In general population based cohorts from three European countries including over 13,000 children, we found a strong and consistent pattern of associations between core ADHD symptoms and scholastic impairment. These results are in line with previous research showing comorbidity between ADHD and learning problems [26, 27], and extend previous findings by confirming an association between core ADHD symptoms and impairment in general population samples. The same associations are similar to what has been shown for clinical cases of ADHD indicating that ADHD diagnoses are just extreme values from a continuous distribution (e.g. like hypertension).
Prior work has focused on literacy skills and found an association with ADHD . However, we found that mathematics was also strongly associated with core symptoms, even when taking into account indices related to disadvantage (maternal education and family structure) and child gender. The association between inattention and scholastic impairment was more pronounced than the association between hyperactivity symptoms and impairment. Similarly, others  have found reading difficulties were more strongly associated with the ADHD inattentive than the hyperactive subtype.
Girls had a lower prevalence of severe symptoms in all cohorts, which is in line with the gender disparity in ADHD diagnosis. The larger cohorts showed significantly stronger relative associations between core symptoms and impairment among girls as compared to boys.
A major change in the DSM-IV from earlier versions was the addition of impairment requirement. Symptoms must generate impairment in order to be considered a psychiatric disorder. Several reports show that many children are impaired but do not reach the threshold number of symptoms and, therefore, do not meet the diagnostic criteria . Impairment should weigh heavily in the diagnostic process even when children have fewer symptoms than required . Our data do not allow us to identify children meeting the ADHD diagnosis and many of the children that scored high on the core symptoms we recorded most likely do not meet the full criteria. Still, children who scored high on the core symptoms and have impairment are in need of support whether or not they fulfill diagnostic requirements. Children with psychiatric symptoms are more likely to dropout of secondary school  and children with ADHD have poor long-term academic underachievement . Further, scholastic impairment is related to delinquency [34, 35]. Leaving children's needs unmet (behaviorally or pharmacologically) could result in many of them not reaching their full potential.
Initial screening procedures could be conducted in schools as teachers are in a good position to first detect learning problems and notice behavioral deviations because children with ADHD symptoms often disrupt the classroom. Ideally screening for ADHD symptoms could be a part of the school health surveys. Alternatively, teacher ratings on core symptoms could be incorporated into routine progress reports given to parents. Early support in the form of academic assistance, teaching coping strategies to children (e.g. organizational skills training), and teachers' monitoring behaviors in the classroom would be important first steps. This would also facilitate an alliance between teachers and parents to work on shared goals. This strategy is in keeping with a greater emphasis being placed on support rather than assessment . Multimodal interventions alleviate symptoms and help in reducing impairment [37, 38]. Until then, however, parents or clinicians should request this information from teachers and make it available when determining whether further evaluation is warranted. This study is in line with previous research showing the value of teacher and parent rating scales and indicates that teacher ratings of a few core symptoms and impairment are informative and can be an initial step in following the ADHD evaluation guidelines [39–41].
Prevalence of core symptoms was related to living conditions in all three cohorts. Our results converge with previous reports showing that disadvantaged children are more likely to receive high ratings on behavioral problems [42, 43]. Thus, policy makers can increase surveillance of ADHD core symptoms and scholastic impairment among disadvantaged children who may be more vulnerable for negative outcomes.
The finding that inattention was lowest in Finland (even after stratification on living conditions) merits further research to discern whether etiological factors or other unmeasured factors related to living conditions are at the root of the observed difference. It is unclear to what extent methodological differences between cohorts could have contributed to the differences.
There are some limitations to consider. First, methodology differed somewhat between the cohorts e.g., data collection took place 6 years earlier in the Finnish cohort. This difference in time of data collection may explain why Finnish children were rated less inattentive than in the other two cohorts. It may be possible that awareness of ADHD symptoms has been steadily increasing within the last decade and may have affected teachers' willingness to endorse symptoms. However, it seems unlikely that increased awareness would pertain only to inattention and not to hyperactivity symptoms. A Swedish community study , which collected data at the time of the Finnish data collection (i.e. 6 years earlier), found similar prevalence rates as those presently reported for the Swedish cohort. Our Finnish prevalence rates are practically identical to results found from a later date from a study conducted in another part of Finland . Because Swedish and Finnish children were the same age, differences in inattention rate cannot be attributed to developmental effects. Thus, neither developmental differences nor time effects between cohorts can fully explain prevalence differences.
Second, we were limited to teacher-reported core symptoms and scholastic impairment. It may be that teachers tend to rate a child poorly in one area if he or she is rated poorly in another or that behavioral ratings bias perceptions of scholastic performance, e.g. due to a negative halo effect. However, teachers within each cohort rated various behaviors and outcomes pertaining to well-being and development, consequently the connection between the three core symptoms that we study and scholastic impairment was not likely to have been made. The majority of children in our study were first-graders, therefore, official grades or national test scores are not available with which to compare teacher ratings. Nonetheless, our measure has ecological validity as teachers are in the best position to rate academic performance and do so in reality. Teacher ratings of behavior have been found to reflect age-appropriate evaluations . Symptoms may present differently according to environment and it is suggested that inattention and hyperactivity symptoms are more reliably observed in a school setting . In this respect, core ADHD symptoms at school may be more relevant for scholastic impairment than in other environments. Recently, Caroll and colleagues  found that child literacy impairments were related equally to both teacher and parent ratings of ADHD symptoms using the SDQ. Further, there is a school effect related to child performance , however, teachers in our samples were not confined to any particular school, but came from large geographic areas.
Third, we related only three core symptoms to scholastic performance rather than clinical diagnosis. Our results show that functional impairment is not only limited to children who fulfill all the clinical criteria for ADHD, but suggest that increases in symptomatology and impairment go hand-in-hand in the general population.
Fourth, we used cross-sectional data although we expect a relation between ADHD symptoms and scholastic impairment to develop over time as has been suggested in clinical studies [3, 48]. It is likely that there is a dual pathway connecting behavioral symptoms and scholastic difficulties. Our follow-ups were initiated during the acquisition of scholastic skills (Swedish and Finnish cohorts). Thus, there may be a greater likelihood for the causal pathway to be from hyperactivity-inattention symptoms to scholastic deficits, which is consistent with findings showing inattention symptoms contribute to later reading difficulties .