Sleep Problems in Children With Attention-Deficit/Hyperactivity Disorder: Title and subTitle BreakPrevalence and the Effect on the Child and Family FREE

Although sleep problems are common in children with attention-deficit/hyperactivity disorder (ADHD), their effect has received very little attention. This is a major omission. Attention-deficit/hyperactivity disorder is the most common mental health disorder in childhood, affecting up to 11% of Australians aged 6 to17 years.¹ Sleep problems are often readily treatable in the general pediatric population without medication, so—if shown to make a major contribution to morbidity—strategies for effective sleep management could potentially have important implications in reducing adverse outcomes of ADHD.

Up to 50% of parents of children with ADHD report difficulties with their children's sleep, including difficulties initiating sleep (delayed onset sleep or bedtime resistance) and maintaining sleep (frequent nocturnal awakenings or restlessness).^{2 - 7} Daytime sleepiness, tiredness on waking, and nightmares are also more common in children with ADHD than controls.^{2 - 10} Objective studies using polysomnography, an actigraph, and video monitoring have shown that children with ADHD have increased sleep latency, decreased rapid eye movement sleep percentages, and increased nocturnal activity.^{5 ,11} Other sleep problems identified in children with ADHD include sleep-disordered breathing (≤ 25%)¹² and restless legs syndrome or periodic limb movement disorder (≤ 36%).^{3 ,13 - 14} Some studies suggest that parents of children with ADHD overreport sleep problems, while other studies show underreporting when compared with objective studies.^{4 ,6 ,14} Physiological differences in the sleep of children with and without ADHD are also unclear.¹⁵

Compared with children without ADHD, those with ADHD have greater difficulties in behavioral, social, and academic functioning^{16 - 17} and poorer quality of life (QOL).^{18 - 19} Their parents experience more parenting stress,^{16 - 17} and their mothers are more likely to report symptoms of anxiety and depression.^{16 ,20} Families of children with ADHD are more likely to report adversely on family activities and parental emotions than families of children without ADHD.^{17 ,19} However, many of these same symptoms are also identical to those noted in children with sleep problems, with or without ADHD. Children with sleep problems have been shown to have more behavioral, emotional, and social problems and have more difficulties with school attendance.^{21 - 23} Children with insufficient, fragmented, or poor-quality sleep have increased impulsivity, hyperactivity, and aggressiveness as well as problems with mood, academic performance, and neurocognitive functioning.^{21 - 24} It has thus been postulated that children with ADHD and sleep problems could have poorer cognitive and behavioral outcomes than children with ADHD alone.^{3 ,5 ,7 ,14}

However, no study has examined the associations between sleep problems and broader effects on the child with ADHD (eg, daily functioning and school attendance), the primary caregiver (eg, mental health and work attendance), and/or the family (eg, emotional and time impact and family activities). Clinically, each of these areas is commonly affected by ADHD and sleep problems in children, but the effect of sleep problems beyond ADHD alone is unknown. Similarly, it is unknown how effectively pediatricians identify and manage sleep problems in children with ADHD. Given that pediatric sleep problems, at least in children without ADHD, are often amenable to brief behavioral interventions,²⁵ quantifying their effect is an important first step toward developing a systematic approach to the identification and treatment of sleep problems in children with ADHD.

We therefore aimed to determine, in a clinical sample of schoolchildren with clinician-diagnosed ADHD, the prevalence of sleep problems and their associations with (1) child health-related QOL, daily functioning, and school attendance, (2) primary caregiver mental health and work attendance, and (3) family impacts. We hypothesized that sleep problems would be common in children with ADHD and would be associated with poorer outcomes in each of these areas. We also aimed to determine whether pediatricians routinely ask about children's sleep, and, if a sleep problem is identified, whether caregivers consider that effective treatment was offered.

Primary caregivers of Victorian schoolchildren with ADHD (aged 5 to 18 years) completed a survey on their children's sleep and behavior between May and November 2006. The children's pediatricians or child psychiatrists diagnosed their ADHD. Families were systematically recruited through (1) behavioral outpatient clinics at the Royal Children's Hospital, Melbourne, Australia (the main tertiary pediatric center for the state of Victoria, population 4.6 million), (2) 8 pediatricians in private practice (selected because we knew they had an interest in children with ADHD), and (3) ADHD support groups. Families with insufficient English to complete the survey were excluded.

Eligible families, preidentified by an independent research assistant (at outpatient clinics and at ADHD support groups) or by the private pediatricians, received a written invitation to participate. Those who did not opt out were approached by telephone for informed consent. Surveys were then mailed to interested families. The Ethics in Human Research Committee at the Royal Children's Hospital approved the study.

The main outcome was the primary caregiver's report of the child's sleep problems during the previous 4 weeks. The primary caregiver was asked “During the past 4 weeks, has your child's sleep been a problem?” Primary caregivers selected from the following options: (1) no, it was not a problem, (2) yes, a mild problem, (3) yes, a moderate problem, or (4) yes, a severe problem. Responses were trichotomized into none vs mild vs moderate or severe sleep problems.²⁶ Caregivers who reported that their children had mild to severe sleep problems were asked “During the past 4 weeks, which of the following sleep pattern/s has/have been a problem for your child?” Problematic sleep patterns recorded included difficulty falling asleep, resisting going to bed, tossing or turning in bed, snoring or difficulty breathing, waking up frequently during the night, difficulty getting up in the morning, and tiredness on waking.²⁶

Secondary outcomes included child, caregiver, and family measures. Child health-related QOL was measured by the Pediatric Quality of Life Inventory 4.0 (PedsQL),²⁷ a 23-item validated questionnaire that provides total, physical, and psychosocial health summary scores, with higher scores indicating better health-related QOL. Child daily functioning was measured by the Daily Parent Rating of Evening and Morning Behavior.²⁸ This is an 11-item scale in which caregivers rate the core ADHD symptoms and behavioral problems typically experienced by children with ADHD during the past 4 weeks, with higher scores indicating poorer daily functioning. The primary caregiver reported on the number of days his or her child missed or was late for school in the previous 6 months. Primary caregiver mental health was measured by the Depression Anxiety Stress Scale,²⁹ a 21-item measure in which adults rate symptoms experienced during the previous 4 weeks. Clinical cutoffs for each of the 3 subscales were calculated, with scores greater than 9 indicative of depression, scores greater than 7 indicative of anxiety, and scores greater than 14 indicative of stress. Also reported were the number of days the caregiver and/or his or her spouse missed or were late to work in the previous 6 months.

Family impacts were assessed using the Emotional Impact, Time Impact, and Family Activities subscales of the Child Health Questionnaire (CHQ) Parent Form 50, with higher scores indicating better functioning.^{30 - 31} Finally, caregivers reported on pediatrician identification of sleep problems during the previous 6 months and, if applicable, advice offered and usefulness of this advice.

The child's age, sex, birth order, comorbid behavioral and psychiatric problems (learning difficulties, anxiety or depression, oppositional defiant disorder, conduct disorder, and autism spectrum or Asperger disorder), and medication use were also sought. The child's ADHD symptom severity was identified as a potential confounder^{18 ,32} and was measured by the ADHD Rating Scale,³³ an 18-item scale based on Diagnostic and Statistical Manual of Mental Disorders (Fourth Edition)³⁴ criteria measuring frequency of core ADHD symptoms during the previous 4 weeks; higher scores indicate greater severity. Sociodemographic data included caregivers' age, educational and employment status, and the census-derived Socio-Economic Indexes for Areas Index of Relative Disadvantage³⁵ assigned to each family as a measure of social and economic welfare based on residential postal code. Higher scores indicate less disadvantage.

Prevalence of no, mild, and moderate or severe sleep problems and of problem sleep patterns in the children were calculated. Differences between outcomes of interest were then calculated for children with no sleep problems vs those with mild sleep problems and those with moderate or severe sleep problems. Linear regression was used to estimate the mean differences between the referent group of no sleep problem and the mild sleep problems and moderate or severe sleep problems group for quantitative outcomes (ie, PedsQL total and physical/psychosocial summary scores, Daily Parent Rating of Evening and Morning Behavior scores, and CHQ subscale scores). Logistic regression was used to estimate the odds ratios for a mental health problem in the primary caregiver (measured by Depression Anxiety Stress Scale cutoffs) between the referent category of no sleep problem and the mild sleep problems and moderate or severe sleep problems category. Analyses were then adjusted for potential confounders^{18 - 19} determined a priori, including children's age, sex, presence of comorbidities (yes or no), medication use (yes or no), severity of ADHD symptoms, and the primary caregiver's educational status (< high school, high school only, or > high school). Finally, a post hoc analysis was conducted to determine which comorbidity was contributing to confounding the relationship between sleep problems and the CHQ subscale scores, controlling for all other comorbidities. All analyses were conducted using Stata, version 9 (Stata Corp, College Station, Texas).

tj;3Of the 330 families eligible for recruitment, 251 were contactable by telephone (of whom, 197 [78%] returned their questionnaire, 45 did not, and 9 declined to participate), and 79 were not contactable but were sent a questionnaire by mail (of whom, 42 [53%] returned it). The overall response rate was 74% (239 of 321 mailed). The mean Socio-Economic Indexes for Areas Index of Disadvantage (for the general population in the state of Victoria, range, 706.95-1151.74; mean, 1020.27 [SD, 59.41]) was higher for responders than nonresponders (1043.26 [SD, 74.78] vs 1022.54 [SD, 61.23], respectively, P = .01). Sample characteristics are detailed in Table 1. In general, most children were male (90.4%) and lived in a 2-caregiver household (76.6%).

Sleep problems were extremely common, affecting 73.3% (n = 175) of all participants. The prevalence of mild sleep problems was 28.5% (n = 68) and moderate or severe sleep problems was 44.8% (n = 107). Table 2 lists the prevalence of different sleep patterns in the mild vs the moderate or severe sleep problem groups. Difficulty falling asleep, resisting going to bed, and tiredness on waking were the most common sleep patterns encountered. Forty-five percent (n = 108) of caregivers reported that their pediatricians had asked about their children's sleep and, of these, 65 (60%) reported receiving advice, which 44 caregivers (68% of 65, or 25% of those reporting sleep problems) rated as helpful.

Children with mild and moderate or severe sleep problems had lower PedsQL total and psychosocial scores than those without sleep problems, and this relationship held true after adjusting for confounders (P < .001) (Table 3). Children with moderate or severe sleep problems were also more likely than those without to miss school and be late for school during the past 6 months, with the prevalence of each increasing as sleep problem severity increased (Table 4).

Severity of ADHD symptoms was strongly associated with both moderate or severe sleep problems and physical QOL (the physical PedsQL subscale score reduced by 0.8 points for every 1 unit increase in ADHD Rating Scale score; coefficient for ADHD Rating Scale, − 0.8; 95% confidence interval [CI], − 1.31 to − 0.37). Because of the cross-sectional nature of the data and lack of a priori information, it was difficult to know whether severity of ADHD symptoms should be treated as a confounder for the relationship between sleep problems and their putative adverse outcomes in ADHD children or as a putative outcome in its own right; conservatively, we chose to treat it as a confounder in the main analyses. However, when we reran analyses without ADHD symptom severity as a confounder, we found that all relationships between mild and moderate or severe sleep problems and outcomes strengthened, with the relationships between moderate or severe sleep problems and family outcomes and physical PedsQL scores all achieving statistical significance (P ≤ .04). We also found that moderate or severe sleep problems strongly predicted symptom severity (coefficient, 2.64; 95% CI, 1.08-4.21) after adjusting for all other confounders.

Primary caregivers of children with ADHD who have moderate or severe sleep problems were more likely to have poorer mental health than those without sleep problems (P ≤ .01) (Table 3). After adjusting for confounders, a primary caregiver of a child with moderate or severe sleep problems was 2.7 times more likely to be clinically depressed, stressed, or anxious than if the child had no sleep problem (odds ratio [OR], 2.72; 95% CI, 1.33-5.54). An education beyond high school was protective (OR, 0.38; 95% CI, 0.20-0.70). There was a higher percentage of primary caregivers of children with moderate or severe sleep problems who were late for work compared with those without sleep problems, and their spouses were more likely to be late to work (Table 4).

Families of children with moderate or severe sleep problems had poorer scores on all 3 measures, but this relationship was not significant after adjusting for confounders (Table 3). The presence of 1 or more comorbidities contributed to confounding for all 3 CHQ subscale scores. Of the comorbidities, conduct disorder had the greatest effect on CHQ Time Impact (coefficient, − 21.87; 95% CI, − 40.47 to − 3.28) and CHQ Family Activities (coefficient, − 20.67; 95% CI, − 35.65 to − 5.69) subscale scores, while learning difficulties and autism spectrum or Asperger disorder had the greatest effects on CHQ Emotional Impact subscale scores (coefficient, − 8.65; 95% CI, − 15.27 to − 2.03; and coefficient, − 9.79; 95% CI, − 17.49 to − 1.91), respectively.

Sleep problems were extremely common in this sample of schoolchildren with ADHD, affecting nearly three-quarters. Moderate or severe sleep problems were strongly associated with poorer child psychosocial QOL and daily functioning, primary caregiver mental health, family functioning, and ability to participate in daily responsibilities outside the home (children being late for school and their caregivers being late to work). However, less than half of caregivers reported that their pediatricians had asked about their children's sleep problems.

Our study is the first to quantify the associations between sleep problems in children with ADHD and child, caregiver, and family functioning, taking into account potential confounders, such as ADHD severity and comorbidity. It is also the first to identify pediatrician management of sleep problems in children with ADHD. We had a good response rate and used validated measures for most outcomes.

Our study has some limitations. First, the data were cross-sectional and thus causality cannot be inferred. Second, the diagnosis of ADHD had been made by a clinician, not a standardized diagnostic interview and, as such, some children's disease may have been incorrectly diagnosed. It is possible that some children may have had a primary sleep disorder, which would result in an overestimation of the relationship between sleep problems and outcomes. Third, although validated outcome measures were used, the sleep measures were based on subjective caregiver reports. However, parental report has been shown to be a reliable indicator of problem sleep patterns in children compared with overnight infrared camera recording³⁶ and has shown robust associations with adverse outcomes in the general population.²⁶ Although there are validated measures of sleep in children, we considered these either too long³⁷ or not valid for our study's age range.³⁸ Studies have suggested that parents of children with ADHD overreport sleep problems compared with parents of children without ADHD,^{4 ,6 ,14} but there are few studies directly comparing parental reports and objective measures of sleep in children with ADHD; 1 study that did (97 children with ADHD) found that parents underreported delayed sleep onset (43% of parents reported the problem; 94% was diagnosed on polysomnography) and overreported snoring and sleep-disordered breathing (63% of parents reported the problem; only 7% of cases were diagnosed on polysomnography).⁴ The discrepancies between parental report and objective measures may reflect the discrepancy between parental perceptions as to what constitutes a problem for a family vs a predefined clinical definition. Fourth, responders were more likely to come from higher socioeconomic groups than nonresponders, and we excluded non–English-speaking families. Thus, our results are most likely to generalize to English-speaking families of children with ADHD with less socioeconomic disadvantage. Finally, there was a preponderance of boys (90%) in our sample; however, this is in keeping with 1 of the largest studies of children with ADHD in which 80% of the sample were boys.³⁹

The prevalence of mild and moderate or severe sleep problems was 73%, higher than the 50% reported in a recent review of sleep and ADHD.³ It may be that inclusion of mild sleep problems raised the prevalence in our study. However, given their adverse associations, even mild sleep problems should not be ignored. The prevalence of difficulty falling asleep (71% in the mild group, and 84% in the moderate or severe group) was higher than the reported 48% in a US study⁴ and also higher than the 56% reported in another review.² The reason for this is unclear.

Two other studies have described the relationship between ADHD and health-related QOL in children, with results similar to ours. A Canadian study of 131 children with ADHD aged 6 to 17 years found that both ADHD symptom severity and the presence of 2 or more comorbidities contributed significantly to poorer psychosocial QOL.⁴⁰ In a randomized controlled trial of atomoxetine in 297 children with ADHD, ADHD symptom severity was also found to be associated with poorer psychosocial QOL.³² However, neither of these studies measured sleep. In our study, the presence of a sleep problem remained a predictor of poor psychosocial QOL even after accounting for ADHD symptom severity and comorbidities.

Consistent with our study, the Canadian study found that conduct disorder was associated with poorer family QOL as measured by the CHQ Time Impact subscale scores,⁴⁰ but, in contrast with our findings, learning difficulties were only weakly associated with the CHQ Emotional Impact subscale scores.

Primary caregivers of children with ADHD and sleep problems had poorer mental health than the caregivers of those without sleep problems. Although no other study has reported on the association between sleep problems and caregivers' mental health in children with ADHD, sleep problems in healthy children are a known risk factor for maternal depression and stress.⁴¹

In summary, sleep problems in schoolchildren with ADHD are extremely common and strongly associated with poorer QOL, daily functioning, and school attendance in the child and poorer caregiver mental health and work attendance. Implementation of a sleep intervention in children with ADHD could feasibly improve outcomes beyond treatment of ADHD alone. It is possible that such intervention could reduce the need for medication in some children. Additional research should include longitudinal analyses to determine the temporal direction of the relationship between ADHD and sleep problems and trials to determine the efficacy of routine sleep management. In the meantime, clinicians caring for children with ADHD should ask about their sleep, and if a problem is present, this should be addressed.

Correspondence: Valerie Sung, MBBS, Centre for Community Child Health, Royal Children's Hospital, Flemington Road, Parkville, Victoria, Australia 3052 (valerie.sung@rch.org.au).

Accepted for Publication: December 18, 2007.

Author Contributions: Dr Sung had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Sung, Hiscock, and Efron. Acquisition of data: Sung, Sciberras, and Efron. Analysis and interpretation of data: Sung, Hiscock, Sciberras, and Efron. Drafting of the manuscript: Sung, Hiscock, Sciberras, and Efron. Critical revision of the manuscript for important intellectual content: Sung and Hiscock. Statistical analysis: Sung, Hiscock, and Sciberras. Administrative, technical, or material support: Sung and Sciberras. Study supervision: Hiscock and Efron.

Financial Disclosure: Dr Efron has been a member of advisory boards for Eli Lilly and Novartis.

Additional Contributions: We particularly thank the following private and outpatient pediatricians who helped with recruitment: Chitra Chandran, MBBS; Katie Heathershaw, MD; Rick Jarman, MD; Lionel Lubitz, MBBS; Annie Moulden, MBBS; Anita Murphy, MBBS; Carl Orkin, MD; and Anne Turner, MBBS.