Sleep disturbances and correlates among a sample of preschool children in rural China

Abstract

This study aimed to explore the prevalence of sleep disturbances among preschool children in rural areas of China and identify the underlying associated factors. A cross-sectional study was conducted in rural areas of Anhui Province, China, from September 2019 to January 2020. The caregivers of children from 26 kindergartens completed the demographic questionnaire, the Children’s Sleep Habits Questionnaire (CSHQ), the Strengths and Difficulties Questionnaire (SDQ), the Self-rating Anxiety Scale (SAS), and the adapted Identification and Management of Feeding Difficulties (IMFeD) tool. One-way analysis of variance, independent-samples t-test, and hierarchical multiple linear regression were conducted to explore the potential influencing factors of sleep disturbances in preschool children. For the specific sleep disturbances based on each subscale of the CSHQ, bedtime resistance (54.9%) and sleep anxiety (47.9%) were the most common types of sleep disturbances, followed by daytime sleepiness (21.3%), sleep duration (20.8%), parasomnias (16.5%), sleep-onset delay (12.6%), night wakings (12.2%) and sleep-disordered breathing (10.3%). The ages of children, children’s emotional/behavioral problems, children’s feeding difficulties, inconsistent parenting attitudes of parents, poor parenting styles, non-parent caregivers, and caregiver anxiety contributed significantly to the CSHQ total score, accounting for approximately 27.1% (R² = 0.271) of the variance. Our study indicates that the prevalence of sleep disturbances among preschool children in rural areas of China is quite high. Furthermore, the potential risk factors are complicated, including factors related to both children and their caregivers.

Introduction

Sleep disturbances are associated with various cognitive and emotional/behavioral problems in children, and the link between them is complex [1–3]. Poor sleep quality could cause a negative impact on the development of children’s physical and mental health. The prevalence of sleep disturbances in different countries varies greatly, especially between the West and East [4]. Studies have indicated that parents were more likely to report sleep problems in their children in Asian countries (24.2%) compared to their counterparts in Western countries (18.4%); in Asian countries, the prevalence of parents reporting sleep problems among their children ranged from 15.1% in Korea to 43.7% in China [5]. Since sleep disturbances could have a detrimental impact on child development, everyday functioning, and quality of life [6], it is important to explore the related factors of sleep disturbances among children.

Furthermore, the sleep patterns among different countries and regions also show different characteristics. In China, children tend to sleep shorter and wake up earlier than their counterparts in the USA; the potential reason may be the differences in school schedules, homework load, and sleep habits between the two countries [7]. A cross-cultural study between Chinese and Japanese preschool children also found differences in sleep patterns between the two countries; although there was no difference in the total severity of sleep disturbance, Chinese preschoolers reported more severe levels of nighttime awakenings and sleep-disordered breathing (SDB), whereas Japanese preschoolers were more susceptible to bedtime resistance [8]. Given that the reason for prevalent discrepancies in different countries and regions is still not clear, more detailed and culture-tailored research is needed to implement interventions in different countries and regions with different sociocultural backgrounds [9]. In China, because of the differences in the socioeconomic status, growth environments, and educational achievements of caregivers between urban and rural areas [10], underlying rural–urban differences in sleep disturbances among children may exist. A previous study explored potential sleep-related factors among urban children in China [9]. Therefore, the characteristics of sleep disturbances among children in rural areas of China need to be investigated to supplement existing research results.

Improper parental supervision, such as inadequate care, parental absence, and overprotection, may have long-term effects on children’s development and health outcomes, such as sleep disturbance, even in adulthood [11, 12]. It is commonly believed that sociocultural factors, including economic considerations and ideological beliefs regarding societal and parental awareness, may play important roles in sleep disturbances among preschool children [4, 9, 13], indicating that parents play an important role in ensuring the quality of their children’s sleep. In addition, children’s lifestyles (including screen time and dietary habits) and body mass index (BMI) were also associated with insufficient sleep duration [14, 15]. Left-behind children may have a higher risk of unhealthy lifestyles and poorer health status [16, 17], which may be closely related to their sleep quality. Therefore, the factors associated with sleep problems among children are complicated and need to be explored comprehensively.

Preschool age is a sensitive developmental stage, and obtaining adequate amounts of sleep is vital to preschool children’s social, emotional, behavioral, and academic development [12, 18]. To date, population-based studies of sleep disturbances among preschool children in rural areas of China are limited, and the influencing factors are still unclear [9, 19, 20]. It is necessary to conduct studies to broadly explore the underlying related factors of sleep disturbances in such a vulnerable group to provide clues for further studies and interventions. Therefore, we conducted a cross-sectional study aimed at understanding the prevalence of sleep disturbances and their associated factors among preschool children in rural areas of China based on a large sample.

Methods

Study design and subjects

We conducted this cross-sectional study from September 2019 to January 2020, and four counties (Funan County in Fuyang, Changfeng County and Feixi County in Hefei, Bowang District in Ma’anshan) were selected as our research sites based on the geographical locations of northern, central, and southern Anhui Province. Kindergartens were enrolled using convenience sampling methods, preschool children from 26 selected kindergartens were enrolled, and their parents or primary caregivers were asked to complete the demographic questionnaire through face-to-face interviews or self-reporting. For caregivers with difficulties completing the questionnaire because of their education levels, face-to-face interviews were conducted. In our study, preschool children were defined as children aged 3–6 years who had not yet attended a primary school. In addition, we excluded participants whose caregivers could not communicate, such as those with cognitive impairment and deafness, and children with severe physical and mental illness were also excluded. After completing the questionnaires, all the items were checked to ensure their integrity. A total of 3,802 questionnaires were distributed. After checking for completeness, 166 invalid questionnaires were excluded, and 3,636 questionnaires were eventually included, with a response rate of 95.6%. This study was approved by the ethics committee of Anhui Medical University (20180402) and followed the principles of informed consent and strict confidentiality.

Assessment instruments

In this study, our questionnaires are well established and consist of several scales to comprehensively evaluate sleep problems among preschool children. The details of the scales included in our questionnaire are as follows.

Demographic questionnaire

We evaluated the sociodemographic information and related variables of children and their families using a questionnaire that assessed the following data: gender, left-behind status, age, birth weight, only child status, average monthly household income, adiposity indicator, screen time level, education level of father, education level of mother, upbringing attitudes of father and mother, main educational methods, and caregivers of children. Left-behind children were defined as children who had been left behind by at least one parent for more than 6 months. The main educational methods were defined as the main way that parents or other caregivers educate their children after they do something wrong. The caregivers of children category indicates whether children’s caregivers are their parents (parents or non-parent caregivers). The adiposity indicator was evaluated by BMI. According to the WHO standard of BMI for preschoolers [21, 22], for the same gender and age, BMI between the 85th and 95th percentiles is overweight, and greater than the 95th percentile indicates obesity. We acquired the height and weight of the children through measurement, and BMI was further generated (BMI = weight/height², kg/m²). Other information was self-reported by the respondent.

The definition of screen time level was set according to the guidelines of the American Academy of Pediatrics (AAP) [23, 24]: normal screen time (≤ 60 min/day), excessive screen time (60 min/day to ≤ 120 min/day), and severely excessive screen time (> 120 min/day).

Children’s sleep habits questionnaire (CSHQ)

The Children’s Sleep Habits Questionnaire (CSHQ) was used to assess the sleep problems of preschool children. The CSHQ was developed in the USA by Owens et al. and is widely utilized to assess and detect sleep problems among children aged 2–10 years [25, 26]. We utilized the Chinese version of the CSHQ [27], which included 33 items describing sleep disturbances. Each item is rated on a three-point scale: “usually” if sleep behavior occurred five to seven times per week, “sometimes” if two to four times per week, and “rarely” if zero to one time per week. Among 33 items, sleep disturbances are evaluated on eight subscales, including bedtime resistance, sleep-onset delay, sleep duration, sleep anxiety, night wakings, parasomnias, sleep-disordered breathing, and daytime sleepiness, with higher total and subscale scores representing more severe disturbances [4]. Parents or other caregivers were asked to recall the sleep behaviors of included children occurring over a “typical” recent week [28]. Given that the CSHQ is not regarded as a diagnostic questionnaire but as a screening tool for existing sleep disturbances, we utilized the definition of previous studies, which defined a total score > 41 as potential clinical sleep disturbances [4, 28]. The cutoff value for subscales was defined as scores > 2 standard deviations above the published community control reference mean values [25] (bedtime resistance > 10.84, sleep-onset delay > 2.31, sleep duration > 5.27, sleep anxiety > 7.79, night wakings > 5.29, parasomnias > 10.61, sleep-disordered breathing > 4.50, and daytime sleepiness > 15.24) according to previous studies [9, 29, 30]. In our study, the Cronbach’s α coefficient was 0.668.

Strength and difficulty questionnaire (SDQ)

The Strength and Difficulty Questionnaire (SDQ) was designed by Goodman et al. to evaluate the behaviors, emotions, and hyperactivity situations of children aged 3–16 years [31, 32]. The Chinese version of the parental SDQ [33] was adopted to assess children’s emotional/behavioral problems. There are 25 items in this scale, which consists of four difficulty subscales (emotional symptoms, conduct problems, hyperactivity/inattention, and peer problems) as well as a strength subscale (prosocial behavior). Each item has a score of 0–2, with a rating of three levels: 0 for not fit, 1 for a little fit, and 2 for a complete fit. The total difficulties score is the sum of the scores of the four difficulty subscales, and higher scores imply more difficulties; conversely, higher scores in prosocial behavior indicate more strength [9]. The Cronbach’s α coefficient was 0.689 in this study.

Self-rating anxiety scale (SAS)

We utilized the Self-rating Anxiety Scale to evaluate the anxious status of caregivers. The Self-rating Anxiety Scale is a self-report scale that includes 20 items covering a variety of anxiety symptoms [34]. Respondents are asked to score each of the items regarding how it applied to them during the past week ((1) none or a little of the time; (2) some of the time; (3) good part of the time; (4) most or all of the time); a higher score indicates a more severe anxiety status [35]. The scores of 20 items are calculated to obtain a raw score ranging from 20 to 80, and the standard score is generated through the raw score multiplied by 1.25; the standard score ≥ 50 indicates the presence of anxiety status, while the standard scores of 50–59, 60–69, and ≥ 70 are considered mild, moderate, and severe anxiety, respectively [36, 37]. The Cronbach’s α coefficient was 0.558 in this study.

Identification and management of feeding difficulties (IMFeD)

The Chinese version of the adapted IMFeD tool was used to explore feeding difficulties among preschool children [38, 39]. This questionnaire consists of 17 items that describe the feeding difficulty status of children through six dimensions: poor appetite (four items, e.g., “Are your children not interested in food and rarely show hunger?’’), food preference (three items, e.g., “Is your child reluctant to try new foods?’’), poor eating habits (four items, e.g., “Does your child always keep food in his mouth and not swallow it?”), fear of feeding (two items, e.g., “Are your children scared when they are ready to eat or when food and tableware appear?”), parental misperception (two items, e.g., “Do you often think your children are not eating enough?”), and organic disease (two items, e.g., “Does your child have a bad appetite and frequent vomiting or diarrhoea?”). Each item is rated through a five-point scale (behaviors happened every day; behaviors happened 5 or 6 days per week; behaviors happened 3 or 4 days per week; behaviors happened 1 or 2 days per week; behaviors never happened). Each item represents a problem in eating behaviors, and caregivers are asked to report how often their children displayed eating disorder behaviors in each week according to each item; if any item is reported to occur always or often (behaviors exceed three days a week), we considered the child to have a problem in that dimension [39–41]. A higher score indicates better eating habits, while a lower score indicates more eating problems [38]. The Cronbach’s α coefficient was 0.887 in this study.

Statistical analysis

The data were entered into EpiData 3.2. SPSS 23.0 software was utilized to perform statistical analysis. We used the mean, standard deviation, frequencies, and percentages to illustrate the distribution of participants in different sociodemographic characteristics. Table 1 shows the assignments of variables in detail. We sorted the included variables into three categories for further analysis: (1) sociodemographic factors (gender, left-behind status, age, birth weight, only child status, and average monthly household income); (2) children’s associated factors (adiposity indicator, screen time level, SDQ scales, and IMFeD scales); and (3) caregivers’ associated factors (education level of father, education level of mother, upbringing attitudes of father and mother, main educational methods, caregivers of children, and anxious status of caregivers). One-way analysis of variance and independent samples t-test was used to compare the total score and the score of each subscale of the CSHQ among different levels of categorical sociodemographic factors. Furthermore, using the total CSHQ score as the dependent variable, hierarchical multiple linear regression was performed to explore the link between various independent variables and the dependent variable by entering independent variables successively; step 1 included the sociodemographic factors, step 2 included the children’s associated factors, and step 3 included the caregivers’ associated factors. Collinearity diagnosis was also performed. P < 0.05 was considered statistically significant.

Table 1.

The assignment of values among variables

Variables	Value Assignment
Gender	1 = Male; 2 = Female
Left-behind status Birth weight	0 = non-LBC, 1 = LBC 1 = Underweight, 2 = Overweight, 3 = Normal
Age(years)	3 = 3 years old; 4 = 4 years old; 5 = 5 years old; 6 = 6 years old
Adiposity indicator (BMI)	0 = Normal, 1 = Overweight, 2 = Obesity;
Education level of father	1 = Junior high school or below; 2 = High school; 3 = College or above
Education level of mother	1 = Junior high school or below; 2 = High school; 3 = College or above
Upbringing attitudes of father and mother	1 = Unanimous; 2 = Occasionally inconsistent; 3 = Often inconsistent
Main educational methods	1 = Natural education, 2 = Persuade education, 3 = Scolding education;
Anxious status of caregivers	0 = Without anxiety, 1 = Mild anxiety, 2 = Moderate anxiety, 3 = Severe anxiety
Caregivers of children	1 = Parents; 2 = Non-parent caregivers
Only child status	1 = Yes; 2 = No
Average monthly household income (CNY)	1 =  < 3,000; 2 = 3,000–4,999; 3 = 5,000–10,000; 4 =  > 10,000
Screen time level	1 = normal screen time (NST, ≤ 60 min/d), 2 = excessive screen time (EST, 60 min/d < screen time ≤ 120 min/d); 3 = seriously excessive screen time (SEST, > 120 min/d)
Sleep disturbances	0 = No; 1 = Yes
The Children’s Sleep Habits Questionnaire (CSHQ)	Raw score
The Strengths and Difficulties Questionnaire (SDQ)	Raw score
Self-rating Anxiety Scale (SAS)	Standard score (Raw score*1.25)
The adapted Identification and Management of Feeding Difficulties (IMFeD)	Raw score

LBC Left-behind children; CNY Chinese Yuan

Results

General characteristics and sleep disturbances of participants

A total of 3,802 questionnaires were distributed to caregivers in our investigation. After checking for completeness, 166 invalid questionnaires were excluded, and 3,636 questionnaires were eventually included. The mean age of the participants was 4.5 (SD = 0.9) years, with a range of 3–6 years. As shown in Table 2, among the participants, approximately 54.2% were male and 45.8% were female. Total CSHQ scores in birth weight, age, upbringing attitudes of father and mother, main educational methods, anxious status of caregivers, and screen time level showed significant differences according to ANOVA and t-tests. Details of the subscales of the CSHQ are given in Tables 2 and 3.

Table 2.

Comparison of CSHQ scores of preschool children with different sociodemographic characteristics

Characteristics	N(%)	Bedtime resistance (Mean ± SD)	Sleep onset delay (Mean ± SD)	Sleep anxiety (Mean ± SD)	Sleep duration (Mean ± SD)	Night wakings (Mean ± SD)	Parasomnias (Mean ± SD)	Sleep disordered breathing (Mean ± SD)	Daytime sleepiness (Mean ± SD)	Total score (Mean ± SD)
Gender
Male	1,970(54.2)	10.90 ± 2.55	1.57 ± 0.71	7.29 ± 2.02	4.26 ± 1.45	4.06 ± 1.21	8.90 ± 2.00	3.47 ± 0.97*	13.19 ± 2.79*	49.42 ± 6.95
Female	1,666(45.8)	10.84 ± 2.51	1.57 ± 0.70	7.29 ± 2.01	4.26 ± 1.41	4.09 ± 1.27	8.86 ± 1.98	3.41 ± 0.90*	13.39 ± 2.80*	49.45 ± 6.83
Left-behind status a
Non-LBC	2,319(63.8)	10.95 ± 2.51*	1.61 ± 0.70**	7.33 ± 1.99	4.23 ± 1.39	4.00 ± 1.20**	8.80 ± 1.92**	3.43 ± 0.91	13.28 ± 2.79	49.31 ± 6.73
LBC	1,317(36.2)	10.73 ± 2.56*	1.52 ± 0.70**	7.22 ± 2.07	4.31 ± 1.49	4.20 ± 1.28**	9.02 ± 2.10**	3.47 ± 0.98	13.27 ± 2.81	49.66 ± 7.16
Birth weight
Underweight	147(4.0)	10.93 ± 2.45**	1.55 ± 0.74	7.20 ± 1.96*	4.42 ± 1.45	4.18 ± 1.45b	9.17 ± 2.39b	3.56 ± 1.14b	12.61 ± 2.88**	49.39 ± 7.28*
Overweight	273(7.5)	10.42 ± 2.52**	1.54 ± 0.67	6.95 ± 1.92*	4.25 ± 1.41	4.03 ± 1.25b	8.81 ± 1.89b	3.45 ± 0.85b	13.05 ± 2.68**	48.45 ± 6.60*
Normal	3,216(88.4)	10.91 ± 2.53**	1.58 ± 0.70	7.33 ± 2.02*	4.25 ± 1.43	4.07 ± 1.22b	8.88 ± 1.98b	3.44 ± 0.93b	13.33 ± 2.80**	49.52 ± 6.89*
Age (years)
3 ~	1,253(34.5)	11.11 ± 2.49**	1.62 ± 0.71*	7.37 ± 1.97**	4.24 ± 1.39	4.19 ± 1.31**b	9.05 ± 2.10**b	3.46 ± 0.97	13.19 ± 2.77**	49.91 ± 6.96**
4 ~	1,247(34.3)	10.98 ± 2.52**	1.57 ± 0.71*	7.34 ± 2.03**	4.28 ± 1.43	4.05 ± 1.22**b	8.89 ± 2.04**b	3.42 ± 0.96	13.51 ± 2.84**	49.77 ± 6.89**
5 ~	946(26.0)	10.57 ± 2.54**	1.52 ± 0.69*	7.20 ± 2.04**	4.24 ± 1.46	3.95 ± 1.17**b	8.70 ± 1.81**b	3.44 ± 0.86	13.14 ± 2.79**	48.61 ± 6.78**
6	190(5.2)	10.17 ± 2.49**	1.52 ± 0.65*	6.90 ± 2.04**	4.38 ± 1.47	4.04 ± 1.18**b	8.66 ± 1.70**b	3.45 ± 0.96	13.03 ± 2.61**	48.27 ± 6.57**
Adiposity indicator (BMI)
Normal	3,029(83.3)	10.91 ± 2.53	1.58 ± 0.71	7.30 ± 2.01	4.25 ± 1.43	4.06 ± 1.23	8.85 ± 1.96	3.43 ± 0.92b	13.29 ± 2.80	49.42 ± 6.84
Overweight	342(9.4)	10.60 ± 2.50	1.54 ± 0.67	7.21 ± 2.07	4.22 ± 1.35	4.11 ± 1.24	9.06 ± 2.19	3.48 ± 1.01b	13.22 ± 2.78	49.33 ± 7.40
Obesity	265(7.3)	10.78 ± 2.54	1.59 ± 0.70	7.25 ± 2.03	4.37 ± 1.48	4.12 ± 1.29	9.05 ± 2.03	3.56 ± 0.98b	13.19 ± 2.76	49.73 ± 6.81
Education level of father
Junior high school or below	2,022(55.6)	10.70 ± 2.55**	1.56 ± 0.71	7.19 ± 2.05**b	4.42 ± 1.48**b	4.18 ± 1.29**b	8.99 ± 2.08**	3.47 ± 0.98b	13.25 ± 2.78	49.65 ± 7.10b
High school	804(22.1)	10.93 ± 2.49**	1.59 ± 0.71	7.36 ± 1.99**b	4.17 ± 1.41**b	3.97 ± 1.16**b	8.76 ± 1.92**	3.42 ± 0.89b	13.26 ± 2.78	49.15 ± 6.75b
College or above	810(22.3)	11.26 ± 2.48**	1.59 ± 0.69	7.47 ± 1.93**b	3.96 ± 1.24**b	3.90 ± 1.14**b	8.74 ± 1.81**	3.40 ± 0.86b	13.36 ± 2.85	49.20 ± 6.49b
Education level of mother
Junior high school or below	2,189(60.2)	10.75 ± 2.53**	1.56 ± 0.71*	7.25 ± 2.05b	4.41 ± 1.48**b	4.17 ± 1.30**b	8.97 ± 2.06**	3.46 ± 0.97*b	13.23 ± 2.81	49.62 ± 7.10b
High school	680(18.7)	10.77 ± 2.51**	1.55 ± 0.68*	7.27 ± 2.03b	4.16 ± 1.38**b	4.00 ± 1.15**b	8.80 ± 1.93**	3.48 ± 0.95*b	13.37 ± 2.70	49.21 ± 6.76b
College or above	767(21.1)	11.33 ± 2.48**	1.63 ± 0.70*	7.44 ± 1.89b	3.92 ± 1.25**b	3.86 ± 1.09**b	8.71 ± 1.81**	3.37 ± 0.81*b	13.33 ± 2.85	49.11 ± 6.39b
Upbringing attitudes of father and mother
Unanimous	1,751(48.2)	10.61 ± 2.50**	1.51 ± 0.68**b	7.10 ± 2.01**	4.17 ± 1.43**	4.04 ± 1.24*	8.70 ± 1.91**b	3.40 ± 0.93**b	13.00 ± 2.76**	48.42 ± 6.76**b
Occasionally inconsistent	1,708(47.0)	11.05 ± 2.51**	1.62 ± 0.71**b	7.42 ± 2.00**	4.32 ± 1.42**	4.08 ± 1.22*	8.99 ± 2.01**b	3.47 ± 0.94**b	13.49 ± 2.77**	50.11 ± 6.73**b
Often inconsistent	177(4.9)	11.78 ± 2.64**	1.79 ± 0.83**b	7.92 ± 2.00**	4.55 ± 1.48**	4.30 ± 1.33*	9.67 ± 2.30**b	3.62 ± 0.95**b	14.07 ± 3.08**	53.01 ± 7.76**b
Main educational methods
Natural education	190(5.2)	10.52 ± 2.54**	1.62 ± 0.76*b	6.97 ± 2.00**	4.60 ± 1.52**	4.13 ± 1.35*b	9.23 ± 2.32**b	3.46 ± 1.01b	13.23 ± 3.00**	49.82 ± 7.54**
Persuade education	3,013(82.9)	10.80 ± 2.50**	1.56 ± 0.69*b	7.26 ± 1.99**	4.23 ± 1.42**	4.04 ± 1.22*b	8.80 ± 1.94**b	3.43 ± 0.92b	13.18 ± 2.74**	49.06 ± 6.73**
Scolding education	433(11.9)	11.54 ± 2.62**	1.66 ± 0.79*b	7.62 ± 2.14**	4.30 ± 1.45**	4.24 ± 1.29*b	9.33 ± 2.09**b	3.53 ± 1.02b	13.98 ± 2.97**	51.87 ± 7.21**
Caregivers of children
Parents	2,719(74.8)	11.02 ± 2.49**	1.61 ± 0.70**	7.41 ± 1.98**	4.23 ± 1.41*	4.03 ± 1.21**	8.84 ± 1.93*	3.44 ± 0.92	13.33 ± 2.73	49.51 ± 6.71
Non-parent caregivers	917(25.2)	10.45 ± 2.58**	1.48 ± 0.70**	6.94 ± 2.08**	4.35 ± 1.50*	4.18 ± 1.30**	9.00 ± 2.15*	3.44 ± 1.00	13.14 ± 2.99	49.22 ± 7.39
Only child status
Yes	1,144(31.5)	11.02 ± 2.48*	1.64 ± 0.72**	7.39 ± 2.00	4.16 ± 1.42**	3.97 ± 1.14**	8.91 ± 1.88	3.46 ± 0.90	13.28 ± 2.73	49.50 ± 6.43
No	2,492(68.5)	10.81 ± 2.55*	1.54 ± 0.69**	7.25 ± 2.02	4.31 ± 1.43**	4.12 ± 1.28**	8.87 ± 2.04	3.43 ± 0.95	13.28 ± 2.83	49.41 ± 7.09
Average monthly household income (CNY)
< 3000	430(11.8)	10.67 ± 2.59	1.52 ± 0.72	7.25 ± 2.05	4.31 ± 1.44**b	4.28 ± 1.35**b	9.09 ± 2.27*b	3.43 ± 0.99b	13.26 ± 2.98	49.78 ± 7.53
3000–4999	1,181(32.5)	10.84 ± 2.50	1.56 ± 0.71	7.31 ± 2.00	4.40 ± 1.51**b	4.16 ± 1.27**b	8.94 ± 2.06*b	3.45 ± 0.95b	13.25 ± 2.75	49.69 ± 6.94
5000–10,000	1,454(40.0)	10.99 ± 2.57	1.61 ± 0.71	7.31 ± 2.05	4.15 ± 1.36**b	3.98 ± 1.17**b	8.77 ± 1.80*b	3.42 ± 0.88b	13.31 ± 2.81	49.22 ± 6.73
> 10,000	571(15.7)	10.81 ± 2.40	1.54 ± 0.67	7.24 ± 1.94	4.21 ± 1.40**b	3.97 ± 1.19**b	8.89 ± 2.07*b	3.51 ± 1.00b	13.27 ± 2.74	49.19 ± 6.68

^a The current left-behind status;

^b Welch test is used when the homogeneity of variance is not satisfied in one-way ANOVA;

LBC Left-behind Children; BMI Body Mass Index; CNY Chinese Yuan; SD standard deviation.

*: p < 0.05; **: p < 0.01

Table 3.

Comparison of CSHQ scores of preschool children with different characteristics in screen time level and anxiety status of caregivers

Characteristics	N(%)	Bedtime resistance (Mean ± SD)	Sleep onset delay (Mean ± SD)	Sleep anxiety (Mean ± SD)	Sleep duration (Mean ± SD)	Night wakings (Mean ± SD)	Parasomnias (Mean ± SD)	Sleep disordered breathing (Mean ± SD)	Daytime sleepiness (Mean ± SD)	Total score (Mean ± SD)
Anxiety status of caregivers
Without anxiety	3,304(90.9)	10.83 ± 2.52**	1.57 ± 0.70	7.23 ± 2.00**	4.21 ± 1.41**a	4.01 ± 1.20**a	8.74 ± 1.87**a	3.39 ± 0.88**a	13.17 ± 2.74**a	48.93 ± 6.61**a
Mild anxiety	257(7.1)	11.55 ± 2.56**	1.60 ± 0.73	8.03 ± 2.01**	4.66 ± 1.51**a	4.58 ± 1.40**a	10.27 ± 2.60**a	3.97 ± 1.31**a	14.43 ± 3.09**a	54.70 ± 7.46**a
Moderate anxiety	65(1.8)	10.34 ± 2.43**	1.58 ± 0.75	7.60 ± 1.93**	4.92 ± 1.56**a	4.83 ± 1.64**a	10.32 ± 2.29**a	3.86 ± 1.16**a	13.71 ± 2.91**a	53.14 ± 7.42**a
Severe anxiety	10(0.3)	12.10 ± 3.31**	2.00 ± 0.94	8.00 ± 3.09**	4.60 ± 2.01**a	4.70 ± 1.42**a	10.70 ± 1.77**a	3.90 ± 1.10**a	16.00 ± 3.02**a	57.60 ± 10.53**a
Screen time level
Normal screen time (NST)	936(25.7)	10.79 ± 2.47	1.53 ± 0.69a	7.28 ± 1.99	4.20 ± 1.39*a	4.03 ± 1.23	8.71 ± 1.80**a	3.40 ± 0.89a	13.08 ± 2.68**	48.78 ± 6.45**a
Excessive screen time (EST)	1,662(45.7)	10.85 ± 2.51	1.59 ± 0.69a	7.27 ± 2.03	4.23 ± 1.41*a	4.08 ± 1.23	8.85 ± 2.03**a	3.45 ± 0.95a	13.25 ± 2.79**	49.32 ± 6.97**a
Seriously excessive screen time (SEST)	1,038(28.5)	10.98 ± 2.60	1.59 ± 0.73a	7.33 ± 2.02	4.36 ± 1.49*a	4.10 ± 1.26	9.08 ± 2.07**a	3.47 ± 0.96a	13.50 ± 2.89**	50.22 ± 7.08**a

^a Welch test is used when the homogeneity of variance is not satisfied in one-way ANOVA; Normal screen time (NST), ≤ 60 min; Excessive screen time (EST), 60 < EST ≤ 120 min; Seriously excessive screen time (SEST), > 120 min; SD standard deviation.

* p < 0.05; ** p < 0.01

As shown in Table 4, 3,252 subjects scored above the CSHQ cutoff, indicating potential sleep disturbance. The prevalence of sleep disturbances among preschool children is 89.4%. For the specific sleep disturbances based on each subscale, bedtime resistance (54.9%) and sleep anxiety (47.9%) are the most common types of sleep disturbances, followed by daytime sleepiness (21.3%), sleep duration (20.8%), parasomnias (16.5%), sleep-onset delay (12.6%), night wakings (12.2%), and sleep-disordered breathing (10.3%).

Table 4.

Sleep disturbances among preschool children (N = 3636)

	Mean of score	SD	Cutoff value	Numbers above the CSHQ cutoff	Prevalence %
CSHQ subscales
Bedtime resistance	10.87	2.53	10.84	1,997	54.9
Sleep onset delay	1.57	0.70	2.31	458	12.6
Sleep duration	4.26	1.43	5.27	756	20.8
Sleep anxiety	7.29	2.01	7.79	1,741	47.9
Night wakings	4.07	1.24	5.29	444	12.2
Parasomnias	8.88	1.99	10.61	600	16.5
Sleep disordered breathing	3.44	0.94	4.50	376	10.3
Daytime sleepiness	13.28	2.80	15.24	775	21.3
Total (CSHQ)	49.44	6.89	41.00	3,252	89.4

CSHQ the Children’s Sleep Habits Questionnaire; SD standard deviation

Hierarchical multiple linear regression analysis of sleep disturbances with different variables

Table 5 displays the final results of hierarchical multiple linear regression analysis (results of step 3). According to the results of collinearity diagnosis, our model did not contain collinearity problems. In step 1, younger age and higher birth weight (> 4,000 g) significantly predicted higher CSHQ scores. In step 2, higher scores of hyperactivity/inattention, conduct problems, emotional symptoms, and peer problems on the SDQ suggested higher CSHQ scores, and lower scores of prosocial behavior on the SDQ, as well as food preference, fear of feeding, parental misperception, and organic disease on the IMFeD, were significantly associated with higher CSHQ scores. In step 3, younger age, higher score of hyperactivity/inattention, conduct problems, emotional symptoms, and peer problems in SDQ, and inconsistent upbringing attitudes of father and mother, improper educational methods, caregivers of children, and higher level of caregivers’ anxiety all suggested higher CSHQ scores, and the lower score of prosocial behavior in SDQ, as well as food preference, fear of feeding, parental misperception, and organic disease in IMFeD, were associated significantly with higher CSHQ scores. Sociodemographic factors (step 1) account for approximately 1.0% (R² = 0.01, p < 0.01), which indicates that sociodemographic variables explain 1.0% of the variance in the CSHQ score. Children’s associated factors (step 2) and caregiver-associated factors (step 3) explained 23.8% and 2.3% of the variation in the CSHQ score, respectively. In the final model (step 3), the ages of children, children’s emotional/behavioral problems, children’s feeding difficulties, upbringing attitudes of fathers and mothers, main educational methods, non-parent caregivers, and caregiver anxiety contributed significantly to the CSHQ total score, accounting for approximately 27.1% (R² = 0.271) of the variation.

Table 5.

Associations between sociodemographic, anxious status of caregivers, emotional/behavioral problems as well as feeding difficulties and CSHQ based on hierarchical linear multiple regression analysis

Variable	Sleep disturbances
	Standardized coefficient	B (SE)	95% Confidence interval
Sociodemographic factors
Female	0.008	0.117(0.201)	− 0.277, 0.511
LBC	0.009	0.129(0.231)	− 0.324, 0.582
Age (years)	− 0.051	− 0.388(0.112) **	− 0.608, -0.169
Birth weight
Underweight	− 0.005	− 0.161(0.502)	− 1.145, 0.823
Overweight	− 0.027	− 0.708(0.377)	− 1.446, 0.031
Normal (ref.)	/	/	/
Only child status	− 0.017	− 0.251(0.219)	− 0.680, 0.179
Average monthly household income (CNY)
< 3,000	0.009	0.200(0.395)	− 0.574, 0.975
3,000–4,999	0.012	0.176(0.312)	− 0.436, 0.787
5,000–10,000	0.011	0.158(0.295)	− 0.421, 0.737
> 10,000 (ref.)	/	/	/
Children’s associated factors
Adiposity indicator (BMI)
Obesity	0.003	0.078(0.384)	− 0.675, 0.832
Overweight	0.001	0.024(0.341)	− 0.646, 0.693
Normal (ref.)	/	/	/
Screen time level
NST	− 0.026	− 0.410(0.274)	− 0.948, 0.128
EST	− 0.016	− 0.217(0.240)	− 0.688, 0.253
SEST (ref.)	/	/	/
SDQ scales
Hyperactivity/inattention	0.078	0.254(0.054) **	0.148, 0.359
Prosocial behavior	− 0.038	− 0.122(0.052) *	− 0.225, -0.020
Conduct problems	0.097	0.445(0.080) **	0.287, 0.602
Emotional symptoms	0.166	0.632(0.065) **	0.505, 0.759
Peer problems	0.037	0.160(0.070) *	0.023, 0.297
IMFeD scales
Poor appetite	− 0.035	− 0.069(0.040)	− 0.147, 0.009
Food preference	− 0.070	− 0.141(0.037) **	− 0.214, -0.067
Poor eating habit	− 0.032	− 0.055(0.036)	− 0.126, 0.015
Fear of feeding	− 0.053	− 0.248(0.095) **	− 0.434,  − 0.061
Parental misperception	− 0.046	− 0.154(0.065) *	− 0.282,  − 0.026
Organic disease	− 0.106	− 0.712(0.122) **	− 0.951,  − 0.473
Caregivers’ associated factors
Education level of father
Junior high school or below	− 0.014	− 0.200(0.358)	− 0.901, 0.501
High school	− 0.020	− 0.331(0.348)	− 1.013, 0.351
College or above (ref.)	/	/	/
Education level of mother
Junior high school or below	0.018	0.253(0.359)	− 0.451, 0.957
High school	− 0.001	− 0.025(0.361)	− 0.732, 0.682
College or above (ref.)	/	/	/
Upbringing attitudes of father and mother
Often inconsistent	0.052	1.658(0.485) **	0.708, 2.608
Occasionally inconsistent	0.048	0.669(0.208) **	0.261, 1.076
Unanimous (ref.)	/	/	/
Main educational methods
Natural education	− 0.028	− 0.865(0.524)	− 1.891, 0.162
Persuade education	− 0.038	− 0.695(0.325) *	− 1.333, -0.057
Scolding education (ref.)	/	/	/
Caregivers of children
Non-parent caregivers	− 0.077	− 1.223(0.268) **	− 1.749, -0.697
Parents(ref.)	/	/	/
Anxious status of caregivers
Severe anxiety	0.045	5.900(1.888) **	2.199, 9.601
Moderate anxiety	0.008	0.395(0.761)	− 1.098, 1.888
Mild anxiety	0.105	2.811(0.399) **	2.029, 3.594
Without anxiety (ref.)	/	/	/

LBC left-behind children; normal screen time (NST), ≤ 60 min; excessive screen time (EST), 60 < EST ≤ 120 min; seriously excessive screen time (SEST), > 120 min; BMI body mass index; CNY Chinese Yuan; CSHQ the Children’s Sleep Habits Questionnaire; SDQ the Strength and Difficulty Questionnaire; IMFeD Identification and Management of Feeding Difficulties; SE, standard error.

**: p < 0.01; *: p < 0.05

Discussion

This study described the prevalence of sleep disturbances and explored the associated factors comprehensively based on a large sample in rural areas of China. The prevalence of sleep disturbances among preschool children in rural areas of China in our study was 89.4%, which is comparable to previous reports in Asian peers [8, 9, 29, 42]. The most common types were bedtime resistance and sleep anxiety, which are consistent with previous studies [9, 29, 42].

Association between factors related to caregivers and children and sleep problems

Due to cultural habits, cosleeping is quite common in Asian countries, including China, Japan [8], and Korea [43], which partly contributes to the scores of bedtime resistance and sleep anxiety, since these subscales share some items related to cosleeping (needs parents in room to sleep; afraid of sleeping alone). In addition, in some rural areas of China, caregivers are more likely to take measures including cosleeping to ensure that the children can sleep peacefully, and different sleep habits between caregivers and children when sleeping together (including irregular sleep schedules and the use of electronic products among caregivers [44]) may be associated with children’s sleep problems. For example, some caregivers are used to watch TV before sleeping, which may influence children’s sleep quality. We found that the relationship of caregivers and children was associated with the CSHQ score of children in the final model. Compared with children nurtured by their parents, children whose caregivers were other relatives tended to have lower CSHQ scores, which indicated that the characteristics of different caregivers needed to be explored in more detail. The potential reasons may be as follows. (1) Other caregivers did not want others to feel that they did not take care of the children well enough, which may bring a certain amount of information bias. (2) Children raised by other relatives may be more accustomed to sleep alone. The correlations between children’s ages and total CSHQ scores may be correlated with developmental changes in the sleep patterns and nervous systems of children [8, 45]. In terms of education styles, our study found that inconsistent upbringing attitudes and scolding education were related to higher CSHQ scores than unanimous attitudes and a more lenient education style (persuading education). Previous studies have proven that inconsistent child-rearing attitudes may increase the risk of emotional/behavioral problems in children, which are related to sleep disturbances [9, 46]. In addition, different upbringing attitudes may make it difficult for children to establish healthy sleep habits because of mixed messages from parents, which could cause confusion and resistance among children [9]. Furthermore, parents or other caregivers should maintain consistency in their educational practices to prevent potential confusion and anxiety in their children, which may relate to their sleep quality. In addition, instead of scolding or beating, education methods should mainly be based on persuasion and setting an example, in case of adverse effects in children’s physical and mental health that are related to their sleep quality.

Our survey found that anxiety levels of caregivers were positively related to children’s CSHQ scores. Previous studies have shown that the association between parental anxiety and sleep problems in children may be bidirectional [1, 18, 47, 48]. Children are likely to be exposed to parental mood disturbances, which could create an environment of uncertainty and insecurity that may contribute to sleep problems [1, 49]. Furthermore, the existence of sleep problems in children may make their parents feel more stressful and thus make the children more likely to feel distressed [50], which could generate or even aggravate their anxiety. The potential correlation between the anxiety status of caregivers and sleep problems among preschool children needs to be further explored in future cohort studies. Our study has implied the importance of psychological interventions for caregivers that could benefit their children.

Association between children’s emotional/behavioral problems and sleep disturbances

We found significant correlations between children’s emotional/behavioral problems and sleep disturbances among preschool children. A previous study showed that emotional/behavioral problems were positively correlated with poor sleep quality and that higher sleep quality could reduce the risk of suicidal ideation in patients with emotional/behavioral problems [51]. In addition, clinical evidence also indicates that sleep and emotion interact, and most psychiatric and neurological disorders expressing sleep disruption tend to display corresponding symptoms of affective imbalance [52, 53]. Furthermore, studies have also implied that the association between emotional/behavioral problems and sleep disturbances could be bidirectional [4, 54, 55]. Compared with previous studies in school-aged children [29] and urban preschool children [9], we also found significant correlations between prosocial behavior and CSHQ scores. The results from our study indicated that the characteristics of sleep disturbances in preschool children in rural areas may differ from those of urban preschoolers, and the potential association between prosocial behavior and sleep disturbances in preschool children should be further explored by high-quality cohort studies.

Association between children’s feeding difficulties and sleep disturbances

In addition, our survey suggested that feeding difficulties, including food preference, fear of feeding, parental misperception, and organic disease, may be associated with children’s sleep quality. Regarding the potential mechanisms, it is possible that food preference could reduce the food variety of children, which may be associated with a series of health problems, including poor sleep quality. Vitamins and minerals, fruits and vegetables, and meat items showed a positive association with sleep quality, whereas high fat and high sugar items showed a negative association, and children with other unhealthy diets, including eating outside the home, eating fast food, eating in front of the TV, or eating alone, were more likely to have shorter sleep and poorer sleep quality [56]. A previous study has shown that regular opportunities for children to interact with their family through family meals may also reduce stress, which could improve children’s sleep quality [57, 58]. In addition, awareness of a healthy diet among parents or other caregivers is also important. Parental misperception about their children’s food consumption could lead to unhealthy feeding behaviors, which may be associated with children’s sleep quality [59]. In rural areas of China, many parents believe that overweight children will be healthier and taller in the future [39, 60], and such overfeeding may lead them to believe that children are difficult to feed [39]. Furthermore, bad appetite, frequent vomiting, or diarrhea caused by specific organic diseases could also reduce children’s food consumption and organism resistivity, which could be related to negative health outcomes, including bad sleep quality. The potential relationship between children’s feeding difficulties and sleep disturbances needs to be further explored in the future. Our study results indicated the potential possibility of building healthy eating habits and establishing a correct concept of feeding for caregivers, especially in rural areas of China, for the improvement of sleep quality in children.

Limitations

However, our study also has some limitations. First, limited by the cross-sectional study design, our study results cannot ensure the causal relationship between sleep disturbances in preschool children and the associated factors discovered. Future well-designed cohort studies are needed to verify the results of our research. Second, since our study sites were selected from rural areas of Anhui Province in China, the generalizability to other populations located in other provinces in China or abroad was limited, and relatively high prevalence of sleep disturbances among preschool children (89.4%) also indicated that the cutoff value of CSHQ total score (41) generated by US sample [28] should be adjusted for Chinese preschool children in future studies since parenting practices, culturally determined values, lifestyle factors, etc. could have a great impact on sleep behaviors and practices of children [13]. To be specific, several cultural and psychosocial factors may contribute to more sleep problems in Chinese children comparing to their US counterparts such as cosleeping, crowded housing, and chronic stress from homework overload or excessive expectations from their caregivers for their children’s future success [7, 61]. Third, based on the number of variables and scales included, our study aimed to explore the potential factors related to sleep disturbances among preschool children, which was limited by specific and deep analysis of certain factors. Fourth, the Cronbach’s alpha values for the CSHQ (0.668), SDQ (0.689), and SAS (0.558) indicated the relatively low reliability of these questionnaires. The potential reason may be the information bias with self-reporting and face-to-face interviews. In face-to-face interviews, some caregivers may not want others to feel that they did not take care of the children well enough, which could cause inaccurate information in some items; in self-reporting, lack of effective supervision could also impact the quality of data. Fifth, limited by the education level of caregivers, parents or other caregivers may not give the information of their children precisely, and different methods to investigate information (self-reporting and face-to-face interview) may cause a certain amount of information bias, since caregivers may have a deviation in understanding the items when answering by themselves. Finally, considering that the CSHQ is a screening tool for existing sleep disturbances, our study results could not certify the existence of sleep disturbances but indicate the likelihood of a potential sleep disorder, which could further provide information serving as the starting point for a more detailed clinical evaluation [28].

Conclusion

In conclusion, the prevalence of sleep disturbances among preschool children in rural areas of China is quite high, and the potential risk factors are multidimensional, including child-related factors (age, emotional/behavioral problems, and feeding difficulties) and caregiver factors (upbringing attitudes, main educational methods, caregivers of children, and anxiety status of caregivers). Future well-designed and large-sampled cohort studies are needed to explore the specific risk factors and targeted interventions for both children and their caregivers.

Author contributions

Dr. YS: designed the study and made the final revision; TZ, KX, HL, XC, and GQ: collected the data; TZ and KX: wrote the manuscript; HL and XC: helped in manuscript writing; YW and JZ: critically reviewed the manuscript and helped in revising the manuscript. All the authors read and approved the final manuscript.

Funding

This work was supported by the National Natural Science Foundation of China (Grant number: 81872704).

Availability of data and material

The raw data required to reproduce these findings cannot be shared at this time as the data also forms part of an ongoing study.

Code availability

N/A.

Declarations

Conflict of interest

The authors declare that they have no conflict of interest.

Ethical approval

This study was approved by Ethics Committee of Anhui Medical University (20180402) and followed the principles of informed consent and strict confidentiality.

Consent to participate

Informed consent was obtained from all individual participants included in the study.

Consent for publication

Written informed consent for publication was obtained from all the participants.