Social and emotional difficulties in children with ADHD and the impact on school attendance and healthcare utilization

Attention-deficit/hyperactivity disorder (ADHD) is a common neuropsychiatric condition in children[1–5] with an estimated prevalence of 3 to 7%[1]. Attention-deficit/hyperactivity disorder is characterized by symptoms of inattention and/or hyperactivity-impulsivity that are more frequently displayed and more severe than typically observed in individuals at a comparable level of development[1], are usually evident in more than one setting (e.g., home and school), and result in impairment in multiple domains of functioning[3, 6, 7]. A rich literature speaks to the burden that ADHD imposes on patients, families, and society as a whole, including negative effects on individual educational[8, 9] and social outcomes[3, 6], negative effects on patient and parent quality of life[7], and increased utilization of and spending on healthcare services[10–17].

Social and emotional difficulties are particularly common and problematic in children with ADHD. Social difficulties present in a variety of forms and can lead to conflicts with family and problems with peers[18–21]. Emotional difficulties often include poor emotional self-regulation, aggression, and reduced empathy[22, 23]. It should be noted that these challenges exist on a continuum. Relatively mild difficulties may fail to come to clinical attention, while in other cases such difficulties can contribute to overt, physician-diagnosed, comorbid mental health disorders, including anxiety, depression, and conduct disorder[22–24]. Major depressive disorder has been reported to occur in 12-50% of children with ADHD in community samples[25–27], and anxiety disorder, established by formal diagnostic interview, was comorbid in one third of ADHD patients enrolled in the commonly cited Multimodal Treatment Study of Children With ADHD[28]. Comorbid mental health conditions, including anxiety and depression, are extremely typical among children with ADHD and have been shown to be associated with greater functional impairment and worse educational outcomes[29–33].

Given the above data, it is important to understand the ways in which co-occurring conditions, including those characterized as social and emotional difficulties, can lead to various types of poor outcomes and functional impairment in children with ADHD, so that caregivers and providers can target interventions appropriately. In this study, we used data from the United States (U.S.) National Health Interview Survey (NHIS) to explore the association between social and emotional difficulties in children with ADHD and select outcomes. Available measures included both parent report of social and emotional difficulties (the brief version of the Strength and Difficulties Questionnaire [SDQ]) and parent report of physician-diagnosed depression, anxiety, and phobias. Unfortunately, teacher ratings and physician diagnoses were not available, and thus independent validation of parent reports was not possible. Available outcomes of interest included school days missed and emergency room (ER) and healthcare provider (HCP) visits over the past 12 months. We hypothesized that the presence of social and emotional difficulties in children with ADHD would be associated with increased school absenteeism and increased healthcare utilization, compared to ADHD children without these difficulties.

Overall, more ADHD children experienced at least 6 HCP visits (31%), compared with experiencing at least 2 ER visits (11%) and missing more than 2 weeks of school (8%). When assessing the impact of co-occurring social and emotional difficulties on school attendance and healthcare utilization, ADHD children with anxiety had significantly greater odds of missing more than 2 weeks of school (3.4 [2.2, 5.1]), having at least 2 ER visits (2.1 [1.2, 3.6]), and having at least 6 HCP visits (2.9 [2.0, 4.4]), compared with those without anxiety. For ADHD children with depression, those with the comorbid condition were 10 times as likely as those without the comorbid condition to miss more than 2 weeks of school (10.1 [5.7, 17.8]); 7 times as likely to have at least 6 HCP visits (7.4 [4.3, 12.7]); and 3.5 times as likely to have at least 2 ER visits (3.5 [2.0, 6.4)]. Similarly, ADHD children with comorbid phobias were 10 times as likely to miss more than 2 weeks of school (10.4 [4.2, 26.2]) as those without phobias, while being 3 times as likely to have at least 6 HCP visits (3.0 [1.3, 7.2]) and 2 times as likely to have at least 2 ER visits (2.4 [1.0, 5.4]). In addition, with each incremental increase in internalizing burden, ADHD children had significantly greater odds of missing at least 2 weeks of school (3.1 [2.4, 4.0]), having at least 6 HCP visits (2.2 [1.7, 2.8]), and having at least 2 ER visits (1.7 [1.3, 2.2]).

For the single, general SDQ item assessing difficulties in emotions, concentration, behavior, or being able to get along with other people (item 6), ADHD children with at least one of these complications experienced significantly greater odds of missing more than 2 weeks of school (4.4 [2.8, 6.9]), experiencing at least 2 ER visits (3.0 [1.8, 5.0]), and having at least 6 HCP visits (3.8 [2.6, 5.4]), compared with those who did not have these difficulties.

For the SDQ items associated with emotional difficulties (items 2 and 3), ADHD children who were worried had significantly higher odds of missing more than 2 weeks of school (3.2 [2.1, 4.8]), experiencing at least 2 ER visits (2.6 (1.4, 4.7]), and having at least 6 HCP visits (2.2 [1.5, 3.1]), compared with those who were not worried. Likewise, ADHD children who were unhappy/depressed experienced significantly greater odds of missing more than 2 weeks of school (3.9 [2.3, 6.4]), experiencing at least 2 ER visits (2.2 (1.3, 3.8]), and having at least 6 HCP visits (2.6 [1.7, 3.8]), compared with those who were not unhappy/depressed.

For the SDQ items associated with social or behavioral symptoms of ADHD (items 1, 4, and 5), children who did not have good attention were about 3 times as likely as children who did have good attention to miss more than 2 weeks of school (2.9 [1.8, 4.6]) and 2.5 times as likely to have at least 2 ER visits (2.5 [1.6, 4.1]), while experiencing at least 6 HCP visits was similar for those with and without good attention (1.5 [1.0, 2.4]). As observed with good attention, ADHD children who were not well behaved had significantly higher odds of missing more than 2 weeks of school (5.5 [2.2, 13.8]) and experiencing at least 2 ER visits (5.2 (2.0, 13.5]), compared with those who were well behaved, while the odds for having at least 6 HCP visits were similar between those who were well behaved and who were not well behaved (1.4 [0.7, 2.9]). ADHD children who got along better with adults experienced similar odds of missing more than 2 weeks of school (0.7 [0.4, 1.3]), having at least 2 ER visits (0.9 (0.5, 1.6]), and having at least 6 HCP visits (1.0 [0.6, 1.5]), compared with those who did not get along better with adults.

When assessing if there was a differential effect of comorbid condition on missed school days and healthcare utilization by ADHD status, three interactions were significant. As stated above, children with ADHD who got along better with adults had lower odds, although not significant, of missing more than 2 weeks of school compared with ADHD children who did not get along better with adults (0.7 [0.4, 1.3]). Conversely, non-ADHD children who got along better with adults experienced significantly higher odds of missing more than 2 weeks of school (1.8 [1.2, 2.7]), compared with those who did not get along better with adults. This diametric relationship resulted in a significant interaction effect (P=0.0490). Significant interactions were also observed for ADHD status and being well behaved (P=0.0060), as well as being worried (P=0.0420), for children experiencing at least 2 ER visits. ADHD children who were not well behaved had significantly greater odds of having at least 2 ER visits, compared with those who were well behaved (5.2 [2.0, 3.5]), while non-ADHD children who were not well behaved had lower odds of having at least 2 visits to the ER, compared with non-ADHD children who were well behaved (0.5 [0.2, 1.2]). On the other hand, both ADHD and non-ADHD children who worried experienced increased odds of having at least 2 ER visits; however, the comparison was significant for the ADHD cohort (2.6 (1.4, 4.7]) and was not significant for the non-ADHD group (1.2 [0.9, 1.7]).

This study adds to the literature which demonstrates that social and emotional difficulties in children with ADHD can contribute to higher rates of unfavorable outcomes. In particular, these data suggest that both parent-observed child social difficulties (e.g., not being “well behaved”) and emotional difficulties (e.g., worry) and parent report of physician diagnosed affective disorders (e.g., depression) can be used to identify children with significantly elevated rates of school absenteeism and ER and HCP utilization. Strikingly, a positive response on a single general item from the SDQ (i.e., item 6, “had difficulties in any of the following areas: emotions, concentration, behavior, or being able to get along with other people”) identifies a subset of children 3 to 4 times as likely as peers answering negatively, to exhibit all three of the examined adverse outcomes. While this general association is compelling, consideration of the other independent measures provides additional insights. The remaining eight items examined can be organized according to the clinical/psychological domain to which they speak: Three to anxious symptoms (i.e., the SDQ “worry” item and the physician-diagnosed “anxiety or stress” and “phobias or fears” items); two to mood (i.e., the SDQ “unhappy/depressed” item and physician-diagnosed “depression”); and three to core ADHD symptoms or social behavior (i.e., the “well behaved,” “good attention span,” and “got along better with adults” SDQ items).

In general, the presence of anxious symptoms had a more pronounced impact on school absenteeism than on ER or HCP utilization. While no further detail as to the nature of the anxiety was available, anxiety-related school avoidance is a well described phenomenon, and, in this regard, it is notable that of the three items, the strongest relationship with school absenteeism was observed for physician-diagnosed phobias (OR 10.4). It is also interesting that parent observation of “worry” (SDQ item 2) was as strongly predictive of increased absenteeism as was report of physician-diagnosed anxiety (OR 3.2 vs. 3.4, respectively).

Consistent with what has been observed in other studies, parent report of physician-diagnosed depression was associated with worse outcomes[7, 41], and it predicted the largest increase in odds of more HCP visits, across all items examined. This result may be due, in part, to the fact that depression is more likely to lead to closer physician follow-up, greater use of pharmacotherapy, and higher rates of specialist referral relative to children with anxiety disorders or phobias, which are generally managed through behavioral therapies. In contrast to the pattern seen with the anxiety items, physician-diagnosed depression was associated with substantially greater odds of both increased school absenteeism (OR 10.1 vs. 3.9) and HCP visits (OR 7.4 vs. 2.6) than was the parental report of a child being “often unhappy, depressed or tearful” (SDQ item 3). These findings suggest that the 16.5% of ADHD children with physician-diagnosed depression are likely a more severely affected subgroup of the 27.4% of children rated positive on SDQ item 3.

The impact of the remaining SDQ items on the outcomes of interest was mixed. A negative response on SDQ item 1 (i.e., “was well behaved, usually did what adults requested”) was actually the strongest predictor of multiple ER visits across all items, and the strongest among SDQ items of school absenteeism; in contrast, poor attention span (SDQ item 5) was more weakly associated with these outcomes. This result is consistent with the fact that children with predominantly inattentive forms of ADHD are more likely to exhibit more subtle problems (e.g., school failure) than their more declarative peers with hyperactivity. Finally, a child’s getting along “better with adults than with other children” (SDQ item 4) did not appear to be associated with any of the outcomes examined, perhaps because of the ambiguous nature of the question (i.e., could be interpreted as a positive or negative attribute). This outcome is further reinforced by the significant interaction between ADHD status and this item in the models that predict school absenteeism (P=0.049). For non-ADHD children, those who got along better with adults tended to miss more school than those who did not; while amongst children with ADHD, those who got along better with adults tended to miss less school. One possible explanation for this finding is that, among children with ADHD, peer rejection is the norm[42]; thus, the ability to “get along better with adults than peers” may indicate positive relationships with teachers in a formalized setting. In contrast, this trait may reflect interpersonal or social deficits in children without ADHD that are associated with increased problems at school.

The use of nationally representative survey data from NHIS represents a particular study strength, permitting generalization of findings to the entire U.S. population of children with ADHD. The sampling design enhances validity by ensuring that participants are selected for inclusion in the study independent of their status for the predictor and outcome variables of interest. Another study strength is the use of the brief SDQ, which has been shown to be a reliable and valid screening instrument for child psychiatric disorders[43, 44].

Our findings should be interpreted in light of several important considerations regarding the measures available within the NHIS survey. First, relatively few items were available to assess emotional and social difficulties, and indeed no information was available regarding the duration or severity of these problems. Furthermore, measures of both emotional and social difficulties and of the outcomes of interest were based on parent report. The use of direct parent report measures of both independent and dependent measures is both a strength and weakness. On the positive side, it permits the collection of data elements that are not available in secondary sources, such as administrative claims. On the negative side, parent report data are subject to recall bias and are necessarily inferior to school attendance and healthcare claims records for the outcomes of interest. In other studies using the SDQ, investigators have reported greater validity and reliability of estimates of emotional and behavioral problems based on reports from multiple informants including parents, teachers, and, for some age groups, children[44, 45]. As parallel assessments from these sources were not available, we were unable to independently verify parents’ assessments. Finally, it should be noted that the thresholds chosen when dichotomizing outcome measures were somewhat arbitrary; these choices are in no way intended to imply that school absences or healthcare utilization above these thresholds were unnecessary or inappropriate in any way.

Several additional study limitations deserve mention. The definition of ADHD status was based on the parent response to a single item. While the validity of this approach is suspect, it should be noted that the prevalence of ADHD in this sample is very close to what would be expected based on estimates from other studies[2, 3] and the prevalence reported in the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV)[1]. Information about the presence of ADHD and mood disorders in parents was also unavailable. Thus, while these factors could clearly serve as an important source of bias, we were unable to examine their impact on the outcomes assessed. Also, a substantial proportion of the sample (9.6%) was missing data on the income variable and, therefore, those subjects were not included in the analysis. A sensitivity analysis was performed using multiple imputations, and the results were consistent with those presented for the non-imputed samples. It is important to note that the sample size was considerably smaller for ADHD children with a reported diagnosis of depression and phobias, as well as for ADHD children who were not well behaved compared to ADHD children without these social and emotional difficulties. Given this information, the comparisons for these particular difficulties should be interpreted with caution. The NHIS also has very limited information regarding the medications the children were taking at the time of the study and, thus, the analyses did not control for medication status, types of medications, or medication adherence. Finally, the current study reports on the results of a large number of statistical tests in which the P-values were not adjusted for multiple comparisons to control the type I error rate. This study was intended to generate hypotheses rather than confirm specific hypotheses. Based on all of the aforementioned limitations, these results would need to be replicated in future studies.

Our findings provide further evidence that the presence of social and emotional difficulties in children with ADHD contributes to the functional impairment observed in this population. In particular, children manifesting these problems are more likely to experience greater school absenteeism and to incur more ER and HCP visits than their unaffected peers. Greater awareness of these associations, together with focused efforts to identify and manage these children appropriately, could lead to improved patient outcomes (e.g., improved school attendance) and to decreased healthcare utilization.