Early adult outcomes for Irish children with behavioural difficulties

Abstract

Background:

There are few longitudinal studies of Irish children and, to date, no studies reporting long-term outcome for those with behavioural disorder in childhood. This paper describes psychological and educational outcomes for a group of Irish children initially assessed at age 11 years and re-examined 10 years later. The period during which the research was conducted coincided with a time of increasing wealth in Ireland and the consequent development of psychological and educational services.

Aims:

To follow up and assess psychological and educational outcomes for a group of young people aged 21 years, half of whom had high levels of behavioural problems at age 11.

Method:

Data were gathered at two time points for 97 children and their families. Childhood measures included psychological and behavioural functioning, IQ, family background and economic circumstances. Outcome measures assessed in young adulthood included psychological functioning, educational attainment and trouble with the law.

Results:

Behavioural deviance at age 11 was found to be highly predictive of negative outcomes in early adulthood including a greater likelihood of involvement in criminal activity and less educational success. The likelihood of educational failure increased with the accumulation of risk factors including economic disadvantage and low IQ.

Conclusions:

These findings, supported in other international studies, underline the importance of behavioural difficulties in childhood for adult outcome, even in an environment of greater service and educational opportunities and access.

Keywords

behaviour disorder childhood disorder adult outcome follow-up study risk factors

Introduction

Early behavioural disturbance has been cited as one of the strongest predictors of later problems, including psychological difficulties, poor educational outcome, involvement in crime and antisocial behaviour (Colman et al., 2009; Fergusson & Lynskey, 1998; Robins, 1966; Simonoff et al., 2004). The fact that a child displays psychopathology in the early years makes it more likely that he/she will continue to experience difficulties into adulthood, and the risk of continuity is greatest for males (Miech, Capsi, Entner Wright, & Silva, 1999; Sourander et al., 2005; Visser, Van der Ende, Koot, & Verhulst, 1999). There are clear associations between early conduct problems and a range of familial factors such as insensitive parenting, family structure, marital discord and parental psychopathology (Berg-Nielsen, Vikan, & Dahl, 2002; Ford, Goodman, & Meltzer, 2004; Vostanis et al., 2006). There is also a strong association between externalizing problems and socioeconomic background (Costello, Compton, Keeler, & Angold, 2003). Such disorders are more strongly associated with lower socioeconomic status (SES) background (than, for example, depression) and exert a strong negative influence on educational attainment and thus on future SES (Miech et al., 1999; Zimmermann et al., 2003).

Two specific developmental outcomes consistently associated with early behavioural disturbances are poor educational attainment and delinquency (Timmermans, Van Lier, & Koot, 2009). Those with severe conduct problems are most likely to leave school without qualifications and are more likely, compared to those with less severe forms of externalizing behaviour, to experience social and economic adversity and mental ill health in adulthood (Colman et al., 2009). There is also an association between childhood conduct disorder and substance abuse in young adulthood (Sourander et al., 2005). According to Patterson, Reid, and Dishion’s (1992) coercion model, children’s early behaviour problems lead to deviant peer affiliation and influences, which escalate into educational underachievement and delinquency.

These associations are well established and powerful such that by the time adolescents with conduct problems reach adulthood ‘they appear to be “selected” into the lower socioeconomic strata through restricted educational attainment’ (Zimmermann et al., 2003, p. 1216). In adulthood these individuals have erratic employment patterns in unskilled jobs and tend to be less integrated socially and this pattern continues into middle age (Colman et al., 2009; Scott, Knapp, & Henderson, 2001). There is also a high correlation between behavioural difficulties in childhood and later offending; and low educational attainment is a key element of this trajectory (Kolvin, Miller, Scott, Gatzanie, & Fleeting, 1990).

This paper is based on a 10-year longitudinal study of young people in Dublin, Ireland, followed up from age 11 to 21 years. The aim of the study was to investigate the implications of early behavioural disturbance and socioeconomic disadvantage on educational outcomes and psychological health in early adulthood. This is the first follow-up study to examine these pathways among Irish youth.

Method

Sample

In the baseline phase of this study a sample of 185 10- and 11-year-old children was randomly selected from a larger epidemiological study of children growing up in West Dublin, Ireland. The children were assessed for psychological health status, behavioural deviance and IQ. Mothers’ psychological health status and family economic circumstances were also assessed. Ten years later, the child study participants, whose mean age was now 21 years (SD = 0.48), were followed up. One hundred and fifty three (82.7%) of the sample were located and 97 of these young people (50 males, 47 females) agreed to participate in the follow-up phase, representing a 52.4% retention rate over a 10-year period. Three participants had died (two of the deaths were drug related). Eighty mothers of the 97 young people also agreed to be re-interviewed as part of the follow-up phase of the research.

Those who participated in the follow-up phase were compared with those who refused to participate on a variety of variables from the baseline phase. No statistical differences were found between follow-up participants and refusers on gender, IQ score, psychiatric symptoms, family history of mental illness, family economic circumstances and employment status of the father. An analysis of participants not located for this stage of the study further revealed few differences between the two groups. Further analysis considered the distinction between participants at follow-up and those who could not be located at follow-up. Almost two-thirds (63.2%) of non-contacts were categorized as behaviourally deviant at age 11 compared with 33% of those who did participate, and this difference was statistically significant (χ² = 10.237, df = 1, p = .001).

Research instruments

During the baseline phase of the study, children’s health status was assessed using the Isle of Wight Parental Interview on Child’s Psychiatric State (Rutter & Graham, 1966). On the basis of this assessment, children were classified according to whether or not they received a psychiatric diagnosis. Behaviour was also assessed using the Rutter B2 Questionnaire (Rutter, 1967), which teachers completed for each child. On the basis of the score on this screening tool, children were classified as either ‘deviant’ or ‘non-deviant’. The Standard Progressive Matrices (Raven, Court, & Raven, 1983), a non-verbal test, was used to provide an indication of intellectual capacity. Information was also gathered from mothers, via semi-structured interview, on a range of social and economic items, and maternal psychological well-being was assessed using the Malaise Inventory (Rutter, Tizard, & Whitmore, 1970).

For the follow-up stage, a screening version of the Structured Clinical Interview for DSM-IV Diagnoses (SCID) was used to assess psychological functioning (First, Gibbon, Williams, Spitzer, & MHS Staff, 1999). The SCID is a semi-structured clinical interview that permits categorization into DSM-IV Axis I diagnoses (First, Spitzer, Gibbon, & Williams, 1996). The screening instrument, administered via a computer, covers the major areas of psychopathology: mood disorders, anxiety disorders, substance use disorders, psychotic symptoms, somatoform disorder and eating disorders. The schedule contains 497 questions, although fewer questions are usually administered as there is an automatic branching program that skips questions if certain symptoms are not present. Following completion, the program automatically generates a complete summary of patient responses and a summary of possible diagnoses. When the computer analysis was completed, the computer-generated diagnoses were checked by three psychiatrists familiar with the instrument. A high level of inter-rater reliability was attained in that they all agreed on case thresholds. The final diagnosis was based on their assessment plus the computer-generated results and the variable used in the analysis is the number of likely diagnoses arising from the SCID assessment. The Beck Scale for Suicide Ideation (Beck & Steer, 1991), a 21-item questionnaire measuring suicidal ideation and behaviour, was administered to detect and measure suicidal ideation and behaviour. Alcohol and substance misuse measurements were based on those used by Gill, Meltzer, Hinds, and Pettigrew (1996) in a national morbidity study in Great Britain. Self-esteem was assessed by the widely used Rosenberg Self-Esteem Scale (Rosenberg, 1965). Locus of control was measured using a scale devised by Pearlin, Menaghan, Lieberman, and Mullan (1981). The remaining areas of interest (employment, educational attainment, physical health, criminal activity) were examined by means of a semi-structured interview devised for the study by the authors. Participants were also asked to report upon whether or not they had experienced a psychological problem in the past year and whether they had previously received treatment for a psychological problem.

Analyses

The results are presented in two parts. The first part uses t tests, χ² analyses and Mann-Whitney tests to compare children who were classified as behaviourally deviant with those who were non-deviant at age 11 on outcome measures of educational attainment, trouble with the law, psychological health and other socio-demographic variables all assessed at the follow-up phase of the study. Comparisons between deviant and non-deviant groups are also carried out in terms of family background characteristics assessed at the baseline phase of the study. In addition, similar comparisons are conducted between children who received a formal psychiatric diagnosis at baseline and those who did not.

In the second part of the results, a series of linear and log-linear regressions are used to examine the predictive relationship between individual and family background characteristics, assessed at both phases of the study, and young adult outcomes. Three outcomes are considered: psychological health, involvement in deviant behaviour in adulthood (trouble with the law) and educational attainment.

Results

Comparison of those classified as behaviourally deviant/non-deviant in childhood

Approximately one-third (32.9%) of the respondents had been categorized as behaviourally deviant at age 11, based on the Rutter B2 Questionnaire. This group (n = 32) was compared with the non-deviant group (n = 65) to examine the association between behavioural deviance in childhood and psychological functioning, trouble with the law and educational attainment in early adulthood. There were more males than females in the deviant group (69% vs 31%) (Table 1).

Table 1.

Comparisons between deviant and non-deviant groups.

	Deviant group (n = 32)	Non-deviant group (n = 65)	Statistic	p
Comparison variable – measured at phase one of study
Receipt of state benefits when child was growing up	58.3%	26.8%	χ² = 7.235	.007
Mothers’ mental health score (Malaise Inventory)	M = 7.03 (SD = 4.44)	M = 4.48 (SD = 4.63)	U = 629.50	.004
Comparison variable – measured at 10-year follow-up
Trouble with the law	46.9%	13.8%	χ² = 12.563	< .001
Early school leaving	65.6%	18.5%	χ² = 21.250	< .001
Attended third-level education	50.8%	15.6%	χ² = 11.115	< .001
Psychological problem in the past year	12.5%	12.3%	χ² = 0.001	NS
Self-esteem	M = 20.56 (SD = 4.92)	M = 21.00 (SD = 3.96)	t = 2.319	NS
Locus of control	M = 13.59 (SD = 2.45)	M = 13.97 (SD = 2.49)	U = 967.00	NS
Young person: Number of likely diagnoses (SCID)	M = 1.67 (SD = 1.65)	M = 1.43 (SD = 1.48)	U = 937.00	NS
Mother: Number of likely diagnoses (SCID)	M = 1.42 (SD = 1.91)	M = 0.77 (SD = 1.21)	U = 556.50	NS

Behavioural deviance in childhood and family background characteristics

The marital status of parents was similar for both the behaviourally deviant and non-deviant groups in that almost all participants had grown up with both parents present. There was a statistically significant difference between the behaviourally deviant and non-deviant groups on mothers’ mental health (assessed by the Malaise Inventory) at the baseline stage of the study. The mothers of those in the behaviourally deviant group had a mean Malaise score of 7.03 (SD = 4.44) in contrast to a mean score of 4.48 (SD = 4.63) amongst other mothers.

There was no significant difference between the groups in relation to the mothers’ present psychological health (assessed by the SCID). One-third of mothers of behaviourally deviant respondents in comparison to 25% of the non-deviant group had experienced psychological problems in the year prior to interview. Twenty-five per cent of mothers in the behaviourally deviant group and 21.4% in the non-deviant group had received psychiatric treatment while their children were growing up, and four mothers of non-deviant respondents and two mothers of the deviant responders had been hospitalized for this. None of these differences were significant. There were no differences in relation to maternal alcohol abuse. Fathers’ psychological status was not assessed formally in the initial part of the study. However, retrospective data collected in the follow-up phase showed that a comparable proportion of fathers in the non-deviant (6%) and deviant groups (7%) had had psychological difficulties in the past and figures for alcohol misuse were not significantly different between the groups (20% for the non-deviant and 27.3% for the deviant group).

There was a significant difference between the groups in relation to indicators of family economic disadvantage. Over half (58.3%) of the behaviourally deviant group’s families, in contrast to 26.8% of the others, had been in receipt of benefits while the child was growing up (χ² = 7.235, df = 1, p = .007).

Behavioural deviance in childhood and socio-demographic and psychological health outcomes

There were no significant differences between deviant and non-deviant groups on the number of likely diagnoses obtained on the SCID or in relation to substance (including alcohol) misuse, suicide ideation, self-esteem or locus of control. There was little difference between those with and without a classification of behavioural deviance in childhood in terms of present psychological status. Only 13% of those classified as deviant (n = 4) and 12% of the others (n = 8) had experienced recent (i.e. in the past year) psychological problems. More behaviourally deviant respondents (25% vs 15% of the non-deviant group) had previously received treatment for psychological problems but this difference was not statistically significant and slightly more of the non-deviant group (four vs none) were receiving treatment at the time of the study. Almost half (46.9%) of the behaviourally deviant group, in contrast to 13.8% of the non-deviant group, had been in trouble with the law and this difference was statistically significant (χ² = 12.563, df = 1, p < .001).

Behavioural deviance in childhood and educational attainment and employment

Two-thirds (66%) of the behaviourally deviant group left secondary school before completing their final examinations (Leaving Certificate) compared with less than a fifth (18.5%) in the non-deviant group – a difference that was significant (χ² = 21.25, df = 1, p < .001). Participation in third-level education also varied considerably. About half (50.8%) of the non-deviant group went to third-level education compared to 15.6% of the group classified as behaviourally deviant and this was also statistically significant (χ² = 11.115, df = 1, p < .001). There were no significant differences between the two groups in relation to employment status although a higher percentage of the non-deviant group (54% vs 34%) were in non-manual socioeconomic categories.

Comparison of those with/without psychiatric diagnoses in childhood

Similar comparisons were conducted between those who had received a psychiatric diagnosis at the baseline phase of the study (n = 29) and those who did not (n = 68) (based on the Isle of Wight Parental interview on Child’s Psychiatric State; Rutter & Graham, 1966). The majority of the diagnoses were enuresis and conduct-related diagnoses. Three children received attention deficit hyperactivity disorder (ADHD)-related diagnoses. There was a slight preponderance of males in the diagnosed group (65% vs 54% of the non-diagnosed group). However, this difference was not statistically significant. There was a significant degree of overlap between those who received a diagnosis and those who were categorized as behaviourally deviant (χ² = 24.22, df = 1, p < .001). Of the 29 children who received a diagnosis, 20 were classified by teachers as behaviourally deviant.

As illustrated in Table 2, comparisons between those with a diagnosis and those without a diagnosis at the baseline phase yielded a similar pattern of findings to the deviant/non-deviant comparisons.

Table 2.

Comparisons between diagnosed and non-diagnosed groups.

	Diagnosis group (n = 29)	Non-diagnosis group (n = 68)	Statistic	p
Comparison variable – measured at phase one of study
Receipt of state benefits when child was growing up	59.1%	27.6%	χ² = 6.851	.009
Mothers’ mental health score (Malaise Inventory)	M = 7.32 (SD = 4.91)	M = 4.48 (SD = 4.38)	U = 575.5	.003
Comparison variable – measured at 10-year follow-up
Trouble with the law	37.9%	19.1%	χ² = 3.864	NS
Early school leaving	51.7%	26.5%	χ² = 5.776	.016
Attendance at third-level education	17.2%	48.5%	χ² = 8.352	.004
Psychological problem in the past year	10.3%	13.2%	χ² = 0.157	NS
Self-esteem	M = 19.34 (SD = 4.27)	M = 21.50 (SD = 4.16)	t = 2.319	.023
Locus of control	M = 13.31 (SD = 2.17)	M = 14.07 (SD = 2.57)	U = 887.00	NS
Young person: Number of likely diagnoses (SCID)	M = 1.86 (SD = 1.60)	M = 1.37 (SD = 1.50)	U = 771.50	NS
Mother: Number of likely diagnoses (SCID)	M = 1.32 (SD = 1.64)	M = 0.83 (SD = 1.39)	U = 523.50	NS

Mothers of those with a diagnosis scored higher on the Malaise Inventory and families of those with a diagnosis were more likely to have been in receipt of state benefits in comparison to those without a diagnosis at age 11. There were no differences between those with and without a diagnosis on maternal and young person’s psychological health (as assessed by the SCID in the follow-up phase), on young person’s locus of control or on the number of times the young person had been in trouble with the law. However, those with a diagnosis at age 11 had significantly lower self-esteem at follow-up, and a greater proportion of them left school early and did not proceed to third-level education.

Prediction of early adult psychological health outcomes and trouble with the law from child and family characteristics at age 11

The prediction of psychological health outcomes and trouble with the law from child and family characteristics assessed at age 11 was tested using linear regression analyses (Table 3). The number of likely diagnoses on the SCID tool and the number of times the young person reported to have been in trouble with the law were the dependent measures. Behavioural deviance was indicated by the total score received on the teacher-rated Rutter B2 scale. Other independent variables included family in receipt of state benefits and mother’s Malaise score (both measured at baseline). Self-esteem and locus of control (assessed at follow-up) were also included in the model as possible protective factors.

Table 3.

Linear regression models for predicting number of likely SCID diagnoses and number of times in trouble with the law.

	Number of likely SCID diagnoses	Number of times in trouble with the law
Family in receipt of state benefits	0.007 (0.346)	0.084 (0.234)
Maternal depression (score on Malaise Inventory)	0.037 (0.031)	0.005 (0.022)
Behavioural deviance (total score on Rutter B2)	−0.028 (0.022)	0.045** (0.016)
Self-esteem (assessed at follow-up)	−0.124** (0.045)	0.004 (0.032)
Locus of control (assessed at follow-up)	−0.014 (0.077)	−0.043 (0.056)
Adjusted R²	0.167	0.117

Standard errors in parentheses

p < .01

In the first model, only self-esteem was negatively associated with psychological outcome, suggesting that those with lower self-esteem had a higher number of likely SCID diagnoses. None of the variables measured at baseline were associated with psychological outcome. In the second model, only behavioural deviance was positively associated with level of trouble with the law.

Prediction of early adult educational attainment from child and family characteristics at age 11

A series of log-linear saturated and unsaturated models were tested to examine associations between child and family characteristics at age 11 and level of educational attainment 10 years later. Prior to building the models, bivariate associations between the various predictors and the outcome variable were examined.

The relationship between behavioural deviance and different levels of educational attainment is presented in Table 4 and indicates that those who had completed only part of second-level education had a higher number of deviant symptoms (χ² = 16.612, df = 2, p < .001). Based on Kruskal-Wallis tests of significance, those who had left school prior to completion of second-level education had higher scores on the Rutter B2 Scale (M = 9.97, SD = 7.99) indicating higher levels of behavioural deviance at age 11, than those who continued on to third-level education (M = 2.84, SD = 3.87).

Table 4.

Results of comparisons according to education level.

Comparison variable	Education level			Kruskal-Wallis statistic	p
Comparison variable	Second level part	Second level all	Third level	Kruskal-Wallis statistic	p
Rutter B2 score	M = 9.97 (SD = 7.99)	M = 6.00 (SD = 7.25)	M = 2.84 (SD = 3.87)	H = 16.612	< .001
IQ score	M = 30.78 (SD = 8.56)	M = 31.08 (SD = 11.59)	M = 38.19 (SD = 7.77)	H = 13.406	.001
Behaviour Categorization (n = 97)
Deviant (n = 32)	66% (n = 21)	19% (n = 6)	16% (n = 5)	c2 = 21.937	< .001
Non-deviant (n = 65)	18% (n = 12)	31% (n = 20)	51% (n = 33)
Family circumstances (n = 80)
In receipt of state benefits (n = 29)	55% (n = 16)	31% (n = 9)	14% (n = 4)	c2 = 15.162	< .001
Not in receipt of state benefits (n = 51)	22% (n = 11)	22% (n = 11)	56% (n = 29)

IQ also had an important relationship with educational attainment. Those who entered third-level education had higher IQ scores than those who had stopped schooling before the end of second level (χ² = 13.406, df = 2, p = .001). Based on Kruskal-Wallis tests of significance, those who had left school prior to completion of second-level education had lower IQ scores (M = 30.78, SD = 8.56) than those who continued on to third-level education (M = 38.19, SD = 7.77). Furthermore, when IQ scores were re-categorized into above average and average/below average, a significant difference between IQ categorization and educational outcome was apparent (χ² = 15.452, df = 2, p < .001). Those with above average IQ were very likely (over 57%) to go on to third level while over 51% of those with average/below average IQ left the educational system without completing second level.

Differences in educational attainment based on the presence or absence of economic disadvantage (measured by whether the family had been in receipt of state benefits when the child was growing up) were also examined and found to be significant (χ² = 15.162, df = 2, p = .001). Respondents from families who received state benefits were more likely to complete only part of second level education, while those from families not in receipt of state benefits were more likely to go on to third level.

Three factors assessed at age 11 – behavioural deviance, IQ and family in receipt of state benefits – emerged as statistically related to educational attainment. Two of these factors were also interrelated in that respondents with a deviant categorization tended to come from families who were in receipt of state benefits. The respondents’ IQ categorization was independent of the other three variables. A predictive model for educational attainment was developed, based on these key variables, using log-linear and logit models. The dependent variable was educational attainment and the independent variables were family in receipt of state benefit, IQ categorization (average/below average and above average), and behavioural deviance. Inspection of the coefficients and the Z values in Table 5 indicates that behavioural deviance was a stronger predictor of educational attainment than family in receipt of state benefits or IQ categorization.

Table 5.

Education, behavioural deviance, receipt of benefits, IQ: parameter estimates for specified logit model.

Variables	Coefficient	Standard error	Z value
Main effects
Education level	.3471	.213	1.626
	−.1821	.253	−0.718
Behavioural deviance	.6316	.205	3.07
	−.3697	.247	−1.49
Receipt of state benefits	.3813	.205	1.86
	.1973	.200	0.982
IQ categories	.3625	.201	1.796
	.1405	.190	0.736

Goodness-of-fit test statistics

Likelihood ratio χ² = 8.73531, df = 8, p = .365

Pearson χ² = 6.52279, df = 8, p = .589

Deviance, family in receipt of state benefits and IQ all had a significant, independent, direct effect on educational achievement and these differences were clearly evident at all three levels of education. Eighty-five per cent of respondents who were classified as behaviourally deviant, whose family were in receipt of state benefits during their childhood, and who were average/below average in IQ, did not complete second-level education. Conversely, only 8% of those outside these categories failed to complete second-level schooling. Overall, all non-deviant sub-categories were less likely than deviant sub-categories to leave school without completing second-level education (Table 6). The differences in relation to third-level education are stark. Eighty-five per cent of the behaviourally deviant group who were in receipt of benefits and with an average/below average IQ did not complete second level, compared to 43% of non-deviant respondents who had similar attributes (i.e. average/below average IQ and family in receipt of state benefits). This finding indicates clearly that a categorization of behavioural deviance, assessed at age 11, is a very sensitive predictor of educational attainment.

Table 6.

Log-linear model: categorization of behavioural deviance, family in receipt of state benefit, and IQ categorization, by level of education achieved.

			Second level part	Second level all	Third level
Deviant	State benefit	IQ average/ below average	85%	12%	3%
	State benefit	IQ above average	69%	16%	15%
	No state benefit	IQ average/ below average	68%	14%	18%
	No state benefit	IQ above average	35%	11%	53%
Non-deviant	State benefit	IQ average/ below average	43%	46%	10%
	State benefit	IQ above average	25%	42%	33%
	No state benefit	IQ average/ below average	24%	37%	39%
	No state benefit	IQ above average	8%	19%	73%

Two other factors, family in receipt of state benefit and IQ categorization, also had significant predictive powers, but were somewhat less important than a behaviourally deviant categorization. Respondents with average/below average IQ were less likely than those with above average IQ to pursue third-level education. Those groups in receipt of state benefits were more likely to leave the educational system without completing second-level education. However, those in receipt of state benefits with above average IQ presented a similar pattern of educational attainment to those with an average/below average IQ from families not in receipt of state benefits. This finding underlines the importance of socioeconomic factors in terms of educational attainment. A similar pattern was evident for both the behaviourally deviant and non-deviant groups but the behaviourally deviant group was less likely to complete second-level and go on to third-level education.

Discussion

Early behavioural difficulties are strongly associated with later psychological difficulties, poor educational attainment and involvement in crime (Colman et al., 2009; Fergusson & Lynskey, 1998; Robins, 1966; Simonoff et al., 2004). The findings reported here support this general trend. Irish children who were assessed as behaviourally deviant at age 11 were less successful educationally and more likely to have been in trouble with the law than the non-behaviourally deviant group. The association between early behavioural disturbance and later involvement in deviant behaviour was well established in the present study, in line with previous research (Fergusson & Lynskey, 1998). One limitation of the study was that information was not collected regarding the age of the young person at the time of their first contact with the police. Such information could be useful in understanding trajectories of deviance from middle childhood through adolescence and into adulthood.

In line with previous research, the association in this study between early behavioural difficulties and lower educational attainment strengthened with the accumulation of risk factors – specifically low IQ and economic disadvantage (Colman et al., 2009; Costello et al., 2003; Zimmerman et al., 2003). In Colman et al.’s (2009) study, severe externalizing behaviour predicted early school leaving and Kolvin et al. (1990) identified early school problems amongst a delinquent group and, conversely, the protective influence of educational achievements and high IQ in a control group. Miech and colleagues’ (1999) long-term follow-up study of a Finnish birth cohort demonstrated an association between poor school achievement and future mental health. Another Finnish study (Sourander et al., 2005) with a comparable age span to the present study supports the link between early conduct disorder, educational underachievement and mental disorder in early adulthood. Although it did not emerge as significant in this study, co-morbidity of conduct disorder and conditions such as ADHD are common and this increases the risk of negative outcomes (Babinski, Hartsough, & Lambert, 1999).

The link between economic disadvantage and behavioural problems in children is well established in the literature (Miech et al., 1999; Zimmermann et al., 2003). Conduct disorder occurs four times more often in families with unskilled occupations than in professional families. Major contributing factors include family-based stressors such as mental illness, marital disharmony and violence (Scott et al., 2001). According to Kolvin et al. (1990), the probability of negative outcomes for the child increased markedly with the degree of family deprivation. Resonant with the findings of this study, other researchers have described a cycle of disadvantage set in motion by early behavioural problems and the economic costs to society of these disorders (Scott et al., 2001; Simonoff et al., 2004). In line with the present study, Simonoff and colleagues cited the important role of IQ and reading level in assessing long-term outcome in children and the mediating role of early school leaving in negative adult outcomes. Similarly, Fergusson and Horwood’s (1998) investigation indicated that early disruptive behaviour makes it more likely that there will be an early exit from school and hence the individual is less likely to attain a stable occupational status.

The lack of association between early behavioural disturbance and subsequent psychological disorders as indicated by the SCID screening measure was somewhat surprising and diverges from well-established links in the literature (Colman et al., 2009; Sourander et al., 2005). This finding may indicate a limitation of the SCID screening measure to discriminate cases of psychological disorder from non-cases. In the current study, self-esteem emerged as the only significant predictor of adult psychological outcomes, suggesting that high levels of self-esteem may act as a protective factor against psychological disorder (Rutter, 1985). Of course, it is also possible that there is a bidirectional process in operation whereby a positive sense of self is more likely in the absence of psychological disorder. This finding points to the importance of interventions to promote self-esteem and self-efficacy in order to buffer against potential psychological distress.

In conclusion, until now no longitudinal research has considered the association between early childhood behavioural disturbance and psychological and social outcomes in early adulthood within an Irish context. The findings of this study indicate that behavioural deviance in childhood or early adolescence impairs individuals in terms of educational achievement and that this situation is exacerbated in the contexts of comparatively low IQ and economic disadvantage. Conduct disorder is one of the most common of childhood psychiatric disorders and is a strong predictor of adult health and related problems. This is borne out in this study of Irish children. A key strength of the study is its longitudinal dimension over a relatively long period of time. The sample size is small and over the 10-year follow-up study, a retention rate of 52% was achieved. This compares well with other community-based longitudinal studies conducted over a similar period (Dalgard, Bjork, & Tambs, 1995; Serbin et al., 1998). The findings from this and comparable community-based follow-up studies illustrate the challenges facing individuals within low socioeconomic contexts who display early behavioural disturbance. In identifying the school as a key site for the detection of conduct disorder, the study points to possible areas of intervention. Early diagnosis, along with intervention strategies that include teachers, parents and children, are more likely to result in educational success/retention, which emerges as a key factor in preventing negative outcomes for those with conduct disorder.

Footnotes

Acknowledgements

The authors would like to acknowledge the assistance of Professor Michael Fitzgerald, Trinity College Dublin, in initiating and developing this study and the Department for Social and Family Affairs for funding the research.

References

Babinski

L. M.

Hartsough

C. S.

Lambert

N. M.

(1999). Childhood conduct problems, hyperactivity-impulsivity, and inattention as predictors of adult criminal activity. Journal of Child Psychology and Psychiatry and Allied Disciplines, 40, 347–355.

Beck

A. T.

Steer

R. A.

(1991). Manual for Beck Scale for Suicide Ideation. San Antonio, TX: Psychological Corporation.

Berg-Nielsen

Vikan

Dahl

(2002). Parenting related to child and parental psychopathology: A descriptive review of the literature. Clinical Child Psychology & Psychiatry, 7, 529–552.

Colman

Murray

Abbott

R. A.

Maughan

Kuh

Croudace

T. J.

Jones

P. B.

(2009). Outcomes of conduct problems in adolescence: 40 year follow-up of national cohort. British Medical Journal, 338, a2981.

Costello

E. J.

Compton

S. N.

Keeler

Angold

(2003). Relationships between poverty and psychopathology. A natural experiment. Journal of the American Medical Association, 290, 2023–2029.

Dalgard

O. D.

Bjork

Tambs

(1995). Social support, negative life events and mental health. British Journal of Psychiatry, 166, 29–34.

Fergusson

Horwood

(1998). Early conduct problems and later life opportunities. Journal of Child Psychology & Psychiatry, 39, 1097–1108.

Fergusson

Lynskey

(1998). Conduct problems in childhood and psychosocial outcomes in young adulthood: A prospective study. Journal of Emotional & Behavioural Disorders, 6, 2–18.

First

Gibbon

Williams

Spitzer

, & MHS Staff (1999). SCID screen patient questionnaire extended (SSPQ-X) for Windows. Toronto: Multi-Health Systems.

10.

First

Spitzer

Gibbon

Williams

(1996). Structured Clinical Interview for DSM-IV Axis I Disorders, Clinical Version (SCID-CV). Washington, DC: American Psychiatric Press.

11.

Ford

Goodman

Meltzer

(2004). The relative importance of child, family, school and neighbourhood correlates of childhood psychiatric disorder. Social Psychiatry and Psychiatric Epidemiology, 39, 487–496.

12.

Gill

Meltzer

Hinds

Pettigrew

(1996). Psychiatric morbidity among homeless people. OPCS Surverys of psychiatric morbidity in Great Britain, Report 7. London: HMSO.

13.

Kolvin

Miller

F. J. W.

Scott

D. M.

Gatzanie

S. R. M.

Fleeting

(1990). Continuities of deprivation? The Newcastle 1000 Family Study. Aldershot: Avebury.

14.

Miech

R. A.

Capsi

Entner Wright

B. R.

Silva

P. A.

(1999). Low socioeconomic status and mental disorders: A longitudinal study of selection and causation during young adulthood. American Journal of Sociology, 104, 1096–1131.

15.

Patterson

G. R.

Reid

J. B.

Dishion

T. J.

(1992). Antisocial boys. Eugene, OR: Castalia.

16.

Pearlin

L. I.

Menaghan

E. G.

Lieberman

M. A.

Mullan

J. T.

(1981). The stress process. Journal of Health & Social Behaviour, 22, 337–356.

17.

Raven

J. C.

Court

Raven

(1983). Manual for Raven’s Progressive Matrices and Vocabulary Scales. Research supplement no. 1. London: H. K. Lewis & Co. Ltd.

18.

Robins

(1966). Deviant children grown up: A sociological and psychiatric study of sociopathic personality. Baltimore, MD: Williams & Wilkins.

19.

Rosenberg

(1965). Society and the adolescent self-image. Princeton, NJ: Princeton University Press.

20.

Rutter

(1967). A children’s behavioural questionnaire for completion by teachers: Preliminary findings. Journal of Child Psychology & Psychiatry, 8, 1–11.

21.

Rutter

(1985). Resilience in the face of adversity. Protective factors and resistance to psychiatric disorder. British Journal of Psychiatry, 147, 598–611.

22.

Rutter

Graham

(1966). Psychiatric disorder in 10- and 11-year-old children. Proceedings of the Royal Society of Medicine, 59, 382–387.

23.

Rutter

Tizard

Whitmore

(1970). Education, health and behaviour. London: Longman.

24.

Scott

Knapp

Henderson

(2001). Financial cost of social exclusion: Follow-up study of antisocial children into adulthood. British Medical Journal, 323, 191–194.

25.

Serbin

Cooperman

J. M.

Peters

P. L.

Lehoux

P. M.

Stack

D. M.

Schartzman

A. E.

(1998). Intergenerational transfer of psychosocial risk in women with childhood histories of aggression, withdrawal, or aggression and withdrawal. Developmental Psychology, 34, 1246–1262.

26.

Simonoff

Elander

Holmshaw

Pickles

Murray

Rutter

(2004). Predictors of antisocial personality: Continuities from childhood to adult life. British Journal of Psychiatry, 184, 118–127.

27.

Sourander

Multimaki

Nikolakaros

Haavisto

Ristkari

Helenius

Almqvist

(2005). Childhood predictors of psychiatric disorders among boys: A prospective community-based follow-up study from age 8 years to early adulthood. Journal of the American Academy of Child & Adolescent Psychiatry, 44, 756–767.

28.

Timmermans

Van Lier

P. A. C.

Koot

H. M.

(2009). Pathways of behavior problems from childhood to late adolescence leading to delinquency and academic underachievement. Journal of Clinical Child & Adolescent Psychology, 38, 630–638.

29.

Visser

J. H.

Van der Ende

Koot

Verhulst

F. C.

(1999). Continuity of psychopathology in youths referred to mental health services. Journal of the American Academy of Child & Adolescent Psychiatry, 38, 1560–1568.

30.

Vostanis

Graves

Meltzer

Goodman

Jenkins

Brugha

(2006). Relationship between parental psychopathology, parenting strategies and child mental health. Social Psychiatry & Psychiatric Epidemiology, 41, 509–514.

31.

Zimmerman

Wittchen

H. U.

Hofler

Pfister

Kessler

R. C.

Lieb

(2003). Primary anxiety disorders and the development of subsequent alcohol use disorders: A 4-year community study of adolescents and young adults. Psychological Medicine, 33, 1211–1222.