DEVELOPMENTAL MEDICINE & CHILD NEUROLOGY
COMMENTARIES
Best practice when using the Strengths and Difficulties Questionnaire with extremely preterm children: are two informants better than one? KARLI TREYVAUD 1,2 1 Murdoch Childrens Research Institute, Melbourne, Vic.; 2 Department of Paediatrics, University of Melbourne, Melbourne, Vic., Australia. doi: 10.1111/dmcn.12448 This commentary is on the original article by Johnson et al. on pages 453–459 of this issue.
Most mental health practitioners would argue that it is important to acquire information on clients, especially children, from multiple informants. The article by Johnson et al.1 assessed the diagnostic accuracy of the Strengths and Difficulties Questionnaire (SDQ), a brief validated behavioural measure, when completed by parents and teachers of school-aged extremely preterm children. Preterm children have higher rates of social-emotional difficulties and psychological disorders compared with children born at term,2 and it is important that these high-risk children are identified for formal assessment and treatment. Johnson et al. revealed that when used as a clinical screening tool, an abnormal rating on the SDQ by either a parent or teacher (combined rating) gave the most accurate prediction of emotional disorders, conduct disorders, and attentiondeficit–hyperactivity disorder (ADHD). In contrast, when wanting to obtain the most accurate prediction of the prevalence of psychological disorders (such as when used as an outcome in research), the use of pervasive ratings was best, i.e. when both parent and teacher provided an abnormal rating. As one might expect, based on symptom criteria needed to assign certain psychiatric diagnoses and as highlighted by Johnson et al. in the study, there were differences in which approach (combined or pervasive ratings) worked best for different disorders. For example, pervasive ratings were better predictors of ADHD, while parent ratings were better predictive of autism spectrum disorder (ASD). This suggests that when ASD is the outcome under investigation, additional non-parental ratings may not necessary add value. Knowing this, if we continue to use only parental ratings of child psychological problems with the SDQ, we need to be aware that the validity of our results may be reduced. Furthermore, there are other potential biases to consider, such as research suggesting that parents tend to overestimate psychological problems in their children compared
with teachers,3 a pattern also reported by Johnson et al. Parental report of psychological problems in children appears to also be influenced by the mental health of the parents themselves, with several studies demonstrating an over-estimation of child psychological difficulties by parents with a history of mental health problems.4 This may be particularly relevant for parents of high-risk children such as those born preterm, who report higher levels of psychological distress in the early and school-age years,5 and should be considered when interpreting results. However, other research suggests that mothers with a history of anxiety or depression may actually be more accurate at perceiving internalizing symptoms in their children.3 Thus while parental mental health difficulties do not necessarily limit the validity of parental perceptions of their child’s social-emotional development, given the increased risk of mental health problems in parents of preterm children, the use of the use of multiple informants to assess child psychological problems remains important. Obtaining multi-informant SDQ ratings provides the clinician and researcher with the data needed to be as specific and accurate as possible when estimating psychological disorder prevalence or screening for psychological disorders in extremely preterm children. While many of us might have suspected that indeed two informants are better than one in this context, there are challenges associated with collecting non-parental ratings of children’s socialemotional development. These include generally lower response rates from non-parents (as seen in the Johnson et al. study), requirements to obtain ethical research approval from multiple sources (e.g. individual state education departments), and the lack of non-parental raters for preschool children who are not in formalized care outside of the home. It seems that if we want to aim for best practice with the SDQ, the challenge for researchers and clinicians is to find ways around these potential barriers. In summary, the study by Johnson et al. has implications for the use of the SDQ with preterm children in clinical and research contexts. Not only does this study highlight the importance of using more than one informant when assessing child social-emotional development, it reminds us to think critically about how and for what purpose we use the popular and psychometrically sound SDQ.
REFERENCES 1. Johnson S, Hollis C, Marlow N, Simms V, Wolke D. Screening for childhood mental health disorders using the
multi-informant reports. Dev Med Child Neurol 2014; 56: 453–59.
2. Johnson S, Marlow N. Preterm birth and childhood psychiatric disorders. Pediatr Res 2011; 69: 11R–8R.
Strengths and Difficulties Questionnaire: the validity of
© 2014 Mac Keith Press
413
3. Bora S, Pritchard VE, Moor S, Austin NC, Woodward
4. Chilcoat HD, Breslau N. Does psychiatric history bias
5. Treyvaud K. Parent and family outcomes following very
LJ. Emotional and behavioural adjustment of children
mothers’ reports? An application of a new analytic
preterm or very low birth weight birth: a review. Semin
born very preterm at early school age. J Paediatr Child
approach J Am Acad Child Adolesc Psychiatry 1997; 36:
Fetal Neonatal Med 2013; doi:10.1016/j.siny.2013.10.008.
Health 2011; 47: 863–9.
971–9.
[Epub ahead of print]
Looking for predictors of functional outcomes in children with cerebellar abnormalities PATRICE JISSENDI-TCHOFO Department of Radiology, Pediatric Neuroradiology Section, CHU Saint-Pierre, Brussels, Belgium. doi: 10.1111/dmcn.12446 This commentary is on the original article by Al-Maawali et al. on pages 460–467 of this issue.
The role of the cerebellum in cognitive, emotional, and behavioral functions is now well established as evidence based on the results of numerous studies, starting with the description by Schmahmann and Sherman of the cerebellar cognitive affective syndrome.1 This assumption was supported by cognitive and behavioral disturbances found in patients with cerebellar lesions. Further, the demonstration of the anatomical and functional links between the cerebrum and the cerebellum, thanks to diffusion tensor imaging and functional magnetic resonance imaging (fMRI) studies, has helped clinicians to better understand the cerebro-cerebellar connectivity. Since then, it has been challenging to find and understand the correlations between cerebellar congenital or acquired lesions and clinical signs, because of the variety of cerebellar involvements on the one hand, and the heterogeneity of clinical presentations on the other. However, it was assumed that the categorization of cerebellar disturbances in relation to specific morphological abnormalities could help in predicting the outcomes in children. Consequently, several studies focused on populations of children showing similar cerebellar malformations or injury. Neuropsychological features were thus described in children with defined cerebellar malformations such as agenesis, hypoplasia, dysplasia, and Joubert syndrome.2–4 In these studies, the methods used to assess the parameters of cognitive functions varied according to the study populations, dealing with different scales of age-related tests, also adapted to the language of the children and the school system of the country. The intellectual measurement tools commonly included the Diagnostic and Statistical Manual of Mental Disorders, 4th edition; the Wechsler Intelligence Scale for Children, 3rd edition; and the Wechsler Preschool and Primary Scale of Intelligence, 3rd edition. Whereas social and behav414 Developmental Medicine & Child Neurology 2014, 56: 413–419
ioural abilities were assessed using the Vineland Adaptive Behavior Scales, and other specific scales (McCarthy, Mullen, Peabody, Child Behavior Checklist, Pediatric Quality of Life Inventory, etc.). However, the validity of the results remains debatable for children too young or severely affected to undergo the classic psychometric tests. As mentioned above, the imaging features of cerebellar abnormalities turned out to be poor predictors of neurodevelopmental outcomes because of the lack of clear and significant correlations. When looking forward to predicting the functional outcomes in children with cerebellar abnormalities, the impact of these lesions on the daily living skills of children at a given age and over a certain period of life should be considered. Interestingly, the study by Al-Maawali et al.5 provides additional information in that field with a focus on cerebellar atrophy of childhood onset. The study was carried out in a prospective fashion, enrolling 44 patients over a period of 5 years, using the Activities of Daily Living score system to assess speech and swallowing, cutting food and dressing, walking and frequency of falls, bladder function and hygiene, as well as ability to sit. They found that larger size of the transverse cerebellar diameter was significantly correlated with better functional outcome. They also found that children with associated supratentorial anomalies were more affected than those with isolated cerebellar atrophy, and that motor signs such as ability to sit and frequency of falls were the most remarkable. Beyond these reported findings, the study provides additional parameters both on the side of imaging and psychomotor evaluation. The MRI two-dimensional measurements are easy to perform in daily clinical imaging, although less precise than surface assessment of fissure enlargement, which needs some specific image post-processing. Nevertheless, applying these measurements on larger cohorts of children with cerebellar abnormalities, including malformations as well as acquired injury, may help in approaching the regional anatomical modifications which can be further correlated with developmental and daily living disabilities, and thus serve as predictors of functional outcomes in those children. The ability to predict these outcomes will certainly lead to better management of these conditions by using supportive motor and cognitive therapies.
COMMENTARIES
Best practice when using the Strengths and Difficulties Questionnaire with extremely preterm children: are two informants better than one? KARLI TREYVAUD 1,2 1 Murdoch Childrens Research Institute, Melbourne, Vic.; 2 Department of Paediatrics, University of Melbourne, Melbourne, Vic., Australia. doi: 10.1111/dmcn.12448 This commentary is on the original article by Johnson et al. on pages 453–459 of this issue.
Most mental health practitioners would argue that it is important to acquire information on clients, especially children, from multiple informants. The article by Johnson et al.1 assessed the diagnostic accuracy of the Strengths and Difficulties Questionnaire (SDQ), a brief validated behavioural measure, when completed by parents and teachers of school-aged extremely preterm children. Preterm children have higher rates of social-emotional difficulties and psychological disorders compared with children born at term,2 and it is important that these high-risk children are identified for formal assessment and treatment. Johnson et al. revealed that when used as a clinical screening tool, an abnormal rating on the SDQ by either a parent or teacher (combined rating) gave the most accurate prediction of emotional disorders, conduct disorders, and attentiondeficit–hyperactivity disorder (ADHD). In contrast, when wanting to obtain the most accurate prediction of the prevalence of psychological disorders (such as when used as an outcome in research), the use of pervasive ratings was best, i.e. when both parent and teacher provided an abnormal rating. As one might expect, based on symptom criteria needed to assign certain psychiatric diagnoses and as highlighted by Johnson et al. in the study, there were differences in which approach (combined or pervasive ratings) worked best for different disorders. For example, pervasive ratings were better predictors of ADHD, while parent ratings were better predictive of autism spectrum disorder (ASD). This suggests that when ASD is the outcome under investigation, additional non-parental ratings may not necessary add value. Knowing this, if we continue to use only parental ratings of child psychological problems with the SDQ, we need to be aware that the validity of our results may be reduced. Furthermore, there are other potential biases to consider, such as research suggesting that parents tend to overestimate psychological problems in their children compared
with teachers,3 a pattern also reported by Johnson et al. Parental report of psychological problems in children appears to also be influenced by the mental health of the parents themselves, with several studies demonstrating an over-estimation of child psychological difficulties by parents with a history of mental health problems.4 This may be particularly relevant for parents of high-risk children such as those born preterm, who report higher levels of psychological distress in the early and school-age years,5 and should be considered when interpreting results. However, other research suggests that mothers with a history of anxiety or depression may actually be more accurate at perceiving internalizing symptoms in their children.3 Thus while parental mental health difficulties do not necessarily limit the validity of parental perceptions of their child’s social-emotional development, given the increased risk of mental health problems in parents of preterm children, the use of the use of multiple informants to assess child psychological problems remains important. Obtaining multi-informant SDQ ratings provides the clinician and researcher with the data needed to be as specific and accurate as possible when estimating psychological disorder prevalence or screening for psychological disorders in extremely preterm children. While many of us might have suspected that indeed two informants are better than one in this context, there are challenges associated with collecting non-parental ratings of children’s socialemotional development. These include generally lower response rates from non-parents (as seen in the Johnson et al. study), requirements to obtain ethical research approval from multiple sources (e.g. individual state education departments), and the lack of non-parental raters for preschool children who are not in formalized care outside of the home. It seems that if we want to aim for best practice with the SDQ, the challenge for researchers and clinicians is to find ways around these potential barriers. In summary, the study by Johnson et al. has implications for the use of the SDQ with preterm children in clinical and research contexts. Not only does this study highlight the importance of using more than one informant when assessing child social-emotional development, it reminds us to think critically about how and for what purpose we use the popular and psychometrically sound SDQ.
REFERENCES 1. Johnson S, Hollis C, Marlow N, Simms V, Wolke D. Screening for childhood mental health disorders using the
multi-informant reports. Dev Med Child Neurol 2014; 56: 453–59.
2. Johnson S, Marlow N. Preterm birth and childhood psychiatric disorders. Pediatr Res 2011; 69: 11R–8R.
Strengths and Difficulties Questionnaire: the validity of
© 2014 Mac Keith Press
413
3. Bora S, Pritchard VE, Moor S, Austin NC, Woodward
4. Chilcoat HD, Breslau N. Does psychiatric history bias
5. Treyvaud K. Parent and family outcomes following very
LJ. Emotional and behavioural adjustment of children
mothers’ reports? An application of a new analytic
preterm or very low birth weight birth: a review. Semin
born very preterm at early school age. J Paediatr Child
approach J Am Acad Child Adolesc Psychiatry 1997; 36:
Fetal Neonatal Med 2013; doi:10.1016/j.siny.2013.10.008.
Health 2011; 47: 863–9.
971–9.
[Epub ahead of print]
Looking for predictors of functional outcomes in children with cerebellar abnormalities PATRICE JISSENDI-TCHOFO Department of Radiology, Pediatric Neuroradiology Section, CHU Saint-Pierre, Brussels, Belgium. doi: 10.1111/dmcn.12446 This commentary is on the original article by Al-Maawali et al. on pages 460–467 of this issue.
The role of the cerebellum in cognitive, emotional, and behavioral functions is now well established as evidence based on the results of numerous studies, starting with the description by Schmahmann and Sherman of the cerebellar cognitive affective syndrome.1 This assumption was supported by cognitive and behavioral disturbances found in patients with cerebellar lesions. Further, the demonstration of the anatomical and functional links between the cerebrum and the cerebellum, thanks to diffusion tensor imaging and functional magnetic resonance imaging (fMRI) studies, has helped clinicians to better understand the cerebro-cerebellar connectivity. Since then, it has been challenging to find and understand the correlations between cerebellar congenital or acquired lesions and clinical signs, because of the variety of cerebellar involvements on the one hand, and the heterogeneity of clinical presentations on the other. However, it was assumed that the categorization of cerebellar disturbances in relation to specific morphological abnormalities could help in predicting the outcomes in children. Consequently, several studies focused on populations of children showing similar cerebellar malformations or injury. Neuropsychological features were thus described in children with defined cerebellar malformations such as agenesis, hypoplasia, dysplasia, and Joubert syndrome.2–4 In these studies, the methods used to assess the parameters of cognitive functions varied according to the study populations, dealing with different scales of age-related tests, also adapted to the language of the children and the school system of the country. The intellectual measurement tools commonly included the Diagnostic and Statistical Manual of Mental Disorders, 4th edition; the Wechsler Intelligence Scale for Children, 3rd edition; and the Wechsler Preschool and Primary Scale of Intelligence, 3rd edition. Whereas social and behav414 Developmental Medicine & Child Neurology 2014, 56: 413–419
ioural abilities were assessed using the Vineland Adaptive Behavior Scales, and other specific scales (McCarthy, Mullen, Peabody, Child Behavior Checklist, Pediatric Quality of Life Inventory, etc.). However, the validity of the results remains debatable for children too young or severely affected to undergo the classic psychometric tests. As mentioned above, the imaging features of cerebellar abnormalities turned out to be poor predictors of neurodevelopmental outcomes because of the lack of clear and significant correlations. When looking forward to predicting the functional outcomes in children with cerebellar abnormalities, the impact of these lesions on the daily living skills of children at a given age and over a certain period of life should be considered. Interestingly, the study by Al-Maawali et al.5 provides additional information in that field with a focus on cerebellar atrophy of childhood onset. The study was carried out in a prospective fashion, enrolling 44 patients over a period of 5 years, using the Activities of Daily Living score system to assess speech and swallowing, cutting food and dressing, walking and frequency of falls, bladder function and hygiene, as well as ability to sit. They found that larger size of the transverse cerebellar diameter was significantly correlated with better functional outcome. They also found that children with associated supratentorial anomalies were more affected than those with isolated cerebellar atrophy, and that motor signs such as ability to sit and frequency of falls were the most remarkable. Beyond these reported findings, the study provides additional parameters both on the side of imaging and psychomotor evaluation. The MRI two-dimensional measurements are easy to perform in daily clinical imaging, although less precise than surface assessment of fissure enlargement, which needs some specific image post-processing. Nevertheless, applying these measurements on larger cohorts of children with cerebellar abnormalities, including malformations as well as acquired injury, may help in approaching the regional anatomical modifications which can be further correlated with developmental and daily living disabilities, and thus serve as predictors of functional outcomes in those children. The ability to predict these outcomes will certainly lead to better management of these conditions by using supportive motor and cognitive therapies.
