In the present case-control study of primary school children, having a CI was associated with a higher risk of psychiatric disorder as assessed by the Kiddie-SADS-PL. The percentage of psychiatric disorders decreased and the general level of functioning increased as a function of higher FSIQ-level both in the CI-and the NCI-group. The protective effect of a normal to high FSIQ-level was not more substantial in children with CI, supporting an overall protective effect of normal to high intellectual function on children's mental health.
More than half of the children in the CI-group met the criteria of a psychiatric disorder, compared to a third of the children in the NCI-group. Thus, even when the two groups were matched on FSIQ-level, the overall percentage of psychiatric disorders was still significantly higher in children with CI. The estimated risk of psychiatric disorder in this case-control sample of children with CI is in accordance with the twofold increased risk of mental health problems in children with CI shown in a study of the whole population of the BCS . Although the risk is similar, the overall percentage of psychiatric disorders is higher, as expected due to the selection of participants to this stage of the BCS.
The present study showed that children with a FSIQ-level between 70 and 84 had a similar risk estimate of psychiatric disorders as children with a FSIQ-level below 70 - a risk that was significantly higher than for children with a FSIQ-level of 85 or above. This finding is in accordance with the results of Goodman and collaborators, showing that healthy children with low IQ within the normal range (defined as WISC-R FSIQ in the range 70-89) had more behavioural problems compared to those with higher IQ-scores . It is also consistent with the findings of Dekker and collaborators, showing that children with borderline intellectual disability (IQ-range 60-80) and those with moderate intellectual disability (IQ-range 30-60) had a similar rate and estimated risk of mental health problems that was significantly higher than for children with a higher level of intellectual function .
A protective effect of normal to high intellectual function was found both in the CI- and the NCI-group. Such an overall effect was contrary to what we expected from the stressors associated with CI and from the findings of Perrin et al. . The differences between the results in Perrin and collaborators' and the present study may partly be ascribed to methodological factors. First of all, Perrin et al. had the focus on children with an IQ-score above 80, as it was measured by an unstandardized test of intellectual function (the PPVT). Secondly, the measures of mental health, the recruitment procedures and characteristics of the samples are quite different in the two studies. In Perrin et al.'s study, the healthy children were recruited from public and private schools, while the children with CI were recruited through generalist and specialist pediatric offices. The children participating in the present study were part of the same case-control sample selected from the BCS-population, with the same percentage of screen positive and screen negative children in the CI- and the NCI-group. Furthermore, the two groups in our study were matched on FSIQ-level. Consequently, the CI- and the NCI-group in the present study were probably more similar on critical variables than the corresponding groups in Perrin and collaborators' study.
Strengths and limitations
The main strength of the study was the use of a standardized test of intellectual function (WISC-III) and a validated clinical interview generating DSM-IV diagnoses (Kiddie-SADS-PL). Moreover, the study sample was drawn from a population of children from the second largest city of Norway and included both screen positive and screen negative children. An additional strength was the use of a comparison group matched on FSIQ-level.
Some limitations should be mentioned. First of all, the use of categorical IQ measures reduced the statistical power of our analyses. The categorical levels were included due to our focus on children with an IQ-level within the normal range and higher, and it should be mentioned that an analysis of the full range of FSIQ-scores did not change the results concerning the impact of IQ on mental health problems. Secondly, the IQ-distribution of the CI-group was skewed. Only 3 children had an IQ-level above the normal range (>115), compared to 12 children in the NCI-group. This skewness probably reflects what the case is for children with CI as a group: compared to their peers, they have a higher frequency of general and specific learning disabilities, which in turn is associated with lower mean IQ . Finally, although the protective effect of normal to high IQ was not more substantial in children with CI in the present study, it is still an important protective factor in relation to risk of mental health problems in this group of children. However, IQ only explained some of the association between CI and mental health. In future studies we will include other factors considered important for the mental health of children with CI.
The present study showed that children with a normal to high FSIQ-level had better mental health than children with a very low and low FSIQ-level. The frequencies of psychiatric disorders were somewhat higher in the CI-group compared to the NCI-group within all three FSIQ-levels. Paediatricians and others working with children with CI should be aware of this increased risk of mental health problems and the need of psychological support not only for children with low IQs, but also for children with an IQ-score within the normal range of intellectual function.