Studies have shown that children with ADHD followed through adulthood have deficits in multiple domains compared with their non-ADHD peers. However, these functional outcomes are associated with substantial variability, warranting further investigation. This study, which is the longest prospective follow-up of children with ADHD to date, examined whether characteristics present in childhood and adolescence were associated with adult functional outcomes.
The study recruited 207 boys with ADHD aged 6–12 years (mean age 8.36 ± 1.63 years) who were referred to a child psychiatric clinic due to behavioural problems.* The boys were assessed longitudinally during three follow-ups, at a mean age of 18.0 ± 1.3 years (range 16–22; n=195), 25.3 ± 1.4 years (range 22–30; n=176) and 41.4 ± 2.9 years (range 30–47; n=135), respectively. Trained clinicians who were blinded to all previous data conducted the interviews at each follow-up.
Predictive characteristics were selected stepwise from the large number of measures recorded in childhood and adolescence. Childhood predictors included parents’ social-economic situation (SES), children’s full-scale IQ, reading level and severity of ADHD, conduct problems, oppositional defiant behaviour, eruptive aggression, immature behaviour and social functioning. Adolescent predictors included educational and work goals, global job and social functioning, parental alcohol or non-alcohol substance-use disorder and parental antisocial personality disorder. In adulthood, multiple functional domains were examined, including overall function, educational attainment and occupational ranking. Linear regression analysis was then used to test the relationship between characteristics in childhood and adolescence and functional outcomes in adulthood.
Within the overall sample, childhood ADHD ratings were increased (mean 2.3 ± 0.4; 0–3 scale) and conduct problems were decreased (0.8 ± 0.4; 0–3 scale). At adulthood, the average Global Assessment Scale score was 63.5 ± 13.6, which reflects relatively good functioning with mild symptoms or some difficulty. Mean years of education was 13.3 ± 2.1 and the average Hollingshead occupational rating for best job ever held was 3.9 ± 1.6 (where 8 = higher executive and 1 = unemployed). Mean occupational functioning was in the ‘3 = average’ to ‘4 = good’ range (3.2 ± 1.2 at 17 years; 2.8 ± 1.0 at mean age 25 years; 3.3 ± 1.3 at mean age 33 years; 3.6 ± 1.3 at mean age 41 years). Additionally, average levels of social functioning at the three age follow-ups were 3.6 ± 1.3, 2.9 ± 0.7 and 3.2 ± 1.2, respectively.
Results demonstrated that a number of childhood and adolescent characteristics were significant predictors of functional outcomes in adulthood:
- Overall function: Conduct problems in childhood were associated with poorer overall function in adulthood (b = –6.53; standard error [SE] 2.91; p = 0.03). Conversely, high educational goals in adolescence were associated with improved functioning (b = 3.20; SE 1.16; p = 0.007).
- Occupational functioning: Occupational functioning improved slightly over time (b = 0.01; SE 0.01; p < 0.05); a better trajectory was also observed in children with a higher IQ (b = 0.01; SE 0.01; p < 0.01) and in adolescents with better social functioning (b = 0.18; SE 0.06; p = 0.002). Conversely, severity of conduct problems in childhood was associated with poorer occupational functioning over time (b = –0.40; SE 0.18; p < 0.05).
- Social functioning: The level of social functioning worsened over time (b = –0.01; SE 0.01; p < 0.05); IQ was the only significant childhood predictor of social functioning (b = 0.01; SE 0.01; p < 0.05), whilst job functioning in adolescence (b = 0.13; SE 0.05; p = 0.008) was positively related to social functioning over time.
- Educational attainment: Higher SES (b = 0.48; SE 0.20; p < 0.05), IQ (b = 0.06; SE 0.02; p = 0.002) and reading achievement (b = 0.03; SE 0.01; p < 0.05) in childhood predicted better educational attainment in adulthood, whereas conduct problems in childhood (b = –0.96; SE 0.48; p < 0.05) and antisocial behaviour in adolescence (b = –0.13; SE 0.04; p = 0.002) were associated with lower adult educational attainment.
- Occupational ranking: Higher IQ in childhood was the only significant predictor of occupational ranking in adulthood (b = 0.06; SE 0.01; p < 0.001).
There were several limitations to this study. The authors could not determine whether the findings reported here generalise to women, individuals from other ethnic or racial backgrounds, or children with ADHD of the predominantly inattentive type. Additionally, some analyses may have been underpowered and, as a result, did not detect significant associations. Furthermore, some associations were not robust enough to adjust for multiple comparisons. It is also important to note that even when single predictors were significantly related to functional outcome, these associations were mostly weak.
The authors of this study concluded that other than childhood IQ, which predicted improved outcomes in multiple aspects of adult functioning, there were no childhood or adolescent characteristics that consistently predicted later outcomes. The authors stated that these findings supported those reported in previous studies, which also identified few common predictors of adult outcomes in children with ADHD. Providing additional support to children with ADHD who have a lower IQ may improve their adult functional outcomes; however, predicting the course of childhood ADHD remains problematic and requires further investigation.
Find out more about predicting the course of childhood ADHD here
*Study inclusion criteria were history of behavioural problems, increased teacher and parent ratings of hyperactivity, behavioural problems at home and school, verbal IQ ≥85 and English-speaking parents. The children’s clinical picture was consistent with the Diagnostic and Statistical Manual of Mental Disorders – 5th Edition definition of ADHD combined presentation, as they had increased teacher ratings of inattention and hyperactivity/impulsivity symptoms, and symptoms were impairing and cross-situational, and present before 12 years of age. Exclusion criteria included children with neurological or significant medical disorders, psychosis or conduct disorder
Ramos-Olazagasti MA, Castellanos FX, Mannuzza S, et al. Predicting the adult functional outcomes of boys with ADHD 33 years later. J Am Acad Child Adolesc Psychiatry 2018; 57: 571-582.e1.