were found in terms of concentra-
tion, impulsivity, and distractibility, but with a small effect for overall cognitive functioning. This paper was of particularly high methodological quality, scoring the highest methodological rating of all included studies. However, the confounding factor of time with therapist was not controlled for.
Moderate ⁄ low quality The Amsterdam Memory and Attention Training for Children (AMAT-C) is a child-specific cognitive rehabilitation programme that focuses on sustained attention, focused attention, divided attention, memory strategies, and repetition. The training takes place on a one-to-one basis and is relatively intensive. Van’t Hooft et al.47 studied the efficacy of the AMAT-C programme in an RCT. Six months after intervention, improvements in selective attention and verbal working memory were sustained in the treatment group. However, owing to limited information, effect sizes could not be calculated. Therefore, although significant effects were found, it is difficult to ascertain the extent of the AMAT-C intervention effects. Sjo et al.50 applied the AMAT-C programme in a school setting to a group of children with ABI. Despite a significant improvement in attention after intervention, the effect size was small, although a medium effect size was found for improvements on measures of learning and memory. This was a pilot study and, not surprisingly, there were significant methodological flaws including the absence of a comparison group and a very small sample size (n=7). The findings, therefore, have to be taken as tentative. Brett and Laatsch51 also studied the effectiveness of an individualized school-based cognitive rehabilitation programme for children with ABI. Improvements were found after intervention in measures of memory functioning, and these effects were found to be large. However, the absence of a comparison group makes it difficult to exclude natural recovery as the mechanism of change. There was also no mention of the severity of ABI in the participants, which makes it difficult to consider the generalizability of the results. Braga et al.52 looked at cognitive outcomes using two treatment conditions: a direct clinician-delivered intervention group and an indirect family treatment group. The same professionals treated both groups. However, in the indirect family supported group, the family members implemented the intervention with bi-weekly support from care managers after initially receiving intensive training. In the direct cliniciandelivered clinic-based intervention, the clinician used conventional rehabilitation procedures and conducted all interventions without a parent present. In both groups, parents attended a support group. On a measure of intellectual function the direct clinician-delivered treatment group showed improvement across the treatment, with a small effect size, which was not statistically significant. The indirect familysupported treatment group showed greater improvement in intellect with a moderate effect size that was statistically significant. The authors speculate that findings may have been influenced if a more intense intervention was provided by parents in everyday life or if lower levels of parental stress resulted Review 695
696 Developmental Medicine & Child Neurology 2011, 53: 692–701
Before and after trial, no comparison group. Looked at the feasibility of the AMAT-C programme being integrated into the school setting
RCT. Two groups: clinician-delivered or family-supported intervention (SARAH). Both groups received 1y of intensive individualized rehabilitation. Aimed to determine if parents can be trained to deliver rehabilitation exercises effectively to improve physical and cognitive outcomes Before and after trial no comparison group to assess the effectiveness of a school individualized cognitive rehabilitation programme that focussed on (1) alertness, attention, and concentration, (2) perception and memory, (3) executive processes
Group comparison study. Had a comparison group but not randomly assigned. One group received attentionspecific neuropsychological training for 6mo; the comparison group did not receive treatment RCT. Cognitive training programme (AMAT-C) administered by teachers or parents for 30min per day for 17wk with once-weekly contact from therapist. Comparison group had an interactive activity for the same time. 6mo follow-up study (2007)
Description of intervention
n=7 children with ABI (no information on severity) Aged 8–16y. 10mo–8y after injury
n=10 high-school students with ABI (no mention of severity). All were 1–16y after injury
87 children who had sustained a moderate to severe TBI. Aged 5–12y. 6–30mo after injury
n=65 children with attentional problems following severe TBI. (n=40 treatment group; n=25 comparison group). Aged 6–18y. 6–10mo after injury n=38 children with attentional and memory deficits after ABI. Aged 9–17y. 1–5y after injury. Mild, moderate, and severe TBI, encephalitis, anoxia, brain malignancies
Sample
Attention: Benton Visual Form Discrimination test Stroop test Intellectual functioning: TONI-2 WISC-IV (picture completion and freedom from distractibility) Memory: Wide Range Assessment of Memory and Learning Problem solving: Wisconsin Card Sorting test Attention ⁄ executive functioning: TEACh BRIEF Learning and memory: WISC-II
Attention: Visual and Auditory - Reaction Time Tests Gordon diagnostic System Stroop test. Binary choice Test Coding Trail making test Memory: Digit span test 15-word test Rey–Osterrieth Complex figure Test. Rivermead Behavioural Memory test Intellectual functioning: WISC-III
Intellectual functioning: WISC-R or WAIS-R Attention: Continuous Performance test II
Cognitive outcome measuresa
Attention ⁄ executive functioning: Overall g=0.34, small Learning and memory: Overall g=0.58, medium
Attention: Overall g=0.45, small– medium Intellectual functioning: Overall g=0.42, small Memory: Overall g=1.02, large Problem solving: Overall g=0.02, small
Clinician-delivered group overall g=0.18, small Family-supported group overall g=0.66, medium
Effect sizes could not be calculated
Intellectual functioning: g=0.1, small Attention: g=1.35, large
Effect sizesb
Significant improvements in terms of learning and memory.
After treatment, students demonstrated significant improvements in general memory ability.
Children in the familysupported treatment group experienced superior cognitive outcomes than the direct clinician-led treatment group. Improvements were statistically and clinically substantial.
At follow-up, the treatment group showed significantly more improvement in terms of attention. No significant difference in intellectual functioning. The cognitive training group showed improved complex attention and memory functions. These improvements were sustained at 6mo after intervention.
Findings
Titles of outcomes measures: WISC-R, Wechsler Intelligence Scale for Children, Revised; WAIS-R, Wechsler Adult Intelligence Scale, Revised; TEACh, Test of Everyday Attention for Children; BRIEF, Behaviour Rating Inventory of Executive Functioning; TONI, Test of Nonverbal Intelligence. bEffect sizes: g=0–0.5, small; g=0.5–0.8, medium; g>0.8, large. TBI, traumatic brain injury; RCT, randomized controlled trial; ABI, acquired brain injury. AMAT-C, Amsterdam Memory and Attention Training for Children.
a
Sjo et al.50 35%, low
Brett and Laatsch51 54%, moderate
Braga et al.52 73%, moderate
Van’t Hooft et al.47,48 73%, moderate
Galbiati et al.49 85%, high
Study and quality rating
Table I: Description and methodological quality ratings of included studies: cognitive outcomes
Review 697
Sample
Before and after trial (no comparison group). Evaluation of an intervention programme for treating adolescents with pragmatic communication deficits secondary to ABI. The programme ran for 6wk, 4d per week, for 1h each day, the intervention focused on four main modules: initiation conversation; topic maintenance; turn taking; active listening
n=6 children aged 14–17y with ABI and pragmatic deficits and subsequent difficulties in social interaction. 4mo–9y after injury
n=39 families of RCT. children who had sustained a moderate Investigated a 14-session online to severe TBI. computer-based training family Aged 8–13y. problem-solving intervention, 6–19mo after injury providing training in problem solving, communication, and behaviour management and meetings. Therapist meeting every 2wk. Internet resource comparison group. Families continued to receive any psychosocial care they were receiving RCT. n=32 families of children aged 5–16y who had sustained Studied the efficacy of a sevena moderate to severe TBI. session manualized family-centred problem-solving intervention, delivered All children were 4–12mo after injury. over 6mo to participating families. Intervention focused on a five-step process: aim; brainstorm; choose; do it; evaluate. Usual care comparison group. Before and after trial (no comparison n=6 families with a group) to assess the feasibility and child who had sustained a moderate to preliminary efficacy for an online family severe TBI more than 15mo previously. problem-solving treatment. The Aged 5–16y intervention addressed cognitive appraisals, coping, and family communication through self-guided web pages and one-to-one video-conference sessions with a trained therapist
Description of intervention
Child behaviour: CBCL CBCL internalizing CBCL externalizing Parental adjustment: BSI-GSI: Parent–child interaction: CBQ Child adjustment: BRIEF HCSBS Parent–child interaction: IBQ PARQ (school conflict scale): Family functioning: FAD–GF Parent adjustment: FBII PSI GSISC Anxiety Inventory Centre for Epidemiological Studies Depression Scale Pragmatic skills: RICE-RSPCS CPS Behaviour: VABS – socialization
Child adjustment: CBCL HCSBS
Psychosocial outcome measuresa Findings
RICE-RSPCS: g=1.91, large CPS: g=0.72, large VABS: (socialization) g=0.24, small
CBCL: Total g=0.12, small Internalizing g=0.72, medium Externalizing g=0.04, small Parent adjustment: Overall g=0.55, medium Parent–child interaction: Overall g=0.51, medium Child adjustment: Overall g=0.64, medium Parent–child interaction: Overall g=0.44, small Family functioning: Overall g=0.15, small Parent adjustment: Overall g=0.83, high
Significant group changes in the RICE-RSPCS and CPS measures. No statistically significant difference in VABS socialization domain score.
Improvements in antisocial behaviours, and reductions in conflict with parents regarding school issues but no improvement in self-reported depressive symptoms.
In the treatment group, there were significant reductions in child behaviour problems; particularly internalizing symptoms. No significant differences in terms of parent– child conflict.
The treatment group CBCL: Total g=0.24, small reported better child selfInternalizing g=0.45, small– management ⁄ compliance at follow-up. Overall effect sizes medium large for children of lower socioExternalizing g=0.01, small economic status and children HCSBS: Social competence and antisocial older than 11y behaviour overall g=0.42, small
Effect sizesb
Titles of outcomes measures: CBCL, Child Behaviour Checklist; HCSBS, Home and Community Social Behaviour Scale; BSI-GSI, Brief Symptom Inventory, Global Severity Index; CBQ, Conflict Behaviour Questionnaire; IBQ, Interaction Behavior Questionnaire; PARQ, Parent-Adolescent Relationship Questionnaire; FAD-GF, General Functioning Scale of Family Assessment Device; FBII, Family Burden of Injury Inventory; PSI, Parenting Stress Inventory; GSISC, Global Severity Index of Symptom Checklist; RICE-RSPCS, Rehabilitation Institute of Chicago Rating Scale of Pragmatic Communication Skills; CPS, Communication Performance Scale; VABS, Vineland Adaptive Behaviour Scale. bEffect sizes: g=0–0.5, small; g=0.5–0.8, medium; g>0.8, large. RCT, randomized controlled trial; TBI, traumatic brain injury; ABI, acquired brain injury
a
Wiseman-Hakes et al.54 42%, low
Wade et al.19 73%, moderate Wade et al.53 58%, moderate
Wade et al.44 73%, moderate
Wade et al.43 77%, high
Study and quality rating
Table II: Description and methodological quality ratings of included studies: psychosocial outcomes
from parents feeling more involved in their child’s care. This article has several strengths in its design, including being a well-described RCT with those measuring outcomes blind to the treatment condition.
Psychosocial outcomes In terms of quality rating of psychosocial outcomes, of the four studies one was of ‘high’ quality, two were of ‘moderate’ quality, and one was of ‘low’ methodological quality. High quality Wade et al.43 looked at the efficacy of an online family problem-solving intervention using an RCT design. The comparison group had the same access to high-speed Internet and the chat rooms and online support groups as the treatment group but did not have the 14 intervention sessions or the sessions with the therapist. Significant differences were not found between treatment and comparison groups for behaviour or externalizing difficulties. However, small to medium effects were found for internalizing difficulties. After treatment, children in the treatment group were more compliant with parental requests. The authors suggest that these changes may have a positive impact in terms of psychosocial functioning, as the children may have greater control in social situations. However, it is important to note that there were no direct measures of social competence and all measures were parent completed. This study has several methodological strengths and the authors go some way to teasing out the characteristics of individuals who may benefit most from the intervention (children of lower socio-economic status and those older than 11y). Weaknesses include a sample skewed towards less severe injuries, and two families who dropped out of treatment had children with more significant social deficits at baseline. Moderate ⁄ low quality In another RCT, Wade et al.44 studied the efficacy of a relatively brief family problem-solving intervention of seven sessions. Improvements were found for internalizing difficulties such as anxiety ⁄ depression, withdrawal, and for parent– child interactions, with medium effect sizes noted. There were several methodological strengths, although all behavioural measures were parent completed. Wade et al.19,53 also conducted a single group before and after trial evaluating the effectiveness of the online version of the family problem-solving intervention. Families were also given a weekly video-conference session with a therapist. They found this intervention to be feasible for both the child with ABI and the family members. The intervention improved child and parental adjustment. However, there were several limitations to the study, including a sample of six children with TBI and no comparison group. Also, only two of the measures were completed by the child and on one, mean depressive symptoms actually rose slightly, albeit remaining in the normal range. Wiseman-Hakes et al.54 evaluated an intervention for pragmatic communication deficits. They hypothesized that improved pragmatic skills would have a ‘ripple effect’ for social 698 Developmental Medicine & Child Neurology 2011, 53: 692–701
competency. The sample size was small and included a heterogeneous group of in-patients and outpatients, closed head injury, and ABI; time after injury ranged from 3 months to 9 years. The programme was intensive and significant results were observed in terms of pragmatic communication skills. However, the measures of pragmatic communication skills were observational. As there was no comparison group, the rater was unlikely to be blind to the treatment, introducing a potential for bias. No difference was found between pre- and post-treatment on the Vineland Adaptive Behavioural Scale, with no clear evidence for a ‘ripple effect’ on social competence. It is also noteworthy that proxies, and not the adolescents themselves, completed measures of social competence. Therefore, although this article provides some evidence for improvement in pragmatic communication skills, there is little evidence for this having a positive impact on psychosocial functioning in everyday life.
DISCUSSION Cognitive outcomes: main results One high-quality study provides evidence for improvement in attention,49 and effect sizes were large, and statistically significant. There was also evidence in one study that improvements in attention were sustained at 6-month follow-up.48 There is also evidence for the efficacy of interventions for memory and learning difficulties,47,50,51 and that improvements are sustained at 6 months after intervention.48 However, the methodological quality of these studies varies, with two of the three studies not including a comparison group, making it difficult to conclude that the intervention itself led to improvement. The evidence for interventions improving general intellectual functioning is not as encouraging. Galbiati et al.49 found no post-intervention change on measures of IQ. However, Braga et al.52 did find superior outcomes using the Wechsler Intelligence Scale for Children, third edition (WISC-III) in their family-supported intervention compared with the direct clinician intervention. This suggests that even if a treatment programme is individualized, it is important for professionals to be integrated within a family system and that the involvement of the family members in cognitive intervention programmes could be of key importance. There is also evidence that the WISC-III is insensitive as a measure of the specific neuropsychological vulnerabilities that occur after brain injury.54 Future research should perhaps supplement these measures of intellectual ability. The literature suggests that for interventions to be effective they need to be provided intensively. Cognitive outcomes: future research It is difficult to determine if post-intervention gains in attention skills can be generalized beyond the clinical setting. Future research needs to consider school and home settings. There is also need to determine the ‘active ingredients’ of the interventions that are most effective, including matching interventions to individual child and family characteristics. Further good quality studies are required that include the use of RCTs and control for confounds such as time spent with the thera-
pist. Of the studies reviewed, only one looked at persistence of effects at follow-up;48 again, more research is required to ascertain longer-term maintenance of functional gains. Studies by Brett and Laatsch51 and Braga et al.52 are encouraging as they found positive outcomes for programmes implemented in the school and home settings. More research is needed looking at the efficiency of cognitive intervention programmes administered by family members or educational staff. This may follow training and supervision from clinicians or therapists, and may prove to be a cost-effective method of administration.
Psychosocial outcomes: main results There are some encouraging findings, with key outcomes relating to improvements in internalizing symptoms, such as depression ⁄ anxiety and withdrawal,43,44 and in parent–child interactions and parental adjustment.44 Improvements in child adjustment19,53 are difficult to attribute fully to the intervention itself, as there was no control or comparison group. When considering which factors have an impact on the effectiveness of these interventions, it appears that children of lower socio-economic status and children injured over age 11 years may benefit most from online family intervention.44 There is a key role for families, with three out of the four studies involving the family in the intervention. Psychosocial outcomes: future research In some studies that involve the whole family,19,43,44,53 the injured child’s views can become lost as most data were collected from parents. This has prompted some authors, for example Theunissen et al.,55 to conclude that there is a need for research to include child report and teacher-rated measures to gain a more holistic perspective. Overall, more research is needed with other treatment comparison groups and with larger sample sizes. The psychosocial outcomes also relate mostly to childhood TBI, as only one of the total 82 participants over the four studies had a non-traumatic ABI. Strengths, limitations, and future directions To our knowledge, this is the first systematic review of its kind that focuses specifically on the post-acute recovery phase. This is at a time when there appears to be less support, yet a high need for it as the gap in cognitive and psychosocial development between children with ABI and their non-injured peers widens.7 Other reviews in the area have either been practitioner reviews or have focused on less methodologically highquality studies. The present study has limitations, particularly relating to cognitive outcome domains being grouped together. As the evidence base develops further, this will allow
more valid stratification and analysis of specific cognitive domains. It also appears that most of the studies used a wide age range, and research that tailors interventions for specific age ranges would be of real interest for practitioners. This may be particularly important for early intervention for younger children to maximize their developmental trajectory. Another important factor concerns time since injury, which ranged from 4 months to 16 years in the studies reviewed; it may be useful to explore the effect that time since injury has for the optimum effects of the intervention. Finally, this review is limited by the lack of good-quality outcome studies of paediatric rehabilitation. Conducting and evaluating interventions in this area is undoubtedly hampered by methodological issues such as choosing an appropriate comparison and ⁄ or treatment group, and specifying the ‘active components’ of alternative interventions; accounting for education and other supports; or denying access to an intervention where the expertise and resource may be greater than for existing care or other control conditions. These issues are problematical ethically and can often prevent the adoption of fully randomized controlled designs.
Practical applications A fundamental issue concerns the basic availability of postacute rehabilitation, which is heavily dependent on the national and local context of healthcare systems. Within the UK, paediatric cognitive rehabilitation is often provided after personal injury litigation by therapists in private practice through insurance companies, or on an ad hoc basis by learning support professionals at school. Neither of these contexts has been traditionally disposed towards systematic evaluation using the rigorous designs necessary for publication in scientific journals. Understanding these issues should, however, focus efforts to use and deploy the resources available within health and education teams locally as well as in planning larger-scale collaborations. In conclusion, the current literature provides limited but encouraging results, mainly in terms of effective interventions for alleviating attentional and internalizing behaviour difficulties. The adoption of multicentre collaboration may allow many of the methodological concerns above mentioned to be addressed, and may inform the development of appropriately evidenced guidelines both for service providers and for service commissioners. ACKNOWLEDGEMENTS This research was partly funded by NHS Education Scotland. We wish to thank Dr Tom Kelly for comments helpful to the discussion.
REFERENCES 1. Slomine B, Locascio G. Cognitive rehabilitation for children
3. Laatsch L, Harrington D, Hotz G, et al. An evidence-based
5. Devinsky O, Perrine K, Llinas R, Luciano DJ, Dogali M.
with acquired brain injury. Dev Disabil Res Rev 2009; 15:
review of cognitive and behavioral rehabilitation treatment
Anterior temporal language areas in patients with early
133–43.
studies in children with acquired brain injury. J Head Trauma
onset of temporal lobe epilepsy. Ann Neurol 1993; 34: 727–
Rehabil 2007; 22: 248–56.
32.
2. Fletcher JM, Ewing-Cobbs L, Miner ME, Levin HS, Eisenberg HM. Behavioral changes after closed head injury in children. J Consult Clin Psychol 1990; 58: 93–8.
4. Limond J, Leeke R. Practitioner review: cognitive rehabilita-
6. Anderson V, Catroppa C, Morse S, Haritou F, Rosenfeld J.
tion for children with acquired brain injury. J Child Psychol
Functional plasticity or vulnerability after early brain injury?
Psychiatry 2005; 46: 339–52.
Pediatrics 2005; 116: 1374–82.
Review 699
7. Anderson V, Northam E, Hendy J, Wrennal J. Developmental neuropsychology: a clinical approach. Hove and New York: Taylor & Francis Psychology Press, 2001. 8. Ylvisaker M, Adelson PD, Braga LW, et al. Rehabilitation and ongoing support after pediatric TBI: twenty years of progress. J Head Trauma Rehabil 2005; 20: 95–109. 9. Babikian T, Asarnow R. Neurocognitive outcomes and recovery after pediatric TBI: meta-analytic review of the lit-
congenital and acquired brain injury. Anal Interv Dev Disabil 1985; 5: 107–28. 25. Burke WH, Wesolowski MD, Buyer DM, Zawlocki RJ. The rehabilitation of adolescents with traumatic brain injury: outcome and follow-up. Brain Inj 1990; 4: 371–8. 26. Carey JC, Wade SL, Wolfe CR. Lessons learned: the effect of prior technology use on web-based interventions. Cyberpsychol Behav 2008; 11: 188–95.
40. Thomas-Stonell N, Johnson P, Schuller R, Jutai J. Evaluation of a computer-based program for remediation of cognitive-communication skills. J Head Trauma Rehabil 1994; 9: 25–37. 41. Wade SL, Carey J, Wolfe CR. An online family intervention to reduce parental distress following pediatric brain injury. J Consult Clin Psychol 2006; 74: 445–54. 42. Wade SL, Oberjohn K, Burkhardt A, Greenberg I. Feasibil-
27. Chapman SS, Ewing CB, Mozzoni MP. Precision teaching
ity and preliminary efficacy of a web-based parenting skills
10. Limond J, Dorris L, McMillan TM. Quality of life in
and fluency training across cognitive, physical, and academic
program for young children with traumatic brain injury.
children with acquired brain injury: parent perspectives 1-5
tasks in children with traumatic brain injury: a multiple base-
years after injury. Brain Inj 2009; 23: 617–22.
line study. Behav Interv 2005; 20: 37–49.
erature. Neuropsychology 2009; 23: 283–96.
J Head Trauma Rehabil 2009; 24: 239–47. 43. Wade SL, Carey J, Wolfe CR. The efficacy of an online cog-
11. McKinlay A, Grace RC, Horwood LJ, Fergusson DM,
28. Chevignard M, Toure H, Brugel DG, Poirier J, Laurent-
nitive–behavioral family intervention in improving child
MacFarlane MR. Long-term behavioural outcomes of
Vannier A. A comprehensive model of care for rehabilitation
behavior and social competence following pediatric brain
pre-school mild traumatic brain injury. Child Care Health
of children with acquired brain injuries. Child Care Health
Dev 2010; 36: 22–30.
Dev 2010; 36: 31–43.
injury. Rehabil Psychol 2006; 51: 179–89. 44. Wade SL, Michaud L, Brown TM. Putting the pieces
12. Andrews TK, Rose FD, Johnson DA. Social and behavioural
29. Depompei R, Gillette Y, Goetz E, Xenopoulos-Oddsson A,
together: preliminary efficacy of a family problem-solving
effects of traumatic brain injury in children. Brain Inj 1998;
Bryen D, Dowds M. Practical applications for use of PDAs
intervention for children with traumatic brain injury. J Head
12: 133–8.
and smartphones with children and adolescents who have
13. Prigatano GP, Gupta S. Friends after traumatic brain injury in children. J Head Trauma Rehabil 2006; 21: 505–13.
traumatic brain injury. NeuroRehabilitation 2008; 23: 487– 99.
Trauma Rehabil 2006; 21: 57–67. 45. Morris SB, Deshon RP. Combining effect size estimates in meta-analysis with repeated measures and independentgroups designs. Psychol Methods 2002; 7: 105–25.
14. Janusz JA, Kirkwood MW, Yeates KO, Taylor HG. Social
30. Dise-Lewis JE, Lewis HC, Reichardt CS. BrainSTARS: pilot
problem-solving skills in children with traumatic brain
data on a team-based intervention program for students who
46. Rohling ML, Faust ME, Beverley B, Demakis G. Effectiveness
injury: long-term outcomes and prediction of social compe-
have acquired brain injury. J Head Trauma Rehabil 2009; 3:
of cognitive rehabilitation following acquired brain injury: a
tence. Child Neuropsychol 2002; 8: 179–94.
166–77.
meta-analytic re-examination of Cicerone et al.’s (2000, 2005)
15. Ylvisaker M, Turkstra L, Coehlo C, et al. Behavioural inter-
31. Dumas HM, Haley SM, Bedell GM, Hull EM. Social func-
ventions for children and adults with behaviour disorders
tion changes in children and adolescents with acquired brain
47. Van’t Hooft I, Andersson K, Bergman B, Sejersen T, von
after TBI: a systematic review of the evidence. Brain Inj
injury during inpatient rehabilitation. Pediatr Rehabil 2001;
Wendt L, Bartfai A. Beneficial effect from a cognitive train-
2007; 21: 769–805.
4: 177–85.
ing programme on children with acquired brain injuries dem-
systematic reviews. Neuropsychology 2009; 23: 20–39.
16. Max JE, Robin DA, Lindgren SD, et al. Traumatic brain
32. Dumas HM, Haley SM, Rabin JP. Short-term durability and
injury in children and adolescents: psychiatric disorders at
improvement of function in traumatic brain injury: a pilot
48. Van’t Hooft I, Andersson K, Bergman B, Sejersen T, von
two years. J Am Acad Child Adolesc Psychiatry 1997; 36:
study using the Paediatric Evaluation of Disability Inventory
Wendt L, Bartfai A. Sustained favorable effects of cognitive
1278–85.
(PEDI) classification levels. Brain Inj 2001; 15: 891–902.
training in children with acquired brain injuries. NeuroReha-
17. Costeff H, Groswasser Z, Landman Y, Brenner T. Survivors
33. Emanuelson I, Wendt LV, Hagberg I, et al. Early commu-
of severe traumatic brain injury in childhood. II. Late resid-
nity outreach intervention in children with acquired brain
ual disability. Scand J Rehabil Med Suppl 1985; 12: 10–5.
injury. Int J Rehabil Res 2003; 26: 257–64.
onstrated in a controlled study. Brain Inj 2005; 19: 511–8.
bilitation 2007; 22: 109–16. 49. Galbiati S, Recla M, Pastore V, et al. Attention remediation following traumatic brain injury in childhood and adoles-
18. Cole WR, Paulos SK, Cole CAS, Tankard C. A review of
34. Hotz GA, Castelblanco A, Lara IM, Weiss AD, Duncan R,
family intervention guidelines for pediatric acquired brain
Kuluz JW. Snoezelen: a controlled multi-sensory stimulation
50. Sjo M, Spellerberg S, Weidner S, Kihlgren M. Training of
injuries. Dev Disabil Res Rev 2009; 15: 159–66.
therapy for children recovering from severe brain injury.
attention and memory deficits in children with acquired brain
19. Wade SL, Wolfe CR, Brown TM, Pestian JP. Can a web-
Brain Inj 2006; 20: 879–88.
cence. Neuropsychology 2009; 23: 40–9.
injury. Acta Paediatr 2010; 99: 230–6.
based family problem-solving intervention work for children
35. Landis J, Hanten G, Levin HS, et al. Evaluation of the error-
51. Brett AW, Laatsch L. Cognitive rehabilitation therapy of
with traumatic brain injury? Rehabil Psychol 2005; 50: 337–
less learning technique in children with traumatic brain
brain-injured students in a public high school setting. Pediatr
45.
injury. Arch Phys Med Rehabil 2006; 6: 799–805.
Rehabil 1998; 2: 27–31.
20. Beaulieu C. Rehabilitation and outcome following pediatric
36. Marcantuono JT, Prigatano GP. A holistic brain injury reha-
52. Braga LW, Da Paz AC, Ylvisaker M. Direct clinician-deliv-
traumatic brain injury. Surg Clin North Am 2002; 82: 393–
bilitation program for school-age children. NeuroRehabilita-
ered versus indirect family-supported rehabilitation of chil-
408.
tion 2008; 23: 457–66.
dren with traumatic brain injury: a randomized controlled trial. Brain Inj 2005; 19: 819–31.
21. Warschausky S, Kewman D, Kay J. Empirically supported
37. McCauley SR, Pedroza C, Chapman SB, et al. Event-based
psychological and behavioral therapies in pediatric rehabilita-
prospective memory performance during subacute recovery
53. Wade SL, Wolfe C, Brown TM, Pestian JP. Putting the
tion of TBI. J Head Trauma Rehabil 1999; 14: 373–83.
following moderate to severe traumatic brain injury in
pieces together: preliminary efficacy of a web-based family
children: effects of monetary incentives. J Int Neuropsychol
intervention for children with traumatic brain injury. J Pedi-
22. Catroppa C, Anderson V. Traumatic brain injury in childhood: rehabilitation considerations. Dev Neurorehabil 2009; 12: 53–61. 23. Anderson V, Catroppa C. Advances in post-acute rehabilitation after childhood acquired brain injury: a focus on cognitive, behavioural and social domains. Am J Phys Med Rehabil 2006; 85: 767–87. 24. Bagnato SJ, Neisworth JT. Efficacy of interdisciplinary assessment and treatment for infants and preschoolers with
Soc 2010; 16: 335–41.
atr Psychol 2005; 30: 437–42.
38. Ponsford J, Willmott C, Rothwell A, et al. Impact of early
54. Wiseman-Hakes C, Stewart ML, Wasserman R, Schuller R.
intervention on outcome after mild traumatic brain injury in
Peer group training of pragmatic skills in adolescents with
children. Pediatrics 2001; 108: 1297–303.
acquired brain injury. J Head Trauma Rehabil 1998; 13: 23–
39. Suzman KB, Morris RD, Morris MK, Milan MA. Cognitive-
38.
behavioral remediation of problem solving deficits in children
55. Theunissen NCM, Vogels TGC, Koopman MH, et al. The
with acquired brain injury. J Behav Ther Exp Psychiatry 1997;
proxy problem: child report versus parent report in health-
28: 203–12.
related quality of life research. Qual Life Res 1998; 7: 387– 97.
700 Developmental Medicine & Child Neurology 2011, 53: 692–701
APPENDIX 1 Methodological quality appraisals of papers. Score 1 if met, 0 if not met or unable to determine Were specific hypotheses and ⁄ or objectives stated? Were the settings and locations where data was collected stated? Control or comparison group used? Were participants randomly allocated to groups? Is the method of randomization appropriate? Was the total sample size >20 participants? Was the total sample size >40 participants? Were at least some of the measures standardized assessment tools? Were the measures appropriate for age group? Were the inclusion ⁄ exclusion criteria clearly stated? Did the article specify the severity of the brain injury for participants with acquired brain injury and was the method of diagnosis appropriate (e.g. by a medical professional, Glasgow Coma Scale)? Did the injury occur at least 6mo ago (to ensure the results were not a reflection of the recovery process)? Were follow-up data collected after post-intervention data (i.e. to see if effects were maintained post intervention)? Were all participants included in the analysis? If not, was intent-to-treat analysis used? (Award 1 point if a point is granted on the above item) Were those assessing the outcomes blind to the group? Was a power calculation used or sample size justified? Was the intervention described in detail (i.e. how it was administered, etc.) or was there reference to a manual? Were the characteristics of participants clearly described (e.g. demographic information such as age, sex)? Did the results relate to the initial hypotheses? Was statistical analysis appropriate? Were data adequately described (mean, range etc.)? Were effect sizes calculated? Were effect sizes moderate or better (for studies with small sample sizes n
