Sleep and behavioral/emotional problems in children: A population-based study

1. Introduction 

While inadequate/insufficient sleep promotes daytime sleepiness in adults, in children it could enhance inattentive/hyperactivity behaviors [1]. Indeed, poor sleep may cause prefrontal cortical dysfunction and interfere with executive functions, such as working memory, emotional regulation and behavioral inhibition, and therefore may explain the inattentive/hyperactive behaviors in school age children [1].

Parental reports of sleep problems are common in otherwise healthy children receiving routine pediatric care and may represent a “red flag” for the presence of underlying psychiatric, social, or medical problems [2]. For example, Smedje et al. [3] used parental responses to a sleep habits questionnaire and to a behavioral screening form (the Strengths and Difficulties Questionnaire – SDQ) to investigate potential associations between disturbed sleep and behavioral difficulties in 635 children aged 6 to 8 years. These investigators found that the presence of disturbed sleep was associated with behavioral problems and, more importantly, that there were specific associations between particular sleep-related disturbances and particular dimensions of behavior. Indeed, hyperactivity was associated with tossing and turning during sleep and sleep walking, conduct problems were related to bedtime resistance, and emotional symptoms were associated with night terrors, difficulty falling asleep and daytime somnolence [3]. In another study conducted in Finland on a cohort of 5,813 children between the ages of 8 to 9 years, significant associations between both parents and children reports of sleep problems and psychiatric symptoms at school were identified [4]. As with the preceding study, children with reportedly severe sleep problems were more likely to have emotional problems [odds ratio (OR): 2.74], school attendance problems (OR: 2.53), behavioral problems (OR: 2.44) and hyperactivity (OR: 2.02) than those with no or mild sleep complaints.

Considering that perceptual context regarding both sleep–wake patterns and emotional/behavioral problems may be influenced by cultural and social factors, we are unclear whether the published associations between sleep and behavior can be extrapolated to other populations. Therefore, the aim of this study was to investigate the relationship between sleep behaviors and emotional/disruptive problems in a large sample of Portuguese school-aged children.

2. Methods 

The data for this report were drawn from an epidemiologic survey of sleep and wake patterns in primary school children [5]. The project was reviewed and approved by the Regional Director of Education, and parental consent was obtained.

2.1. Survey measures 

2.1.1. Sleep–Waking Questionnaire 

A parental questionnaire to assess child sleep and waking behavior was developed for our epidemiologic survey [5]. The questions concerned sleep/wake times (school days, weekends) total sleep time (school days, weekends), time to fall asleep, number of night awakenings, ability to go back to sleep alone after night awakenings, bedtime habits and difficulties (fall asleep alone at his own bed, fall asleep in parents bed, parental presence to fall asleep, comforting habits to sleep, light on to fall asleep, willing to go to bed at bedtime, bedtime resistance), night time sleep related behaviors (loud snoring, wets bed during sleep, nightmares, walks in sleep, talks in sleep, night terrors, grinds teeth in sleep, fear of sleeping in the dark), daytime sleepiness, irritability and tiredness, factors influencing bedtime (needs to sleep enough for the next day activities, family routine, brothers/sisters bedtime, is sleepy, ends favorite TV program, other), and wake up practices in the morning (parents/family members, alarm clock, by himself/herself, noise, needs to use the toilet, other).

In addition parents were asked whether their child had any problem with sleep, whether they have sought professional help for a sleep problem in the child, about the use of sleep medication, and also to provide information on other medical disorders such as epilepsy, asthma, bronchitis, rheumatism, diabetes, mental disorder, mental retardation or others.

Most questions required responses rated on a 4-point scale from never (coded as 1) to always (coded as 4) or in a yes/no format (see questionnaire in appendix).

The Sleep–Waking Questionnaire’s validity and stability was studied on an independent sample of primary school age children within a one-month interval and the results obtained were found to be satisfactory [5].

To reduce the number of comparisons and avoid biases introduced by multiple statistical tests, a factor analysis using the principal component solution with varimax rotation was performed using the sleep questionnaire items (sleep duration, bedtimes/wake up times during school days/weekends, wake up practices and factors influencing bedtime were not included in this analysis) and a 6-component model was achieved that accounted for 63.4% of the variance. For inclusion in the model, only factors with eigenvalues greater than 1 were considered and only sleep questionnaire items with factor loadings ⩾.4 were retained (the rotated factor matrix is presented in Table 1).

Table 1. Rotated factor matrix of some of the sleep questionnaire items loadings

	Items/factors	Factor 1	Factor 2	Factor 3	Factor 4	Factor 5	Factor 6
	Parents help to sleep (TVE=12.3%)
	Settling to sleep alone	.888	.037	−.070	.135	.090	.031
	Need parents to fall asleep	.783	.037	−.154	.120	.119	.058
	Falls asleep in parents bed	.777	.139	.042	−.022	−.001	.015
	Parasomnias (TVE=11.1%)
	Night terrors	−.024	.793	−.045	.080	.051	−.037
	Sleep talking	.114	.683	−.067	.001	.087	.044
	Nightmares	.025	.617	−.170	.106	.063	.174
	Sleepwalking	.079	.515	−.038	−.014	−.015	.108
	Sleeping difficulties (TVE=10.5%)
	Professional help to sleep	.018	−.178	.788	−.022	−.157	.041
	Sleep problems	−.141	−.178	.765	−.078	−.127	−.015
	Sleep medication	−.048	.023	.699	.000	.222	−.098
	Afraid of dark (TVE=10.4%)
	Lights on to fall asleep	.060	.076	−.010	.930	.044	.003
	Afraid of sleeping in the dark	.150	.064	−.081	.910	.056	.072
	Sleep limit setting (TVE=9.7%)
	Bedtime resistance	.044	.069	−.046	.043	.848	.091
	Unwillingness to go to bed	.125	.056	.006	.052	.842	.030
	Daytime sleep consequences (TVE=9.4%)
	Sleepiness	.017	.023	.035	−.006	−.056	.810
	Tiredness	.021	.134	−.060	.074	.043	.806
	Irritability	.071	.162	−.060	.008	.287	.465

Item loadings for each factor are shown in bold. TVE=Total variance explained by the factor.

Factor 1 included three items that reflected the need for parents to intervene in promoting their child to fall asleep. Factor 2 was termed “Parasomnias,” because it included items such as night terrors, sleep talking, nightmares and sleep walking. “Sleeping difficulties” was the term used for Factor 3 and included items about the child’s sleep problems and the need for medication or professional help. Factor 4, which was termed “Afraid of dark” included items about the need of environmental light to fall asleep and the fear of sleeping in the dark. Factor 5 was termed “Sleep limit setting” factor because it included items about bedtime resistance and unwillingness to go to bed. Finally, Factor 6 included items reflecting daytime consequences of sleep disturbance, such as sleepiness, tiredness and irritability.

In addition to the six factors delineated above, the following variables were assessed as far as their potential association with the various behavioral sub-types: sleep duration during school days and weekends, bedtimes and wake times during school days and weekends, time to fall asleep, night awakenings, snoring, bed wetting, and the ability to go back to sleep alone.

2.1.2. Rutter scale B2 

The Behavior Questionnaire, originally developed by Rutter [6] for completion by teachers (Rutter Scale B2), has 26-items that address a child’s behavior at school. In addition, the teacher is asked for each of the items to rate whether it “certainly applies” (scored=2), “applies somewhat” (scored=1) or “doesn’t apply” (scored=0). A total deviance score is derived from the sum of scores for the individual items, such that cumulative scores may range from 0 to 52. In addition to the total score, three sub-scores may be obtained from the sum of the scores for certain items. The emotional or neurotic sub-score is the summation of the scores of four items (often worried, miserable, fearful, tears on arrival at school), the hyperactivity sub-score is gathered from the sum of scores of three items (restless/overactive, poor concentration, fidgety/squirmy) and the conduct or antisocial sub-score is obtained from the sum of scores of six items (destructive, fights, disobedient, lies, steals, bullies). Rutter found that a cumulative score of nine points or more on the total scale is suggestive of the presence of some psychiatric disorder. The Portuguese version of the Rutter Scale B2 used in this study has been validated and displays similar psychometric properties [7], [8].

In the present study the definition of behavioral problems followed the criteria of Almqvist et al. [9], i.e., children scoring ⩾9 points on the scale and higher points on emotional items than on conduct items were considered to have emotional problems (n=39; 5%). Children scoring ⩾9 points on the scale and 3 or more on hyperactivity items were considered to have hyperactive problems (n=114; 14.6%). Children scoring ⩾9 points on the scale and higher points on conduct items than on emotional items were considered to have conduct problems (n=96; 12.3%).

3. Results 

3.1. Sample characteristics 

Of the 1,381 children approached, 988 questionnaires were returned by the parents (response rate of 71.5%); from the questionnaires 16 children (1.6%) who were >11 years old were excluded to allow for improved homogeneity of the sample cohort. Teacher ratings were missing for 95 (9.6%) of the children, and 98 (9.9%) children who reported significant medical or neurological conditions (epilepsy, asthma, bronchitis, rheumatism, diabetes, mental disorder, mental retardation or others) that could affect sleep were also excluded from further analysis (82 of these were severe asthmatics, 83.7%). Thus, the final sample comprised 779 children (403 girls, 51.7%; 376 boys, 48.3%) with a mean age of 7.9±1.3 years (range: 6 to 11 years), and no age difference between boys and girls was observed. The majority of children were in the 6–7 years and 8–9 years age groups (88.4%), came predominantly from middle to lower social backgrounds, as indicated by parental occupational status (64.7%, social class III and IV), and most of them attended 1st and 2nd primary school grades (52.3%) (see Table 2).

Table 2. Sample characteristics (age, social class and school grade; N=779)

		Total	Girls	Boys
		N=779	N=403	N=376
		n (%)	n (%)	n (%)
	Age (years)
	6–7	328 (42.1)	167 (41.4)	161 (42.8)
	8–9	365 (46.9)	189 (46.9)	176 (46.8)
	10–11	86 (11.0)	47 (11.7)	39 (10.4)
	Social class
	I (high)	190 (24.1)	98 (24.3)	92 (24.5)
	II	85 (10.9)	39 (9.7)	46 (12.2)
	III	338 (43.4)	172 (42.7)	166 (44.1)
	IV (low)	166 (21.3)	94 (23.3)	72 (19.1)
	School grade
	1st	210 (27.0)	114 (28.3)	96 (25.5)
	2nd	197 (25.3)	95 (23.6)	102 (27.1)
	3rd	201 (25.8)	98 (24.3)	103 (27.4)
	4th	171 (22.0)	96 (23.8)	75 (19.9)

Social class: I high, IV low.

3.2. Sleep variables 

Sleep variable frequencies are described in Table 3. Response frequencies to each sleep questionnaire item are also shown.

Table 3. Frequencies of sleep related behaviours/problems, sleep durations and bedtimes/wake-up times in school days and weekends (N=779)

	Items	Response options
		Never n (%)	Occasionally n (%)	Frequently n (%)	Always
	Bedtime behaviors
	Settle to sleep alone	27 (3.5)	78 (10.0)	97 (12.5)	575 (73.8)
	Fall asleep in parents bed	518 (66.5)	198 (25.4)	46(5.9)	14 (1.8)
	Comforting habits to sleep	584 (75.0)	62 (8.0)	47 (6.0)	81 (10.4)
	Light on to fall asleep	477 (61.2)	123 (15.8)	46 (5.9)	128 (16.4)
	Need parents to fall asleep	588 (75.5)	119 (15.3)	42 (5.4)	26 (3.3)
	Bedtime resistance	344 (44.2)	311 (39.9)	101 (13.0)	19 (2.4)
	Willingness to go to bed at bedtime	51 (6.5)	215 (27.6)	255 (32.7)	256 (32.9)
	Nightime behaviors
	Back to sleep alone after night awakenings	669 (85.9)	71 (9.1)	28 (3.6)	6 (.8)
	Loud snoring	480 (61.6)	227 (29.1)	52 (6.7)	11 (1.4)
	Bed wetting	663 (85.1)	91 (11.7)	15 (1.9)	3 (.4)
	Nightmares	284 (36.5)	444 (57.0)	36 (4.6)	4 (.5)
	Sleep walking	705 (90.5)	53 (6.8)	10 (1.3)	2 (.3)
	Sleep talking	300 (38.5)	379 (48.7)	85 (10.9)	8 (1.0)
	Night terrors	611 (78.4)	135 (17.3)	19 (2.4)	2 (.3)
	Teeth grinding in sleep	552 (70.9)	149 (19.1)	57 (7.3)	11 (1.4)
	Fear of sleeping in the dark	513 (65.9)	135 (17.3)	48 (6.2)	74 (9.5)
	Daytime behaviors
	Sleepiness	439 (56.4)	316 (40.6)	20 (2.6)	4 (.5)
	Tiredness	339 (43.5)	382 (49.0)	49 (6.3)	6 (.8)
	Irritability	333 (42.7)	388 (49.8)	50 (6.4)	7 (.9)
	Child Health		Yes	No
	Sleeping problem		35 (4.5)	744 (95.5)
	Sought professional help to sleep		30 (3.9)	749 (96.1)
	Sleep medication		5 (6.0)	774 (99.4)
	Sleep duration/bedtimes and wake up times (M, SD)
	Bedtime school days=21:54 (0:34)		Bedtime weekends=22:54 (0:40)
	Sleep duration school days=9:42 (0:50)		Sleep duration weekends=10:21 (1:02)
	Wake up time school days=7:57 (0:54)		Wake up time weekends=9:31 (1:01)
	Time to fall asleep (n, %)
	<10min=423 (54.3); 10–30min=321 (41.2); >30min=33 (4.2)
	Awakenings during the night (n, %)
	0=530 (68.0); 1 time=213 (27.3); 2 times=26 (3.3); 3 times=3 (.4); >3 times=2 (.3)
	Wake up practices (n, %)
	Alarm clock=26 (3.3); Parents/family members=560 (71.9); Noise=2 (.3); Need to use the toilet=9 (1.2); By himself/herself=176 (22.6); Other=2 (.3).
	Factors influencing bedtime (n, %)
	Family routine=188 (24.1); Sleepiness=126 (16.2); Ends favorite TV program=13 (1.7); Sisters/brothers bedtime=28 (3.6); Has to sleep enough for next day activities=396 (50.8); Other=14 (1.8).

M, Mean; SD, standard deviation; %, percentage; n, number. The total may vary from 779 or 100% within each item due to missing values.

The next step was to compare children with emotional/hyperactive and conduct problems with children with less emotional/behavioral problems with respect to the presence of sleep problems or specific sleep features (see Table 4).

Table 4. Means and Standard Deviations of sleep variables by emotional, hyperactivity and conduct groups within the total sample and gender sub-samples

		Emotional groups						Hyperactivity groups						Conduct groups
		Total		Girls		Boys		Total		Girls		Boys		Total		Girls		Boys
		CG	EG	CG	EG	CG	EG	CG	HG	CG	HG	CG	HG	CG	CPG	CG	CPG	CG	CPG
		n=740	n=39	n=383	n=20	n=357	n=19	n=665	n=114	n=364	n=39	n=301	n=75	n=683	n=96	n=371	n=32	n=312	n=64
	Sleep factors/items	M (sd)	M (sd)	M (sd)	M (sd)	M (sd)	M (sd)	M (sd)	M (sd)	M (sd)	M (sd)	M (sd)	M (sd)	M (sd)	M (sd)	M (sd)	M (sd)	M (sd)	M (sd)
	Parents help to sleep	4.2 (1.87)	4.0 (1.56)	5.9 (1.93)	5.9 (1.55)	4.2 (1.82)	3.8 (1.61)	4.3 (1.88)	4.1 (1.74)	4.3 (1.92)	4.1 (1.83)	4.2 (1.83)	4.1 (1.70)	4.3 (1.88)	4.0 (1.67)	4.3 (1.90)	4.1 (1.98)	4.2 (1.86)	4.0 (1.50)
	Parasomnias	5.8 (1.44)	5.9 (1.83)	5.7 (1.38)	6.3 (2.26)	5.8 (1.51)	5.5 (1.31)	5.7 (1.44)	5.8 (1.59)	5.6 (1.39)	6.0 (1.76)	5.8 (1.50)	5.7 (1.49)	5.7 (1.44)	5.9 (1.60)	5.7 (1.40)	6.0 (1.76)	5.8 (1.49)	5.8 (1.52)
	Sleeping difficulties	5.9 (.37)	5.9 (.52)	5.9 (.37)	5.9 (.31)	5.9 (.38)	5.8 (.69)	5.9 (.39)	5.9 (.36)	5.9 (.38)	6.0 (.16)	5.9 (.40)	5.9 (.42)	5.9 (.40)	5.9 (.24)	5.9 (.38)	6.0 (.18)	5.9 (.42)	5.9 (.27)
	Afraid of dark	3.4 (1.97)	3.2 (2.12)	3.4 (1.90)	3.3 (2.20)	3.4 (2.04)	3.1 (2.07)	3.4 (1.96)	3.3 (2.07)	3.3 (1.88)	3.5 (2.15)	3.4 (2.04)	3.2 (2.03)	3.4 (1.97)	3.3 (1.99)	3.4 (1.97)	3.2 (1.84)	3.4 (2.04)	3.3 (2.07)
	Sleep limit setting	3.8 (1.50)	3.6 (1.65)	3.8 (1.53)	3.7 (1.72)	3.9 (1.46)	3.6 (1.61)	3.8⁎(1.49)	4.1⁎(1.58)	3.8 (1.52)	4.0 (1.70)	3.8⁎(1.44)	4.2⁎(1.51)	3.8⁎(1.50)	4.1⁎(1.51)	3.8 (1.55)	3.9 (1.52)	3.8⁎(1.44)	4.3⁎(1.50)
	Daytime consequences	4.8 (1.31)	5.2 (1.74)	4.8 (1.32)	5.1 (2.00)	4.7 (1.31)	5.3 (1.48)	4.7 (1.34)	4.9 (1.35)	4.8 (.71)	4.9 (0.23)	4.7 (1.32)	4.9 (1.32)	4.8 (1.35)	4.9 (1.29)	4.8 (1.37)	5.0 (1.23)	4.7 (1.32)	4.9 (1.32)
	Time to fall asleep	1.5 (.58)	1.4 (.60)	1.5 (.60)	1.4 (.50)	1.5 (.56)	1.5 (.61)	1.5 (.58)	1.5 (.60)	1.5 (.59)	1.6 (.64)	1.5 (.56)	1.5 (.58)	1.5 (.57)	1.5 (.62)	1.5 (.59)	1.7 (.65)	1.5 (.56)	1.5 (.59)
	Night awakennings	1.4 (.60)	1.3 (.52)	1.4 (.66)	1.3 (.57)	1.3 (.47)	1.3 (.50)	1.4 (.60)	1.4 (.60)	1.4 (.66)	1.4 (.64)	1.3 (.53)	1.4 (.52)	1.4 (.60)	1.4 (.55)	1.4 (.66)	1.4 (.62)	1.3 (.53)	1.4 (.52)
	Snoring	1.5 (.68)	1.5 (.76)	1.5 (.70)	1.5 (.91)	1.5 (.67)	1.5 (.61)	1.5 (.68)	1.5 (.75)	1.5 (.70)	1.5 (.80)	1.5 (.65)	1.5 (.73)	1.5 (.68)	1.6 (.75)	1.5 (.71)	1.5 (.72)	1.5 (.64)	1.6 (.76)
	Bed wetting	1.2 (.44)	1.2 (.58)	1.1 (.35)	1.3 (.75)	1.2 (.52)	1.1 (.32)	1.2 (.44)	1.2 (.52)	1.1 (.37)	1.2 (.47)	1.2 (.50)	1.2 (.55)	1.2 (.43)	1.2 (.56)	1.1 (.37)	1.3 (.51)	1.2 (.49)	1.2 (.59)
	Back to sleep alone	3.8 (.51)	3.8 (.67)	3.8 (.55)	3.7 (.80)	3.8 (.47)	3.8 (.50)	3.8 (.53)	3.9 (.49)	3.8 (.56)	3.8 (.63)	3.8 (.28)	3.9 (.45)	3.8 (.53)	3.9 (.47)	3.8 (.56)	3.8 (.60)	3.8 (.49)	3.9 (.35)
	Sleep duration-SD	9:42 (0:49)	9:56 (1:02)	9:40⁎(0:52)	10:08⁎(1:11)	9:43 (0:45)	9:45 (0:50)	9:40⁎(0:47)	9:54⁎(1:03)	9:40 (0:53)	9:52 (1:01)	9:40 (0:39)	9:55 (1:04)	9:41⁎(0:48)	9:54⁎(1:02)	9:41 (0:53)	9:47 (1:04)	9:41 (0:41)	9:57 (1:02)
	Sleep duration-WE	10:20 (1:01)	10:37 (1:07)	10:23⁎(1:02)	10:55⁎(1:09)	10:17 (1:00)	10:18 (1:01)	10:19⁎(1:00)	10:34⁎(1:10)	10:23 (1:03)	10:35 (1:01)	10:1⁎(0:56)	10:33⁎(1:14)	10:19 (1:00)	10:31 (1:10)	10:24 (1:03)	10:29 (1:05)	10:14 (0:57)	10:33 (1:12)
	Wake up time-SD	7:58 (0:54)	7:55 (0:52)	7:58 (0:58)	7:59 (1:02)	7:57 (0:49)	7:48 (0:40)	7:57 (0:53)	8:02 (1:01)	7:59 (0:57)	7:53 (1:10)	7:54 (0:47)	8:05 (0:56)	7:57 (0:53)	7:59 (1:02)	8:00 (0:58)	7:42 (1:03)	7:54 (0:46)	8:07 (1:00)
	Wake up time-WE	9:31 (1:00)	9:17 (1:08)	9:35 (1:02)	9:24 (1:07)	9:27 (0:58)	9:11 (1:11)	9:31 (1:00)	9:27 (1:06)	9:34 (1:01)	9:37 (1:11)	9:28 (0:58)	9:22 (1:04)	9:31 (1:00)	9:30 (1:04)	9:34 (1:01)	9:41 (1:07)	9:27 (0:59)	9:24 (1:02)
	Bedtime-SD	21:55⁎(0:34)	21:43⁎(0:34)	21:55⁎(0:36)	21:36⁎(0:35)	21:54 (0:32)	21:51 (0:31)	21:55 (0:34)	21:51 (0:35)	21:55 (0:36)	21:45 (0:38)	21:54 (0:32)	21:55 (0:32)	21:55 (0:34)	21:48 (0:34)	21:56⁎(0:36)	21:38⁎(0:36)	21:54 (0:32)	21:53 (0:32)
	Bedtime-WE	22:55 (0:40)	22:52 (0:43)	22:55 (0:41)	22:39 (0:43)	22:54 (0:39)	23:06 (0:40)	22:55 (0:40)	22:53 (0:42)	22:55 (0:41)	22:53 (0:46)	22:55 (0:39)	22:53 (0:40)	22:55 (0:41)	22:48 (0:37)	22:55 (0:42)	22:52 (0:37)	22:56 (0:39)	22:47 (0:38)

SD, school days; WE, weekends; M, mean; sd, standard deviation; CG, control group: children scoring less than 9 points on the Rutter Scale B2; EG, emotional group: children scoring 9 or more points on the Rutter Scale B2 and higher points on emotional items than on conduct items; HG, hyperactivity group: children scoring 9 or more points on the RB2 scale and 3 or more on hyperactivity items; CG, conduct group: children scoring 9 or more points on the scale and higher points on conduct items than on emotional items.

⁎ Significant statistical comparisons between groups are highlighted in bold, t student tests (p<.05).

Children with more emotional sub-type behaviors went to bed earlier in school days (M=21:43, SD=0:34) compared to the group with less emotional problems (M=M=21:55, SD=0:34, t=2.054, df=774, p=.04).

Girls whose teachers reported having more emotional problems in school were described by their parents as sleeping longer during both school days and weekends and going to bed earlier during school days than girls with less emotional problems (Sleep duration school days: M=10:08h, SD=1:11min vs M=9:40h, SD=0:52min, t=−2.219, df=395, p=.027; Sleep duration weekends: M=10:55h; SD=1:09min vs M=10:23h, SD=1:02min, t=−2.174, df=394, p=.030; Bedtime school days: M=21:36h, SD=0:35min vs M=21:55h, SD=0:36min; t=2.238 df=399, p=.020).

Parents of children with the Hyperactivity sub-type behaviors reported having more difficulties in setting sleep limits at bedtime (M=4.11h; SD=1.58min) and that their children slept longer during school days (M=9:54h; SD=1:03min) and weekends (M=10:34h; SD=1:10min) than parents of children with less hyperactivity behaviors (Limit setting factor: M=3.76, SD=1.49; Sleep duration school days: M=09:40h, SD=0:47min, Sleep duration weekends: M=10:19h, SD=1:00min; p values <.05).

Analysis by gender revealed that parents of boys with more hyperactivity problems reported having more sleep limit difficulties at bedtime (M=4.19; SD=1.51) and that their sons slept longer during weekends (M=10:33h, SD=1:14min) than parents of boys with less hyperactive behaviors (Sleep Limit setting factor: M=3.79, SD=1.44; Sleep duration weekends: M=10:13h, SD=0:56min; p values <.05).

Children with more conduct problems in school were more likely to show opposite/defiant behaviors at bedtime (Sleep limit setting factor: M=4.14, SD=1.51) and to sleep longer during school days (M=9:54h, SD=1:02min) than the group of children with less conduct problems (Sleep limit setting factor: M=3.77, SD=1.50; Sleep duration school days: M=9:41h, SD=0:48min, p values <.05). In particular, boys with more conduct problems showed more opposite/defiant behaviors at bedtime than boys with less conduct problem behaviours (M=4.27, SD=1.50 vs M=3.78, SD=1.44, t=−2.415, df=372, p=.016), while girls with more conduct problems went to bed earlier during school days than girls with less conduct problems (M=21:38, SD=0:36 vs M=21:56, SD=0:36, t=2.518, df=399, p=.012).

Boys who either always or frequently displayed bedtime resistance were significantly more likely within the group with hyperactive problems (39%, n=16) than in the group with less hyperactive problems (24.3%, n=42) (χ2=3.648; df=1; p=.046). Similar results were found between the group with conduct problems and the group with less conduct problems with respect to bedtime resistance (41.2%, n=14 vs. 24.4%, n=44, χ2=4.052; df=1; p=.039). The frequency of boys who were never or rarely willing to go to bed (66.0%, n=31) was also significantly higher in the group of boys with conduct problems than in the group of boys with less conduct problems (51.0%, n=104) (χ2=3.477; df=1; p=.044).

In order to investigate if in boys the sleep limit setting factor was still associated with hyperactive/conduct problems when demographic and other sleep variables were controlled, multivariate regressions were then performed. These analyses were carried out separately for boys to control for gender effects and 5 children who were taking sleeping pills were excluded from the analyses. As mentioned in the methods section, children with known diseases were excluded a priori from the data sample. The dependent variables were the Rutter hyperactive sub-score (sum of scores in hyperactive items, quantitative variable) and the Rutter conduct sub-score (sum of scores in conduct items, quantitative variable).The independent variables were the sleep variables that in the preceding analyses were associated with behavioral problems. With respect to demographic variables, social class was included in the analyses since in a previous study, this variable was associated with behavioral problems in both genders [11]. However, age was not included for the present analyses because in our previous study [11] age was not associated with behavioral/emotional problems in boys.

First, a hierarchical regression was conducted using the hyperactive sub-score as a dependent variable and social class and sleep duration during weekends as predictors, then the sleep limit setting factor was introduced in the model after controlling the effects of social class and sleep duration on weekends. Subsequently, another hierarchical regression was conducted using as dependent variable the conduct sub-score and as predictors first social class and then the seep limit setting factor. We found that parents’ sleep limit difficulties predicted hyperactive behaviours in boys after controlling for sleep duration during weekends and social class (R2 change=.02, F(9,42)=7.95, p=.005). With respect to conduct behaviours, parents’ sleep limit difficulties predicted conduct problems in school after controlling for social class (R2 change=.01, F(10,09)=5.31, p=.022).

Thus, boys who have similar social class backgrounds and have similar sleep durations at weekends appear to be more likely to display hyperactive behaviours in school if their parents report sleep limit setting difficulties. Boys with similar social class are also more likely to show conduct problems in school if their parents report sleep limit setting difficulties at home.

4. Discussion 

As a whole, the present study shows that emotional/disruptive behaviors in school (as reported by teachers) were not associated with parental reports of sleep problems such as night terrors, sleep talking, nightmares, sleepwalking, night awakenings, snoring or bed wetting. These findings diverge with those of two previously published surveys in Northern European countries [3], [4], which suggested that children with behavioral/emotional problems in school are reported by their parents as more likely having sleep problems at home, as well as from a Chinese study [12] which shows, in a school-based sample of 2463 children, that daytime inadvertent napping, sleep-disordered breathing and sleep schedules were associated with Attention-Deficit Hyperactivity Disorder (ADHD) related symptoms as assessed by mothers and teachers. Notwithstanding these discrepancies, we also found that parents of children (particularly of boys) with hyperactive and conduct problems in school were more likely to report increased frequency of sleep limit setting difficulties at bedtime than parents of children with less hyperactive or conduct problems, and this is in close agreement with Smedje et al. [3]. However, in the Smedje et al. study [3], hyperactivity was better predicted from the reports on tossing and turning during sleep and sleep walking rather than by the presence of bedtime resistance [3].

In our study, separate exploration of the hyperactive group and the group with conduct problems yielded similar, albeit slightly different findings, whereby bedtime resistance was present in both groups of boys. However, boys with hyperactive behaviors were reported by their parents as showing bedtime resistance but not unwillingness to go to bed, whereas parents of boys with more conduct problems in school reported resistance at bedtime as well as unwillingness of going to bed.

We should emphasize that it was not possible to disentangle conduct problems from hyperactive behaviors in our cohort, since nearly all children with hyperactivity also had, to a varying degree, conduct problems, such that the number of children with pure hyperactive behaviors was too small to enable further statistical analysis.

Therefore, our results suggest that parental reports of sleep limit difficulties at bedtime appear to emerge as an extension of their children’s problems in school, whether such problems are hyperactivity or school conduct. Another explanation for these findings could be that children with hyperactive/conduct problems may have a real problem falling asleep. However, if this was the case it would be expected that they would take longer to fall asleep than other children, and this was not observed in our sample.

In fact, parental difficulty in setting sleep limits at bedtime predicted boys hyperactivity/conduct problems in school, and this association could not be attributed to age, social class, other sleep variables, namely sleep duration, snoring, time to fall asleep and sleep medication consumption, or health status (asthma, bronchitis or mental disease) as these variables were either statistically controlled or excluded a priori from the data sample, or alternatively not associated with behavioral/emotional problems. Moreover, this association was not influenced by psycho-stimulants effects as none of these children were prescribed with these medications.

Another noteworthy finding of our survey pertains to sleep duration. Indeed, during school days, children with more hyperactive or conduct problems slept longer than children with less hyperactive or conduct problems. At weekends, similar results were observed only for children with hyperactive behaviors. Based on the published literature, we would have expected that children with more hyperactive/conduct problems would not be reported by their parents to sleep as much as children with less hyperactive/conduct problems. Indeed, subjective reports of sleep duration in children with ADHD would support such contention [13], [14], [15], even if objective sleep assessments are somewhat contradictory in this regard [16], [17], [18]. One explanation for these a priori conflictive reports could reside in the putative hypoarousal state as initially suggested by Satterfield et al. [19] using electroencephalography (EEG) in children with ADHD. This assumption on the presence of intrinsic sleepiness in ADHD children has received some support as evidenced by the reduced mean daytime sleep latencies in children with ADHD [20], [21].

Another potential explanation for the present findings on parentally-reported increases in sleep duration is that children with either hyperactive or conduct problems may need to sleep longer because they have less well consolidated sleep, such that the proposed metabolic recovery afforded by sleep will take longer to complete [18].

We also found that girls with either emotional or conduct problems were particularly more likely to go to bed earlier during school. It is possible that such a pattern may simply reflect earlier onset of fatigue or sleepiness in the evening due to the daytime behaviors, or alternatively, the parents of these girls might become more tired in the evening because of their child’s behaviors, and consequently anticipate and promote earlier bedtimes.

The majority of studies investigating the relationships between sleep and ADHD indicate a higher prevalence of sleep complaints among children with ADHD compared to control children [13], [22], [23], [24], [25], [26]. Such sleep complaints are manifested as increased bedtime resistance, difficulty in falling asleep, more frequent nighttime awakenings, the presence of sleep-related anxiety, and are commonly reported not only by the parents, but also by the ADHD patients themselves [27].

Results from the present study can not be compared with the above studies as children’s emotional or disruptive behaviors were assessed using teacher reports, rather than through implementation of the more extensive diagnostic procedures as recommended by several sources [28], [29], [30].

In addition, our findings rely on parental reports of sleep problems, and as such are not only markedly subjective, but also convey the inherent bias derived from the interpretation by the parent of their child’s manifestations and do not incorporate more objective tools such as actigraphy or polysomnography .

Thus, the negative findings found in this study may simply represent the fact that parents are unaware of their child’s sleeping difficulties, sleep duration, or actual bedtimes. The older the child is, the poorer the agreement between the child’s self report and the parental report when assessing sleep quality [31]. In fact, children reports tend to indicate more sleep problems than parental reports [32]. Furthermore, the questions about sleep duration, bedtimes and wake up times used in this study were in a general format, and allowed estimates to be construed rather than providing precise assessments of these sleep variables.

Notwithstanding such limitations, this study revealed that in a large community sample of Portuguese school aged children parental difficulties in setting bedtime limits were associated with children with either hyperactivity or conduct problems and this factor might be worthy of further exploration in future studies, particularly considering that bedtime struggles has previously emerged as one of the major factors (along with psychosocial problems in the family and long sleep latency in bed) associated with subsequent development of attention-deficit hyperactivity disorder [33]. Health care professional information to parents about a good sleep hygiene at bedtime and its value for the child’s well-being may possibly benefit children with bedtime resistance and behavioural problems.