Functional Outcomes Among Young People With Trajectories of Persistent Childhood Psychopathology

Key Points Question What functional outcomes in emerging adulthood (ages 17 to 20 years) are associated with persistent childhood psychopathology (ages 9 to 13 years)? Findings In this cohort study of 5141 participants, all types of psychopathology in childhood (internalizing, externalizing, or both) were significantly associated with poor functioning in emerging adulthood. This included poor mental and physical health, social isolation, heavy substance use, frequent health service use, poor subjective well-being, and adverse educational/economic outcomes. Meaning These findings highlight the lasting effects of childhood psychopathology on functional outcomes in emerging adulthood and point to the need for a public health approach to youth mental health.

year-olds (29.5-29.9),we chose to apply the standard adult cut-off for obesity, as defined by the World Health Organisation (≥ 30).

GP visits (>5)
Participants were asked "In the last 12 months, how many times have you seen or consulted, or talked on the phone with <a general practitioner> about your physical, emotional or mental health?".Response options were provided on the numeric-categorical scales shown in Figure S1.The cutpoint of >5 GP visits (i.e. 6 or more) in the past year was based on findings from a UK survey of over 4,000 individuals [3] .It found that the average number of GP visits per year was 5.5 for the whole sample but between 7 and 12 for those with severe mental health symptoms.

Leaving Certificate examinations (<300)
The leaving certificate is the final exam of the Irish secondary school system and the university matriculation examination in Ireland.In the established leaving certificate, students sit exams for a minimum of 6 subjects with ~100 points available for each subject.Students therefore typically score between 0 and 600 points.We chose 300 as the cut-off for low educational attainment as it corresponded roughly to the lowest quintile.A GUI report showed that less than half of those with points in the lowest quintile went onto higher education (42%), post-leaving certificate course (37%), or other further education (18%) [4] .
A small number of participants reported their leaving certificate results at age 17 (n=371), with the vast majority reporting at age 20.These responses were merged into one variable, with the most recent result taken if two responses were provided.

Equivalised income
One socioeconomic variable which was included as a covariate in adjusted analyses was household equivalised income (from age 13 data collection).An equivalence scale was used to assign a "weight" to each household member (weight of 1 to the first adult in the household, 0.66 to each subsequent adult aged 14+ and 0.33 to each child <14 years).The sum of these weights in each household gives the household's equivalised size -the size of the household in adult equivalents.Disposable household income was recorded as total gross household income less statutory deductions of income tax and social insurance contributions.Household equivalised income was the disposable household income divided by equivalised household size.

Parent Education
The education level of primary and secondary caregivers were merged according to the dominance criterion.That is, when valid responses were provided by both parents, the higher level of education was taken; when only one parent had responded, that was taken as the highest level of education.Levels included: none/primary school, lower secondary (e.g.Junior Cert), higher secondary (e.g.Leaving Cert) or technical vocational, non-degree certificate, primary degree, and postgraduate degree.

eMethods 2. Sampling Details
A two-stage recruitment process of the cohort was conducted. [5]First, 910 national schools were randomly selected for involvement, as the primary sampling units.Within these schools, children aged 9 and their families were invited to partake.Second, participants were selected to reflect the distribution of social and economic factors across the Irish population of nine-year-olds at that time.
Additional sample weights were created by comparing the sample with the children aged 9 in the 2006 Irish Census, based on several key socio-demographic variables and readjusted for over/under representation.Similarly, the sample was also re-weighted for attrition between the waves [6] .Thus, the statistics reported in this study are representative of 9-year-olds who were residing in Ireland, who continued to live in Ireland at 17-20 years of age (target population), and who exhibited similar mental health presentations at ages 9 and 13.

eAppendix. Results of Method Triangulation
We repeated the core analyses using alternative groups of child psychopathology, defined by simple cut-offs on the SDQ scales, rather than LCTA.The 80th percentile was used to identify clinically significant cases of psychopathology (sdqinfo.org) [7,8].All group sizes and estimates below are weighted to account for drop-off over time and sampling bias.
Definitions of each group are described in the method and eTable 10 below.Of the 6,039 participants in the sampe sample used for the core analyses, just 2,520 (41.7%) met one of these definitions.The LCTA = Latent Class Transition Analysis (as defined in [9] )

adult outcome: % of sample b Definition
Definitions of Each Outcome Variable and Descriptive Statistics for the Sample Max sample size at age 17 and 20 = 5,141 and 4,024.Differences in Demographic and Clinical Characteristics Between the Included Sample and Participants Who Were Excluded Based on a Changing Childhood Psychopathology Group Between Ages 9 and 13 Years Mean SDQ Scores for the Full Sample (Grey Rows) and Each Childhood Psychopathology Group at Age 9 Years (Top) and Age 13 Years (Bottom) Odds of Poor Functional Outcomes in Those With Any Persisting Psychopathology (ie, High, Externalising or Internalising), Unadjusted and Adjusted for Potential Confounds Odds Ratios of Adverse Adult Outcomes Between the Childhood Psychopathology Groups Fully Adjusted for Socioeconomic Factors and Sex Results of Sex-Stratified Analyses Fully adjusted effect of any childhood psychopathology on each functional outcome for males and females separately.Results of Sex-Stratified Analyses (Externalising Psychopathology)Fully adjusted effect of childhood externalising on each functional outcome for males and females separately.Odds ratios refer to the group difference between those with persistent externalising and those with no childhood psychopathology, adjusted for sex and socioeconomic factors.Results of Sex-Stratified Analyses (Internalising Psychopathology)Fully adjusted effect of childhood internalising on each functional outcome for males and females separately.Odds ratios refer to the group difference between those with persistent internalising and those with no childhood psychopathology, adjusted for sex and socioeconomic factors.a. Group sizes refers to those included in fully-adjusted model (i.e., with data on all covariates); Lower N available for educational/economic outcomes analysis (86 females & 103 males with persistent internalising) ≥ 1 A&E visits in past year @17 15.9% (n=815) 1 or more to "In the last 12 months, how many times have you seen or consulted, or talked on the phone with [A&E] about your physical, emotional or mental health?"≥ 1 A&E visits in past year @20 17.7% (n=711) 1 or more to "In the last 12 months, how many times have you seen or consulted, or talked on the phone with [A&E] about your physical, emotional or mental health?">5GPvisits per year @17 10.9% (n=554) Over 5 to "In the last 12 months, how many times have you seen or consulted, or talked on the phone with [GP] about your <3 to "How many friends do you normally hang around with?" Few friends @20 2.3% (n=94) <3 to "How many friends do you have?"No social support @17 9.7% (n=494) "No" to "Is there an adult in your life you can usually turn to for help and advice?"<5 to "If you were to describe how satisfied you are with your own life in general, how would you rate it on a scale of 0 to 10 (where 0 = extremely unsatisfied; 10 = extremely satisfied."Dissatisfiedwithlife@2018.5% (n=739) <5 to "If you were to describe how satisfied you are with your own life in general, how would you rate it on a scale of 0 to 10 (where 0 = extremely unsatisfied; 10 = extremely satisfied."a.parent-reported b. % of sample with valid/non-missing response to question NEET = Not in Education, Employment or Training eTable 2. All statistics are weighted to account for sociodemographic sampling biases and attrition.Weighted Ns using age 17 weights were: (i) 5,014 (ii) 1,380 and (iii) 1025.Unweighted Ns were: (i) 5,141 (ii) 1,178 and (iii) 898.Parental education, income and single parenthood recorded at participant age 13. a. High SDQ scores defined as scores of 17 or higher (sdqinfo.org)b.Low parental education defined as homes where parent(s) did not complete Leaving Certificate (final secondary school examinations) or an equivalent c.Group difference values are Chi-square statistics or t-values (T-tests) for frequencies and means respectively (*p < .05,**p<.01,***p<.001)eTable3.Bold effects refer to those significant at Bonferoni-corrected threshold (p < .007)±Adjustedforchild'ssex,parent education level, single-parenthood & household income at age 13 eTable 5.Odds ratios refer to the group difference between those with childhood psychopathology (internalising/externalising/high psych.)andthosewithnochildhood psychopathology, adjusted for sex and socioeconomic factors.eTable7.Odds ratios refer to the group difference between those with persistent childhood psychopathology (internalising/externalising/high psych.)andthosewith no childhood psychopathology, adjusted for sex and socioeconomic factors.a.Group sizes refers to those included in fully-adjusted model (i.e., with data on all covariates); Lower N available for educational/economic outcomes analysis (459 females & 576 males with any persistent psychopathology) b.Interaction effect refers to multiplicative interaction between childhood psychopathology and sex, adjusted for main effects of childhood psychopathology, sex and socioeconomic factors.eTable8.a.Group sizes refers to those included in fully-adjusted model (i.e., with data on all covariates); Lower N available was for educational/economic outcomes analysis (334 females & 419 males with persistent externalising) eTable 9.
3,519 participants not included are those who were above/below the 80th percentile at one age, and not at the next.Below, eTable 10 shows the breakdown of the 2,520 individuals into each group (weighted by age 13 sample weights); eTable 11 provides average SDQ scores for these groups; and eTables 12-13 provide odds ratios from adjusted analyses.eTable10.Definitions and Group Sizes for the alternative Childhood Psychopathology Groups Used in the Sensitivity Analysis Mean SDQ Scores for the Alternative Persistent Psychopathology Groups Used in the Sensitivity Analysis (n = 2520) Odds of Adverse Adult Outcomes in Those With Any Persistent Psychopathology From Age 9 to 13 Years Using Alternative Group Definitions for the Sensitivity Analysis Odds Ratios (OR) of Difference Between Childhood Groups in the Odds of Each Adverse Outcome in Young Adulthood (Reference Group: Low Psychopathology)Both original results (bold) and those from the alternative analysis (not bold) are shown.All estimates are adjusted for sex and childhood socioeconomic factors.