Introduction

​

Attention deficit and hyperactivity disorder (ADHD) is a neurodevelopmental disorder, characterized by a persistent pattern of inattentive, hyperactive and impulsive behavior, which can interfere with educational, social and occupational functioning Adolescence ADHD has been frequently linked to increased engagement in various specific risk-taking behaviors (RTBs) including smoking and substance abuse, dangerous driving, gambling  and unprotected sex . In the clinical and developmental literature, risk-taking behavior is defined as deliberate engagement in behaviors that are associated with some probability of undesirable results. Indeed, domain-general risk-taking behavior has also been documented in adolescent ADHD.

Stattin and Kerr defined parental monitoring as the parents’ knowledge about their child’s activities: knowing where, how and with whom their children spend their time. Numerous studies have linked poor maternal monitoring with children RTB and externalizing behavioral problems. Furthermore, this source of information was found to be a substantially stronger predictor of child’s engagement in RTB and externalizing behavioral problems, than other parenting behaviors such as parental control and surveillance. Parental monitoring was found to correlate with specific risk-taking behaviors in ADHD populations as well. Higher parental knowledge was associated with less delinquency and substance use in adolescents with ADHD, compared to controls. Low parental knowledge of the child’s whereabouts was found to moderate the association between ADHD and alcohol use frequency.

​

Though parental monitoring plays an important role in the prevalence and extent of RTB among adolescents with ADHD, there is a possibility that the capacity of parents of youth with ADHD to monitor their children’s whereabouts is influenced by some of the characteristics of the disorder. ADHD was found to challenge parental functioning. Parents of children with ADHD reported less parental satisfaction and efficacy and more parenting stress compared to parents of children without ADHD. In addition, parents of children with ADHD reported less parental warmth and more inconsistent and hostile parenting. In another study, ADHD symptoms were associated with to maternal inconsistent discipline and paternal low involvement. Moreover, in relation to parental monitoring, a recent study by Salari and Thorell has shown that parents’ reports of higher parental knowledge and child disclosure correlated with lower ADHD symptoms.

Taken together, the results of previous research suggest a complex relationship between ADHD symptoms, parenting practices and RTB, according to which ADHD symptoms contribute to lower parental monitoring and knowledge, which in turn may contribute to higher engagement in risk-taking behavior. The aim of the current study was to further examine this model. In particular, we were interested in revealing the direct and the indirect effect of ADHD on RTB, mediated by lower parental knowledge. Additionally, the majority of the literature regarding ADHD-related risk-taking behavior compared individuals who were diagnosed with ADHD to controls who were not diagnosed with the disorder. The existing literature’s reliance on classified diagnostic groups is a limitation, given that taxonomic and genetic studies consistently support a dimensional rather than categorical structure of ADHD. We have recently shown an RTB-ADHD symptoms correlation in a sample of adults from the general population. The current study was therefore interested in assessing the mediation model in relation to ADHD symptomatology level. We hypothesized that parents of children with elevated ADHD symptoms will show reduced knowledge of their child’s behaviors and that this in turn will contribute to higher RTB.

​

Methods

Subjects

The study was approved by the Shaare Zedek Medical Center Institutional Review Board for research on human subjects. A sample of 55 dyads of adolescents with ADHD, ages 13-18, and one of their parents (the vast majority of whom were mothers), were recruited through the outpatient ADHD clinic of the Neuropediatric Unit in Shaare Zedek Medical Center. Inclusion criterion for the ADHD group was a diagnosis made by a child neurologist or psychiatrist. In addition, a group consisted of 37 dyads of adolescents who did not have a history of ADHD diagnosis, and their parents, was recruited from local schools. Exclusion criteria for all participants were history of serious neurological illness (i.e., epilepsy, cerebral palsy), autistic spectrum disorder, psychosis, and attending special education schools for intellectually disabled students. Written informed consent was obtained from all participating parents, and assent was obtained from adolescents. Four subjects, three with and one without a diagnosis of ADHD were removed from the database due to incomplete records.

​

Measures

Severity of ADHD symptoms was measured using the inattention, hyperactivity-impulsivity and oppositional-defiance scales of Disruptive Behavior Disorder Rating Scale. For the correlation and mediation analyses, scores of inattention and hyperactivity-impulsivity were combined into a mean ADHD score, which served as the predictor variable.

The risk behavior questionnaire consists of 16 items probing for different kinds of risky behaviors, e.g. smoking cigarettes, unprotected sex, hitchhiking. Participants are requested to rate the frequency by which they are involved in each of these behaviors in the present (for details refer to Pat-Horenczyk et al.). Reported internal consistency for the risk behavior questionnaire was high (Cronbach’s α = .86). Recently, using this tool elicited significant higher engagement of risk-taking behavior in a similar sample of adolescents with ADHD compared to control group.

For measuring parental practices, as perceived by the children, we used the behavioral control scale of Barber’s parental control questionnaire. This scale consists of three subscales probing excessive behavioral autonomy (e.g. “My mother lets me do anything I like to do”), disengaged family relationships (e.g. “There are very few rules in our family”), and parental monitoring/knowledge (e.g. “How much do your parents really know where you go at night?”), the last of which was hypothesized to be the mediator variable in the model.

​

Procedure

Trained research assistants (graduate psychology students) performed all assessments. After obtaining consents from parents and assents from children, parents completed the DBDRS, and children completed the risk-taking questionnaire and the behavioral control scale.

 

Data analysis

Parental control scales were subjected to reliability analysis and normality testing.  Correlations between the ADHD, risk-taking and parental control scores were examined. Finally, direct and indirect effects of ADHD symptoms on risk-taking behavior were calculated using the multiple mediation approach and SPSS 21.0 macro (PROCESS, Model 6) provided by Hayes. Following procedures recommended by Preacher and Hayes, a multiple mediation model involves (a) an analysis of the total indirect effect – the aggregate mediating effect of all the mediators being examined and (b) an analysis of the specific indirect effect – the mediating effect of a specific mediator. The significance of the indirect effects was tested via bootstrap analysis, which is commonly performed in multiple mediator analyses given its advantage of greater statistical power without assuming multivariate normality in the sampling distribution, assuming only the sample is representative of the population. Mediation is demonstrated via a statistically significant indirect effect (i.e., if the 95% bias-corrected confidence interval for the parameter estimate does not contain zero).

 

Results

Preliminary analysis

Table 1 presents descriptive statistics for the sample. Reliability of the ADHD and parental knowledge scale was in the acceptable range, whereas reliability of family disengaged relationship and excessive behavioral autonomy scales was questionable. None of the total scores withstood Kolmogorov-Smirnov normality testing, and therefore non-parametric statistics was used to describe the sample and analyze the correlations among all variables. Older participants were found to report higher RTB and consequently, age was used as a covariate in the mediation analysis.

 

Correlational analysis

Non-parametric correlation analysis of the primary study variables is presented in Table 2. As expected, level of ADHD symptoms positively correlated with level of engagement in RTB, and negatively correlated with level of parental knowledge. In addition, level of parental knowledge negatively correlated with engagement in RTB. Family disengaged relationship and excessive behavioral autonomy correlated significantly with each other, and disengaged relationship also correlated with higher RTB.

 

Mediation analysis

As can be seen in Figure 1, the bootstrap procedure indicated direct and indirect effects for the contribution of ADHD symptoms on RTB through its effects on parental variables. Together the model accounted for 32.3% of the variability in risk behavior (P<.0001). The regression coefficient between ADHD symptoms and RTB before considering mediators was statistically significant (P<.005). The regression coefficient between ADHD symptoms and parental knowledge, was statistically significant as well (P<.005). The bootstrapped standardized indirect effect mediated by parental knowledge was significant. Calculation of effect ratio (indirect effect divided by total effect) revealed that the indirect effect accounted for 34.3% of ADHD’s contribution to RTB. The indirect effects of ADHD symptoms, mediated by family disengaged relationship and excessive behavioral autonomy were not significant. ADHD symptoms still predicted risk taking after

accounting for the indirect effect (see Table 3 for coefficients and CIs).

 

Discussion

The current study assessed the relationship between adolescents’ ADHD symptoms, parental monitoring and RTB. The results of the study indicated both direct and indirect relationship between ADHD symptoms and RTB. The results suggest that while ADHD symptoms account for significant variability in adolescents’ RTB, this relationship is partially mediated by the parents’ knowledge on the child’s behaviors. The path we suggest from ADHD to RTB, therefore views ADHD symptoms as contributing to more reckless and risky behaviors but also in reducing parental monitoring that could serve as a protective mean.

ADHD symptoms and risk-taking behavior

Parents’ report of ADHD symptoms predicted youth self-reported engagement in a variety of RTBs. These findings are consistent with studies documenting increased risk taking by people with ADHD. However, most of the studies compared groups of participants with and without clinically diagnosed ADHD. Only seldom did researchers investigate the relation between risk-taking behavior and ADHD that was defined dimensionally as a continuum of symptoms. For example, Kollins et al. found that each reported inattention and hyperactivity/impulsivity symptom significantly increased the likelihood of regular smoking, and Pingault et al. (2013) reported on a prospective population cohort that inattention predicted nicotine dependence.

The current study, as well as a recent study that examined RTB and ADHD symptoms in the general adult population, extend the ADHD-related cigarette smoking into RTB in general. In addition, most of the studies reporting increased RTB in subjects with ADHD have focused on specific risk behaviors (see Nigg (2013) for review), and only rarely, studies approached RTB as a general tendency. Recently, our group used domain-general risk-taking behavior scales to reveal ADHD-related increased risk-taking behavior. The present study supports these findings, extending them to the adolescents’ population. 

The results of our study further indicated that ADHD symptoms level predicted RTB beyond parental monitoring. This finding is in line with previous research that linked between ADHD core deficits (such as deficient attention, executive, motivational and affective functioning and risk taking or reckless behavior. It suggests that ADHD symptoms may increase the risk of RTB for youth even when the family dynamic is relatively intact and the parents attempt to monitor the behavior of the child. 

Parental knowledge mediates ADHD-related RTB

In addition to the direct link between ADHD symptoms and RTB, the results of the current study provide further support to the central role parent-child relationship plays in the behavioral consequences of ADHD. The model supported by our path analysis suggests that ADHD symptoms contribute to lower parental knowledge, which in turn leads to increased engagement in RTB. This interpretation of the data is in agreement with findings regarding lower parental monitoring in families of children with higher level of ADHD symptoms.

Stattin and Kerr  have emphasized that the main source for parental monitoring is child disclosure, the voluntary act of sharing information with parents, and not parental solicitation and active surveillance. Stattin and Kerr’s work expanded the concept of parental monitoring to include a relational aspect. As such, parental monitoring is based on the willingness of the adolescent to share details about his life with his parent, and the quality of the relationship that promotes such sharing and disclosure. The negative impact of ADHD symptoms on parental knowledge indicated in our study suggests that ADHD symptoms may contribute to a parent-child dynamic where personal information about the youth’s life is less shared. This finding is consistent with longitudinal studies indicating the negative effect of ADHD symptoms on parenting practices and on effective parenting. In relation to RTB, the negative impact of ADHD symptoms on parenting may reduce the parents’ effectiveness in preventing engagement in RTB, or in setting effective consequences to RTBs that could lower its reoccurrence. ADHD, as indicated in our study, affects child-parent relationship to the point that this relationship does not protect against risk-taking behavior any longer. In that sense, RTB can be considered a “complication” of the primary ADHD symptoms.

 

Limitations

It is important to note however that the relation between child disclosure and RTB are bi-directional, and few longitudinal studies have shown that youth behavior problems predict less child disclosure and parental knowledge overtime. The design of the current study is cross-sectional, and therefore such alternative interpretation cannot be ruled out. Future longitudinal studies will be necessary to determine directionality between ADHD symptoms, parental knowledge and RTB.

Other limitations refer to the validity and reliability of the measures. Engagement in risk-taking behaviors and parental monitoring were assessed using adolescents’ self-report, which was not validated by collateral report. Therefore, there is a possibility that the reported child and parenting behaviors were biased and did not represent the actual RTB the youth engaged with or the actual parental knowledge level. Additional limitation is related to the family disengagement and excessive behavioral autonomy scales that achieved a questionable reliability in our sample, and therefore their role in mediating ADHD-related risk-taking behavior may have been obscured.

 

Clinical implications

The current findings, if replicated, may have important clinical implications. First, it stresses the role of ADHD symptoms beyond parenting components as significant contributors to RTB. Second, the results highlight the role of parental knowledge in preventing adolescent risk-taking behaviors, and steer out an important link between ADHD symptoms and parental knowledge. Clinically, our model suggests two links as targets for intervention to reduce RTB among adolescents with ADHD, namely, decreasing ADHD-related symptoms and improving parental knowledge. Given the chronic nature of ADHD symptoms and the clinical challenges to reduce ADHD core symptoms, improving the parent child relationship and the parents’ capacity to communicate with the child about his whereabouts may serve as an important and effective target for clinical interventions.

 

Acknowledgments

The study was funded by the Mirsky Foundation. The authors want to thank Ms. Moran Yakobi, Ms. Tzlil Zaguri and Ms. Zoya Vainshtein for their excellent help in running the study.