Behavioural and emotional symptoms of preschool children with cerebral palsy: a population-based study
Dr Solveig Sigurdardottir at State Diagnostic and Counselling Centre, Digranesvegur 5, 200 Kopavogur, Iceland. E-mail: firstname.lastname@example.org
Aim To describe behavioural and emotional symptoms among Icelandic preschool children with cerebral palsy (CP).
Method Children with congenital CP, assessed with the Child Behavior Checklist/1½–5 (CBCL/1½–5) and Caregiver-Teacher Report Form (C-TRF), were enrolled in the study. A comparison group was recruited from the general population. Thirty-six children (53% males) with CP were assessed at a mean age of 4 years 11 months (SD 5mo, range 4–6y); 26 (72%) had bilateral distribution of symptoms and 32 (89%) had spastic CP. Thirty (83%) were at Gross Motor Function Classification System levels I or II and six at levels III or IV. For comparison, 110 (43% males) and 120 (48% males) children were assessed with the CBCL/1½–5 and the C-TRF respectively, at a mean age of 4 years 6 months (SD 6mo, range 4–6y).
Results Sixteen children (48%) with CP had high scores on total problems scale of the CBCL/1½–5 and 20 (65%) on the C-TRF compared with 18% of the comparison group, both on the CBCL/1½–5 and the C-TRF (p<0.001). Children with CP had higher scores on all subscales of the CBCL/1½–5 and the C-TRF, except somatic complaints. Attention difficulties, withdrawn, aggressive behaviour, and anxious/depressed symptoms were most pronounced among children with CP.
Interpretation A large proportion of preschool children with CP have substantial behavioural and emotional difficulties, which need to be addressed in their treatment.
List of abbreviations
Child Behavior Checklist/1½–5
Caregiver-Teacher Report Form
What this paper adds
- • Behavioural and emotional symptoms are frequent even in preschool children with CP.
- • These symptoms are evident both at home and at preschool.
- • Attention difficulties, withdrawn, aggressive behaviour, and anxious/depressed symptoms are especially pronounced.
- • Even children with mild motor difficulties have high rates of psychological symptoms.
Cerebral palsy (CP) is the most common motor impairment of childhood. It is caused by a lesion of the brain occurring early in development. CP frequently affects behaviour and emotional well-being of children, most likely as a result of complex interactions between biological factors in the brain and psychosocial surroundings.1
Prevalence of behavioural and emotional problems among children with CP ranges between 25% and 60%,2–4 reflecting differences in age of childhood populations, psychometric measures, and severity of CP. Children with CP have been shown to exhibit diverse externalizing and internalizing symptoms,2,3,5 consistent with the conclusion that brain injury does not result in a specific behaviour pattern.1 Motor impairment can affect behavioural expressions of children with CP, especially during the preschool years.6 Although symptoms may be evident at an early age, mental health disorders are less often diagnosed in preschool children than in older children.7
Assessment of psychological symptoms in early childhood relies heavily on caregivers’ descriptions obtained through checklists.4 Several researchers have described mental health of school-age populations with CP3,5,8 whereas reports on younger children are sparse and symptom consistency across situations has rarely been studied.2,7
Therefore the aim of this study was to describe behavioural and emotional symptoms of preschool children with CP observed by parents and by preschool teachers and compare them with symptoms from a comparison group at the same age.
The data used in this study were mainly abstracted from medical records at the central developmental centre in Iceland, the State Diagnostic and Counselling Centre, where all children with congenital CP in Iceland are comprehensively evaluated between 4 and 6 years of age.9 A part of the evaluation is assessment of mental health, which includes a clinical evaluation and assessment with standardized symptoms questionnaires, mainly the Child Behavior Checklist/1½–5 (CBCL/1½–5) and Caregiver-Teacher Report Form (C-TRF) of the Achenbach System of Empirically Based Assessment.10 The questionnaires are rated by parents (most often the mothers) and the child’s regular preschool carer/teacher. Children born between 1999 and 2004 and who by 15 October 2009 had been examined at the State Diagnostic and Counselling Centre were eligible for participation in this study.
The comparison group comprised children at the same age born between 2003 and 2005 recruited from the general population for an ongoing validation study of the CBCL/1½–5 and C-TRF forms. Data available on 15 October 2009 were used in this study.
A total of 52 children with congenital CP born between 1999 and 2004 were evaluated at the State Diagnostic and Counselling Centre between 4 and 6 years of age. Eight children with severe disability were not included in the study as their mental health assessment was confined to clinical assessment. Additional eight children were excluded; six were assessed with school-age versions of the CBCL and TRF forms and two children had mental health assessment without using the forms. Thus, the study group comprised 36 children (19 males [53%], 17 females [47%]), who were examined at a mean age of 4 years 11 months (SD 5mo). Parents assessed behaviour of 33 children, whereas 31 children were assessed by preschool teachers (28 children had both CBCL/1½–5 and C-TRF ratings whereas five had only CBCL/1½–5 ratings and three had only C-TRF ratings). The excluded children had comparable motor impairment as children in the study group, whereas a slightly smaller proportion of the excluded (25%) than the included children (39%) had an IQ<70.
A total of 525 preschool children, between the ages of 1.5 and 6 years, were randomly selected from ten preschools in Iceland for a validation study of the CBCL/1½–5 and C-TRF forms. For comparison in the present study, we only included data from children aged 4 to 6 years. Ratings by parents for 110 children (43% males) were available (mean age 4y and 5mo, SD 6mo), and ratings by preschool teachers for 120 children (48% males), mean age 4 years and 6 months (SD 6mo). As these data were without personal identities, we do not know the exact proportion of the CBCL/1½–5 and C-TRF ratings that were from the same children.
CP was defined as a group of disorders causing permanent impairment of voluntary movement or posture attributed to non-progressive disturbances occurring in the developing fetal or infant brain.11 Gross motor abilities were classified using the Gross Motor Function Classification System (GMFCS).12 Children with mild gross motor impairment (GMFCS levels I and II) were merged into one group whereas children with moderate to severe impairment (GMFCS levels III and IV) were merged into another group. None of the included children were at GMFCS level V.
The CBCL/1½–5 and C-TRF, designed for children aged 1½ to 5 years, were used to obtain standardized ratings of diverse aspects of behavioural, emotional, and social functioning of the children.10 The forms were translated and published in the Icelandic language in 2002 to 2004 but validation is ongoing. They include 99 problem items rated 0, 1, or 2 (0, not true; 1, somewhat/sometimes true; 2, very/often true). Many of the items are identical although the C-TRF substitutes items pertinent to group situations for items specific to family situations. Seven syndrome scales (six scales for C-TRF) are constituted: emotionally reactive, anxious/depressed, somatic complaints, withdrawn, sleep problems (only on CBCL/1½–5), attention problems, and aggressive behaviour. Three composite or summary scales are computed: internalizing scale (sum of emotionally reactive, anxious/depressed, somatic complaints, and withdrawn), externalizing scale (sum of attention problems and aggressive behaviour), and total problems scale (all syndrome scores).10 The numbers of items constituting each syndrome scale are listed in Table I; the minimum score on each scale is 0 whereas maximum scores are obtained by multiplying the number of items by 2.
Table I. Median scores with interquartile ranges (IR) on the CBCL/1½–5 and C-TRF for Icelandic preschool children with cerebral palsy compared with scores from an Icelandic comparison group
| Emotionally reactive (9)||2 (0–4)||3 (1–6)||0.031|
| Anxious/depressed (8)||1 (1–3)||3 (2–5)||0.003|
| Somatic complaints (11)||1 (0–2)||1 (1–3)||0.395|
| Withdrawn (8)||1 (0–2)||2 (1–3)||0.001|
| Sleep problems (7)||2 (1–3)||2 (1–5)||0.012|
| Attention problems (5)||0 (0–1)||3 (1–6)||<0.001|
| Aggressive behaviour (19)||6 (4–10)||9 (6–15)||0.012|
|Internalizing scale (36)||5 (3–10)||9 (5–16)||0.005|
|Externalizing scale (24)||7 (4–11)||12 (7–22)||0.001|
|Total problems scale (99)||20 (11–32)||34 (19–58)||<0.001|
|C-TRF (number of items)||(n=120)||(n=31)|| |
| Emotionally reactive(7)||0 (0–1)||1 (0–3)||0.001|
| Anxious/depressed (8)||0 (0–1)||2 (1–3)||<0.001|
| Somatic complaints (7)||0 (0–0)||0 (0–0)||0.895|
| Withdrawn (10)||0 (0–1)||3 (1–5)||<0.001|
| Attention problems (9)||0 (0–2)||7 (1–9)||<0.001|
| Aggressive behaviour (25)||0 (0–3)||6 (1–10)||<0.001|
|Internalizing scale (32)||1 (0–3)||6 (2–10)||<0.001|
|Externalizing scale (34)||1 (0–5)||12 (4–19)||<0.001|
|Total problems scale (99)||4 (1–11)||24 (10–43)||<0.001|
According to the reference manual, cutpoints at the 84th and 90th centiles of the internalizing, externalizing, and total problems scales are used to define scores in the borderline clinical and clinical range respectively.10 In the manual it is recommended including the borderline category to improve the basis for decisions about children’s need for help. Therefore, we combined the borderline clinical range with the clinical range and defined high scores on the summary scales as scores equal to or above the 84th centile of the comparison group.10
Other variables (only assessed among children with CP)
Gestational age was classified into term (≥37wks) and preterm (<37wks). Cognitive function was objectively assessed in all children by clinical psychologists.9 The Wechsler Preschool and Primary Scale of Intelligence13 was administered and IQ obtained. For one child a developmental quotient was calculated9 from the Bayley Scales of Infant Development.14 Children with IQ or developmental quotient below 70 were classified as having intellectual impairment. Other associated impairments, i.e. hearing impairment (deafness or hearing threshold at or above 65dB in better ear), visual impairment (functional blindness or visual acuity of 0.3 or less in better eye with correction), and epilepsy (child using antiepileptic medication at the time of assessment), were documented. Further, previous history of maladaptive behaviours or psychiatric diagnoses was documented.
The Data Protection Authority and the National Bioethics Committee in Iceland approved the study as part of a larger study entitled ‘Icelandic children with CP born 1990–2006’ (VSNb2008020038/03.1). The Committee approved the use of clinical information on children, including characteristics of CP and associated deficits, routinely collected at the State Diagnostic and Counselling Centre. Specifically, omission of informed consent was approved by the Bioethics Committee. The study, in which the comparison group originated, was registered by the Data Protection Authority. Parents accepted participation by completing forms for their children and by signing a consent allowing preschool teachers to complete C-TRF for their children.
All data were analysed using SPSS software, version 16.0 (SPSS Inc., Chicago, IL, USA). Median values with interquartile ranges were calculated as the data were not normally distributed. A non-parametric statistical method, the Mann–Whitney U test, was applied for two-group comparisons. The χ2 test was used to analyse differences in proportions between most groups whereas Fisher’s exact test was used in cases when more than one cell in a cross tabulation had fewer than five observations.15 Odds ratios were calculated with 95% confidence intervals (CI) as estimates of the relative risk for high symptom scores among children with CP compared with the comparison group.15 Stratification was used to control for possible confounders, i.e. IQ and gross motor impairment. A two-tailed p value below 0.05 was considered statistically significant.
Among the 36 children with CP, 26 (72%) had bilateral motor impairment and 32 (89%) had spastic CP. Thirty (83%) could walk unaided or with mild support (GMFCS levels I or II) whereas six (17%) children had severe motor impairment (GMFCS levels III or IV). Twenty-three (64%) children were born preterm, 14 (39%) had intellectual impairment (IQ<70), five were visually impaired, one had hearing impairment, and two had epilepsy (one child had both visual impairment and epilepsy). Three children had autism spectrum disorder and two had attention-deficit–hyperactivity disorder.
Children with CP had higher median scores than the comparison group, as judged by parents on all scales of CBCL/1½–5 except somatic complaints (Table I). Differences between groups were pronounced (p≤0.001) for the attention problems, withdrawn, externalizing, and total problems scales. Ratings captured by preschool teachers also showed the same pattern of higher median scores (p≤0.001) for the group with CP on all scales except somatic complaints. Differences between the group with CP and the comparison group were overall more pronounced when rated by preschool teachers than by parents.
In the subgroup of children with IQ≥70, median scores did not differ between those with CP and the comparison group when assessed by parents, except for attention problems where children with CP had higher scores (p<0.001; Table II). However, in the same subgroup, C-TRF ratings revealed higher scores for children with CP on all scales except emotionally reactive and somatic complaints. Children with CP and intellectual impairment (IQ<70) had higher median scores (p<0.01) than the comparison group on all scales except for somatic complaints, as judged by parents and preschool teachers (Table II).
Table II. Median scores with interquartile ranges (IR) on CBCL/1½–5 and C-TRF for Icelandic preschool children with cerebral palsy compared with scores from an Icelandic comparison group; subgroups with IQ≥70 and IQ<70 studied separately
| Emotionally reactive||2 (0–4)||2 (1–5)||5 (2–10)||0.607||0.002|
| Anxious/depressed||1 (1–3)||2 (1–4)||5 (3–6)||0.241||<0.001|
| Somatic complaints||1 (0–2)||1 (0–2)||2 (1–3)||0.971||0.132|
| Withdrawn||1 (0–2)||1 (0–2)||3 (3–5)||0.443||<0.001|
| Sleep problems||2 (1–3)||2 (1–4)||4 (3–6)||0.298||0.003|
| Attention problems||0 (0–1)||2 (1–3)||6 (4–8)||<0.001||<0.001|
| Aggressive behaviour||6 (4–10)||7 (4–10)||18 (13–22)||0.897||<0.001|
|Internalizing scale||5 (3–10)||7 (4–11)||14 (11–23)||0.445||<0.001|
|Externalizing scale||7 (4–11)||9 (5–13)||24 (17–29)||0.299||<0.001|
|Total problems scale||20 (11–32)||25 (16–33)||56 (47–74)||0.162||<0.001|
|C-TRF||(n=120)||(n=20)||(n=11)|| || |
| Emotionally reactive||0 (0–1)||1 (0–1)||3 (1–3)||0.149||<0.001|
| Anxious/depressed||0 (0–1)||1 (0–3)||3 (2–5)||0.045||<0.001|
| Somatic complaints||0 (0–0)||0 (0–0)||0 (0–0)||0.729||0.481|
| Withdrawn||0 (0–1)||3 (0–5)||4 (2–6)||0.001||<0.001|
| Attention problems||0 (0–2)||3 (0–7)||8 (8–16)||<0.001||<0.001|
| Aggressive behaviour||0 (0–3)||2 (0–9)||13 (6–19)||0.014||<0.001|
|Internalizing scale||1 (0–3)||4 (1–8)||9 (7–16)||0.002||<0.001|
|Externalizing scale||1 (0–5)||7 (2–16)||22 (14–36)||0.001||<0.001|
|Total problems scale||4 (1–11)||13 (6–32)||43 (31–59)||0.001||<0.001|
Table III shows that larger proportions of children with CP had high scores on internalizing, externalizing, and total problems scales than the comparison group. Almost half of the study group (48%) had high scores on the total problems scale of CBCL/1½–5 compared with 18% of the comparison group (p<0.001). The differences were even more pronounced when C-TRF ratings were compared: 65% of study group had high scores on total problems scale compared with 18% of the comparison group (p<0.001). Table III also shows the odds ratio for high scores on the summary scales for children with CP compared with those in the comparison group. Odds for high scores on the total problems scale was 4.2 (CI 1.7–10.7) when judged by parents and 8.1 (CI 3.1–21.3) when rated by preschool teachers. When a cut-off at the 90th centile was applied, odds for high scores on the summary scales for children with CP compared with the comparison group were either unchanged (internalizing scale rated by parents) or even higher (data not shown).
Table III. Prevalence and odds ratio (OR) with 95% confidence intervals (CI) of high scores (i.e. ≥84th centile) on summary scales of CBCL/1½–5 and C-TRF in preschool children with cerebral palsy compared with comparison group
| Study group (n=33)||20 (61)||13 (39)||0.007||3.1 (1.2–8.0)||20 (61)||13 (39)||0.003||3.6 (1.4–9.3)||17 (52)||16 (48)||<0.001||4.2 (1.7–10.7)|
| Comparison group (n=110)||91 (83)||19 (17)||93 (84)||17 (15)||90 (82)||20 (18)|
| Study group (n=31)||13 (42)||18 (58)||<0.001||5.5 (2.2–14.0)||12 (39)||19 (61)||<0.001||8.4 (3.2–22.2)||11 (35)||20 (65)||<0.001||8.1 (3.1–21.3)|
| Comparison group (n=120)||96 (80)||24 (20)||101 (84)||19 (16)||98 (82)||22 (18)|
Table IV shows that for the subgroup of children with CP and IQ≥70, the odds for high scores on the summary scales were not increased when assessed by their parents whereas the odds varied between 2.7 (CI 1.0–7.3) and 3.6 (CI 1.3–9.9) on assessment by preschool teachers. For children with CP and IQ<70, the odds for high scores on the summary scales were considerably increased (p<0.001) compared with the comparison group both when assessed by parents and preschool teachers (data not shown). For the subgroup of children with CP and mild gross motor impairment, the odds for high scores varied between 2.7 (CI 1.1–6.7) and 3.9 (CI 1.6–9.5) as judged by parents, and between 5.5 (CI 2.2–13.4) and 7.2 (CI 2.9–18.2) when rated by preschool teachers (Table IV). For children with severe gross motor impairment (GMFCS III or IV), the odds for high scores on summary scales were increased both at home and at preschool (data not shown).
Table IV. Prevalence and Odds ratio (OR) with 95% confidence intervals (CI) for high scores on summary scales of CBCL/1½–5 and C-TRF among two subgroups of children with cerebral palsy compared with comparison group
| Study group; IQ≥70 (n=20)||16 (80)||4 (20)||0.769||1.2 (0.4–4.0)||17 (85)||3 (15)||0.959||1.0 (0.3–3.7)||16 (80)||4 (20)||0.847||1.1 (0.3–3.7)|
| Comparison group (n=110)||91 (83)||19 (17)||93 (85)||17 (15)||90 (82)||20 (18)|
| Study group; GMFCS levels I or II (n=28)||18 (64)||10 (36)||0.032||2.7 (1.1–6.7)||17 (61)||11 (39)||0.005||3.5 (1.4–8.9)||15 (54)||13 (46)||0.002||3.9 (1.6–9.5)|
| Comparison group (n=110)||91 (83)||19 (17)||93 (85)||17 (15)||90 (82)||20 (18)|
| Study group; IQ≥70 (n=20)||12 (60)||8 (40)||0.049||2.7 (1.0–7.3)||12 (60)||8 (40)||0.011||3.5 (1.3–.8)||11 (55)||9 (45)||0.008||3.6 (1.3–9.9)|
| Comparison group (n=120)||96 (80)||24 (20)||101 (84)||19 (16)||98 (82)||22 (18)|
| Study group; GMFCS levels I or II (n=26)||11 (42)||15 (58)||<0.001||5.5 (2.2–13.4)||11 (42)||15 (58)||<0.001||7.2 (2.9–18.2)||10 (38)||16 (62)||<0.001||7.1 (2.9–17.8)|
| Comparison group (n=120)||96 (80)||24 (20)||101 (84)||19 (16)||98 (82)||22 (18)|
In this study we found that 40 to 50% of preschool children with CP had substantial behavioural or emotional difficulties when judged by parents, whereas this applied to 60 to 65% when assessed by preschool teachers. Preschool children with CP were three to four times more likely to show considerable psychological symptoms when rated by parents than the comparison group, and the risk was even higher when rated by preschool teachers. Except for the subscale somatic complaints, children with CP had higher scores on all scales of the CBCL/1½–5 and C-TRF forms than the comparison group.
The differences in behavioural and emotional symptoms of children with CP compared with the comparison group were marked, as indicated by the low p values, making it unlikely that our findings occurred by chance.
The strength of the current study is that data on children with CP were systematically collected at a national developmental centre over 6 years by a few clinicians, allowing for stability in procedures. Another advantage is that behavioural ratings were obtained from two informants, parents and preschool teachers. Case ascertainment was high although a few severely disabled children were excluded on clinical grounds and because questionnaires have not been found to be suitable for children with the most severe disabilities.3,8,16
A limitation of the study is the relatively small sample size, which limits statistical power to detect (sub)group differences. Thus lack of statistically significant differences should be interpreted with caution. Caution is also needed regarding the estimates of prevalence of high scores, as the CIs are large. Lack of detailed information on children in the comparison group may further limit the interpretation of findings. Finally, the cross-sectional design of the study limits any causal or directional conclusions.
Higher scores among children with CP than healthy children could possibly be explained by closer observation of behavioural and emotional symptoms of those with CP as they often need assistance and support in daily life. However, studies including diagnostic assessment of mental health in children with hemiplegia4 and those of very low birthweight17 have indicated that questionnaires can give reliable information about the children’s mental health status, thus making information bias less likely. Adverse parental functioning could be a confounder in this study,3,4,8 but as ratings from preschool teachers were consistent with parental ratings, this is unlikely to explain our results.
Our finding of high odds for scores in the borderline or clinical range (≥84th centile) among children with CP may indicate a causal relation. When a cut-off at the 90th centile was applied, the results were essentially unchanged, further supporting the clinical significance of the findings. Our results are partly consistent with the Isle of Wight study,18 which identified psychiatric disorder in 44% of school-age children with CP and IQ≥50, and a study of 4- to 7-year-old children, which found parent-reported behavioural problems five times more often in children with CP than healthy children.2 Furthermore, a recent study on school-age children with CP found over 40% of the group to have scores in the borderline to abnormal range.3 Thus the new finding in this study is that behavioural and emotional problems are common already in preschool children with CP, when rated both by parents and preschool teachers. In comparison, a study of children with hemiplegia revealed high prevalence of psychological symptoms in those younger than 5; however, in contrast to our results the rate of significant symptoms was higher when the questionnaires were rated by parents (51%) than by preschool professionals (35%).4
Attention problems and being withdrawn were the most problematic symptoms for the group with CP both at home and at preschool, whereas in the preschool environment, aggressive behaviour and anxious/depressed symptoms were also increased. Previous studies based on caregivers’ descriptions obtained through checklists have reported similar findings,2,3,8 and a recent study using neuropsychological measures identified impairments in attention and executive functions among school-age children with CP.19 The somatic complaints subscale is not thought to be a sensitive marker for internalizing symptoms among children with chronic disorders,20,21 which might explain why we did not observe differences in ratings of somatic complaints between children with CP and the comparison population.
Children with mild motor impairment (GMFCS levels I or II) had high rates of problem behaviour across situations as did children with IQ<70. The latter finding is consistent with other studies reporting an association between lower IQ and more psychiatric problems.4,22 In this respect, IQ is thought to be a marker for underlying neurobiological factors that influence psychopathology.4 In contrast, only 15 to 20% of children with CP without intellectual impairment (IQ≥70) had high scores on parental assessment, whereas this applied to 40 to 45% of the group when judged by preschool teachers. We speculate if this finding might indicate that psychosocial surroundings can affect behavioural expressions of children, as the preschool environment appeared to bring out mental health weaknesses of the subgroup of children with CP but without intellectual impairment. The motor impairment itself restricts activities and play of young children, especially within the busy preschool environment, and can lead to frustration and social withdrawal. This may at least partly explain the higher prevalence of maladaptive behaviours within preschool settings than at home.
Although upon assessment large proportions of children with CP exhibit impairing behavioural and emotional symptoms, their problems frequently go unrecognized and untreated.3,4 Therefore, therapeutic programmes need to monitor psychological and social well-being of young children with CP as early identification and intervention may prevent chronic maladjustment.23
Our results imply that preschool children with CP are at high risk for behavioural and emotional problems. Even children with CP and mild motor impairment encounter such problems both at home and at preschool. Attention difficulties, withdrawn, aggressive behaviour, and anxious/depressed symptoms are the main challenges faced by these children. We conclude that treatment programmes of preschool children with CP should take into account their psychological functioning.
The study was funded by the Liaison Committee for Central Norway Regional Health Authority and the Norwegian University of Science and Technology.