1. Subjects
This study adopted a case-control study design. Clinical data and information were extracted from a larger dataset developed in our previous research project that
was approved by the Human Ethics Committee of the Second Xiangya Hospital of Central South University. The children with ASD in the dataset were recruited from the Outpatient Department of the Mental Health Institute at the Second Xiangya Hospital of Central South University, Changsha, Hunan Province and four training schools for autistic children (the Elim Training Center for Children with Autism in Qingdao, Shandong Province, and the Xinyuan Training Center for Children with Autism, the Aimeng Training Center for Children with Autism, and the Qihang Training Center for Children with Autism in Changsha, Hunan province). The subjects in the control group were recruited from two kindergartens in Changsha. After obtaining written informed consent from the children's parents or their legal guardians, the children in the ASD group were independently diagnosed by two senior psychiatric doctors, according to DSM-IV-TR criteria for autistic disorder. The child's past illnesses were investigated, and children who had an organic disease of the nervous system or a psychiatric disorder, such as childhood schizophrenia or intellectual disability, were excluded from the study. Children with ASD with a chromosomal abnormality, such as Fragile X syndrome, and children with ASD and TD children whose parents had a disease were also excluded.
Subjects aged 4 to 10 years in the dataset were selected. The Social Responsiveness Scale (SRS) was used as a complementary diagnostic test in the study, and a total raw score of 56.5 was set as the cutoff score[22]. Subjects who scored above 56.5 in the control group and scored below 56.5 in the case group were excluded. Ultimately, a total of 440 children with ASD and 344 age-matched TD children were included in the study. All of the included subjects were Han Chinese. In the case group, 383 subjects were male and 57 were female; their mean age was 5.13±1.12 years. In the control group, 166 subjects were male and 178 were female; their mean age was 4.84±0.92 years.
2. Data collection and clinical assessment
The parents of the children were asked to complete a self-administered structured questionnaire that included family demographic information. The questionnaire also included some widely used scales to measure some clinical features of the enrolled children and parents.
2.1 Sleep problems
The Children’s Sleep Habits Questionnaire (CSHQ) was used to measure the sleep problems in the subjects in both groups. The CSHQ is a parent-reported questionnaire that uses a 3-point Likert scale and provides a total score and eight subdomain scores on bedtime resistance, sleep-onset delay, sleep duration, sleep anxiety, night waking, parasomnia, sleep-disordered breathing and daytime sleepiness. A higher score indicated more sleep problems. Based on the original report, the total score cutoff was set at 41[23]. However, some studies have indicated that a cutoff of 41 might overestimate sleep problems and suggested adopting the cutoff of 48 instead. Considering 48 as the cutoff value of the CSHQ can conservatively define sleep problems; thus, for the subsequent results, we considered a cutoff value of 48[5].
2.2 Parental quality of life
The Short Form 36 Health Survey Questionnaire Version 2.0 (SF-36v2) was adopted to measure parental QOL. The questionnaire categorizes QOL into 8 dimensions, including physical functioning, role-physical, pain, general health as energy/vitality, social functioning, role-emotional, and mental health. Two new comprehensive dimensions, the physical health summary (PCS) and mental health summary (MCS) scores, can be derived from the SF-36v2. A higher score on the SF-36v2 indicated better QOL. In the current study, we used a Chinese norm (Hong Kong) to provide the T scores for the SF-36v2[24], and the internal consistency and test-retest reliability of the Chinese form of the SF-36v2 is acceptable (coefficients 0.66 to 0.89).
2.3 Childcare burden
The Zarit Caregiver Burden Interview (ZBI) was used to measure childcare burden in the parents. Lu L et al.[25] translated the ZBI into Chinese from the English and Japanese versions, and according to their report, the internal consistency of the Chinese version of the ZBI was 0.875. The ZBI contains 22 items scored on a 5-point Likert scale to measure the health, psychological well-being, social life, finances of the caregivers, and the relationship between caregivers and patients. Higher ZBI scores represented heavier childcare burden.
2.4 Social impairments in ASD
In addition to being used as a complementary diagnostic test, the SRS was also adopted to measure social impairments in the children with ASD. The SRS was developed to assess social behaviors in children[26]. The SRS is a parent-reported 65-item questionnaire that uses a 4-point Likert scale and includes five subdomains: awareness, cognition, communication, motivation, and autistic mannerisms. It is a widely used scale in autism research. It has convergent validity with the Autism Diagnostic Observation Schedule and the Autism Diagnostic Interview-Revised, ranging from 0.35 to 0.58, and external validity with the Vinland Adaptive Behavior Scale and the Child Behavior Checklist, ranging from 0.35 to 0.64[27]. A higher score on the SRS indicates greater impairments in social behavior.
3. Statistical analysis
We used Student’s t-test to compare the demographic characteristics, children’s CSHQ scores, and parental SF-36v2 scores between groups. A chi-square test was used to compare the incidence of sleep problems between the ASD and TD children. To test the impact of children’s sleep problems on their parents’ QOL, two linear regression models were used. The first linear regression model adjusted for the child’s age, parental age and childcare burdens as covariates and was separately conducted in the ASD and TD groups. The second model adjusted SRS scores in the ASD group based on the first model to test whether sleep problems in children with ASD impacted parental QOL independent of social impairments. Additionally, if the second linear regression model found that the CSHQ scores had a significant effect on parental QOL, a pathway analysis would be conducted to explore the relationships among sleep problems, social impairments, childcare burden and parental QOL. As mentioned above, our previous study indicated that the social impairment associated with ASD affected parents’ QOL due to the parental childcare burden. Considering that sleep problems in children can affect their social functioning [10] and increase the childcare burden on their parents [28], we hypothesized that sleep problems (CSHQ score) will exert direct impacts on parental QOL (SF-36v2 score), while social impairment (SRS score) and childcare burden (ZBI score) may act as mediators between parental QOL and sleep problems. An initial pathway recursive model was constructed based on our hypothesis and the findings of a previous study. The CSHQ score was characterized as an exogenous variable, and the SRS score, ZBI score and SF-36v2 score were characterized as endogenous variables.
All statistical analyses were conducted in IBM SPSS Statistics (Version 22.0) and IBM SPSS AMOS (Version 24.0). The significance level of all tests in the study was set at p<0.05 (two-tailed).