Specialized Service Use for Psychiatric and Neurodevelopmental Disorders by Age 14 in Finland

Altogether, 12.9% (N=7,561 of 58,538) of children had used specialized services for psychiatric or neurodevelopmental disorders between birth in 1996 and 2010 (age 14). For a psychiatric or neurodevelopmental disorder, 2,785 children had visited pediatric neurology clinics (cumulative incidence 4.8%), 3,077 had visited child or adolescent psychiatry clinics (cumulative incidence 5.3%), and 2,989 had visited pediatric clinics (cumulative incidence 5.1%). A total of 1,290 children had visited two or more of these specialties (cumulative incidence 2.2%), and 962 had been admitted to a child or adolescent psychiatry inpatient unit (cumulative incidence 1.6%).

The diagnostic groups clustered into three groups with regard to the sex distribution, to the age period when the disorders were first diagnosed, and to lifetime comorbidity in specialized care. First, learning and coordination disorders, autism spectrum disorders, and hyperkinetic disorder were more prevalent among boys and were often diagnosed before school age, and the diagnostic groups overlapped to a certain degree. According to a review of sex differences (10), these disorders occur mostly among males and typically involve neurodevelopmental impairment. In fact, they are often termed neurodevelopmental disorders (11). However, the cumulative incidence for these neurodevelopmental disorders continued to increase until the end of follow-up; that is, for many children these disorders were diagnosed in late childhood, although the symptoms per definition must be present before school age. For example, the cumulative incidence of specialized service use for autism spectrum disorders was .5% by age eight and 1.0% by age 14. According to a systematic review from 2012 (12), the prevalence of autism spectrum disorders in recent European studies has ranged between .3% and 1.2%, with a median of .6%. This indicates that a large proportion of Finnish children with autism spectrum disorders sometimes receives a diagnosis in specialized services, but the diagnosis often is established at school age.

Second, conduct disorders, including oppositional defiant disorder, had a male predominance, in line with other studies (10), but were seldom diagnosed before school age and overlapped to a maximum of 11% with the other diagnostic groups. It is notable that the proportion of children with conduct disorder and a comorbid disorder was low compared with the level of comorbidity in population-based surveys (13,14). Although we cannot rule out that hyperkinetic and internalizing disorders are undiagnosed among children with conduct disorder, it is more likely that the low level of comorbidity among children with conduct disorders relates to the diagnostic classification in ICD-10. There is a specific code for “hyperkinetic disorder associated with conduct disorder” (F90.1) under the chapter of hyperkinetic disorders (F90) in the ICD-10, and there is a separate group for “mixed disorders of conduct and emotions” (F92). Accordingly, hyperkinetic disorders associated with conduct disorder were classified as hyperkinetic disorders, and mixed disorders of conduct and emotions were classified as conduct disorders.

Third, prevalence of anxiety disorders, depression, and stress and adjustment disorders was relatively equal between boys and girls, and these disorders were often diagnosed in late childhood or early adolescence, with lifetime comorbidity of these disorders present in many cases. In line with a previous review (10), anxiety and depressive disorders had a similar sex distribution in prepubertal years but were more prevalent among females after puberty. Depression and anxiety are often termed emotional disorders, whereas stress and adjustment disorders can involve both emotional and disruptive symptoms. The diagnostic pattern of stress and adjustment disorders resembled the pattern seen mostly with emotional disorders. Some children first received diagnoses of stress and adjustment or anxiety disorders and later received a diagnosis of depression, because there was a high proportion of comorbidity between these disorders and the age at which stress, adjustment, and anxiety disorders were diagnosed was earlier than the age for a diagnosis of depression.

Several methodological issues should be considered when comparing the cumulative incidence results with the results from other service use studies. For example, survey-based studies potentially are prone to recall bias, whereas register-based studies are not; register-based studies include only specified services, whereas surveys can collect information for different forms of service; and some registers allow for long-term follow-up at the individual level, whereas other registers do not. A novel finding was that 8% of boys and 3% of girls had used specialized services for learning and coordination disorders between birth and age 14, when the data were treated as longitudinal data, but the corresponding proportion was only .9% for boys and .4% for girls when the register-based data were treated as cross-sectional data at age 14. This difference indicates that most Finnish children with learning and coordination disorders visited specialized care only for a short time and that long-term follow-up designs are necessary for identifying these disorders in register-based research. The longitudinal cumulative incidence estimates of learning and coordination disorders, rather than the cross-sectional one-year prevalence, resemble results from surveys. For example, in a cross-sectional U.S. survey from 1997 to 2008 (15), teachers or health professionals had identified 9% and 5% of three- to 17-year-old boys and girls, respectively, with learning disabilities. Future studies with both survey-based and register-based information are needed to investigate to what extent these kinds of information differ.

The cumulative incidence of specialized service use for conduct disorders, including oppositional defiant disorder, was 1.7% by age 14. The comparison with community-based samples is challenging because of methodological issues, but this figure can be considered low. For example, the three-month prevalence of conduct disorder and oppositional defiant disorder among Finnish eight- to nine-year-old children was 5% in a study based on parental interviews (16), and similar figures have been found in the United Kingdom (17) and the United States (18). Similarly, anxiety disorders in specialized care had a cumulative incidence of 2.2% by age 14, whereas the three-month prevalence of anxiety was 5% at age eight in a sample of Finnish children, including children who received treatment and children who did not (16). In line with previous studies (1,2), this difference indicates that in Finland some psychiatric problems, such as conduct and anxiety disorders, are underrecognized in primary care or are not referred to specialized services.

It is well known that neurodevelopmental disorders occur more often among boys than girls, because of different risk and protective factors between the sexes (10). It is also acknowledged that the sex ratio for service use for disorders might be even higher than the sex ratio for the disorders per se (10). As when comparing cumulative incidence with other studies, comparing sex ratios with other studies is challenging because of methodological differences. For example, the boy-girl ratio for autism spectrum disorders was 3.8:1.0 in our study, compared with 2.5:1.0 in a recent epidemiological study from South Korea (19) and 3.9:1.0 in another epidemiological study from the United Kingdom (20). The ratio in this study indicates that there was no major gender bias in referral of children with autism spectrum disorders to specialized services in Finland, but studies with more detailed data on nontreated cases are needed to confirm these findings and give insight into possible gender biases of the full spectrum of disorders.

The strengths of this study include total nationwide longitudinal data from birth, both inpatient and outpatient units from all specialties, and a uniform diagnostic system (ICD-10). There are also several limitations to consider. The results represent all children with specialized service use for psychiatric and neurodevelopmental disorders in Finland but lack information about children who are not recognized by primary care and are not referred to specialized services. Having such information would allow more definitive conclusions to be drawn in regard to unmet need of services and under- and overdiagnosis. Furthermore, the data required complete information concerning specialized service use, because data on primary care visits are not yet available for research purposes in Finland. In addition, the diagnostic validity of childhood autism has been assessed and has been found to be high (21), but it is not known how well other disorders correspond to research diagnostic criteria. Moreover, the total sample was based on all singleton live births, but emigration, immigration, and deaths were not taken into account in the analyses. However, mortality is very low, emigration ranges between .1 and .4% per year, and immigration ranges between .3 and .8% per year among Finnish people between birth and age 14 (22). It is therefore unlikely that these factors would substantially affect the results.

Every eighth Finnish child in this study had visited specialized services for psychiatric or neurodevelopmental disorders sometime between birth and age 14. The most common diagnosis was learning and coordination disorders, which were diagnosed more than twice as often as anxiety, hyperkinetic, and conduct disorders. It is also evident that many children received their diagnosis at a later age than would be expected from what is known about the typical age of symptom onset. One way to improve the early recognition of childhood psychiatric and neurodevelopmental disorders in the population is to implement systematic screening of mental health problems in child welfare and school health clinics. Such an arrangement would be in line with the systematic assessment and referral policies of, for example, scoliosis and deviant growth patterns during childhood (23). However, this would require children with unmet needs to have access to effective mental health interventions, which currently are mainly available in specialized services. The long-term burden of childhood mental health problems is likely to remain high if only a proportion of children with problems access appropriate treatment. Therefore, early identification and a smooth referral system are important to decrease the long-term burden of these problems in the population. Other important ways to promote mental health in childhood and adolescence include educational programs in school settings to increase the knowledge concerning mental health, pathways to care and stigma (24), and easily accessible interventions in primary care.

This study was funded by National Institutes of Health grant P50MH090966 (Project 1, Dr. Brown and Dr. Sourander, co–principal investigators) and by the Sackler Center for Developmental Psychobiology at Columbia University. Dr. Gyllenberg was funded by the Sigrid Juselius Foundation, by the Finnish Medical Association, and by the Foundation for Pediatric Research in Finland.

The authors report no competing interests.