Discussion
Our findings suggest that the mother's migration background is associated with an increased risk of ASD in preschool children, as well as more severe symptoms and a younger age at diagnosis. The findings suggest that the mother's migration background may influence the development of ASD. Our findings also imply a need to tailor the investigation to this patient group and highlight the need to clarify information and facilitate supportive services.
One of the strengths of our study is that we have identified all children diagnosed within the autism spectrum at St Olav's Hospital in the given age group and time period. Only 8 % of those invited to participate declined or had moved, and therefore could not be included. The Norwegian health service and the close cooperation between the health services and educational institutions (the Educational and Psychological Counselling Service/kindergartens) help ensure that children with a suspected neurodevelopmental disorder are identified and referred, regardless of socioeconomic status or family background. Hence, our dataset is unique and is hardly affected by selection bias.
Limitations to our study include limited information on the mother's country of origin as this was acquired through the child's hospital records. It was also difficult to acquire reliable information on language skills and cognitive function due to the child's young age and difficulties with social interaction associated with this disorder. We therefore have limited information on how many of the children also had a concurrent intellectual disability. We also have limited information on the child's socioeconomic status and we lack genetic information.
Previous studies support our findings of an increased risk for ASD in children of migrant mothers (9–11, 22) . Several studies investigate the prevalence of ASD in this group (23, 24) while we investigated the incidence. Nevertheless, we do not expect this to affect results as our study population is of a young age. Several genetic variants can predispose a child to ASD (25) . However, studies have shown that environmental factors both before and after birth also play a role (16, 26) . Further research is therefore needed on how these can contribute to the risk of developing autism (16) .
Medical records showed that 67 children (86 % of the group with a migrant background) had two parents with a migrant background. However, not all medical records included information on the fathers' country of birth. Only one child had a Norwegian-born mother and a foreign father. A Swedish study (10) found that the mother's migration background increased the risk for ASD independent of the migration background of the father. A Finnish study (11) found no increased risk of ASD among children where only the father had a migrant background.
We found a higher mean ADOS score in children of migrant mothers compared with children of Norwegian-born mothers. This group was also younger at the time of diagnosis. A plausible explanation could be that these children were identified and examined at an earlier age because they had more severe symptoms. An Australian study (12) found that children of mothers who migrated from low-income countries were younger at the time of diagnosis and had an increased risk of intellectual disability. Our findings may indicate greater severity of the disorder in children of migrant mothers. The association between higher ADOS scores and early age of diagnosis was shown in both groups. This indicates that young children with clear signs of developmental disorder are identified and evaluated early, regardless of the mother's country of origin. However, we cannot rule out the possibility that some of the variance in ADOS scores found in this study could be attributed to reduced validity of the ADOS scoring as a diagnostic tool when the child and healthcare professional do not speak the same language or share the same cultural background.
Children of migrant mothers were overrepresented among preschool children with ASD, which may be an indication that migration in this age group has aetiological significance. A review of patient records of children diagnosed with ASD in a geographical region of the United States in the period 2004–14 (27) found that children with verbal language skills and comorbidities such as ADHD were often diagnosed at an older age. The study also found that children raised by migrant mothers in multilingual homes were younger at age of diagnosis than children of mothers without a migrant background.
Symptoms of Asperger syndrome become increasingly apparent in older children, while delayed language development often contributes to early referral (28) . Our dataset has few children with Asperger syndrome as only preschool children were included. Several studies have shown a reduced risk of Asperger syndrome (10, 29) and of high-functioning autism (14) in children of migrant mothers. This supports the hypothesis that there is a distinction between ASDs in these two age groups, and they should therefore be studied separately. In most studies, inclusion criteria are based on the child's year of birth (11, 15, 23, 24, 29) or when the child was examined (13, 22, 23) . Comparing individuals with ASD diagnosed at different times could reveal important information with regard to aetiological factors.
A qualitative study from England of parents with a Somali background who have children with autism (30) highlighted the importance of having a cultural understanding of this disorder. This study illustrates the need to improve the families' understanding of ASD, to provide families with support and to encourage them to seek help at an early stage. It is likely that the challenges described in the study are generalisable to our patients.
The large proportion of children with ASD of migrant mothers in Sør-Trøndelag (56 % of the study population) underlines the importance of adapting the evaluations and follow-up to children and families with a different native tongue and cultural background. Such measures include interpreting services and the translation of information to the child's mother tongue.
Conclusion and implications
Conclusion and implications
This study supports the clinical impression that ASD is overrepresented among children of migrant mothers. The incidence of ASD was 7.7 times higher in children of migrant mothers than children of Norwegian-born mothers. Our findings also suggest that children with ASD of migrant mothers are younger at the time of diagnosis and have more severe symptoms.
This reinforces the need for further studies on the role of maternal migration in the development of ASD. Healthcare professionals should be aware of how a different mother tongue and cultural background impacts the evaluation and follow-up of children with ASD and their families. The rapidly increasing number of children diagnosed with ASD and concurrent overrepresentation of families with a migration background puts further pressure on evaluation and follow-up resources.