Predictors of Poor Sleep Quality in Youth and Young Adults on the Autism Spectrum

Background: Poor sleep quality, primarily insomnia is known to be common in children and younger adolescents with autism, with up to 80% of individuals being affected. Emerging research indicates that sleep difficulties can have chronic course, continuing into older adolescence and adulthood.  Research has started to explore the risk factors behind sleep difficulties in the autism population with findings suggesting that they may be associated with both core symptoms of autism, including sensory sensitivities and repetitive behaviours, and with comorbid conditions such as ADHD, anxiety, depression and gastro-intestinal symptoms. Recent proposals have particularly singled out the key role of anxiety and depression, at least in relation to insomnia symptoms.  

Objectives:   Our aim was to explore the relative contributions of anxiety and depression, sensory sensitivities, emotional awareness and somatic symptoms in predicting sleep quality in young people with autism.

Methods:   Participants were drawn from the Autism CRC Longitudinal Study of School Leavers with autism (www.autismcrc.com.au), which commenced in 2015 and is ongoing. Currently 38 youth with autism, aged 15-25 years, 24 males (M_age = 17.46 years, SD = 2.15) and 14 females (M_age = 18.86 years, SD = 3.23) have completed demographic information, the Autism Quotient-28 item form (autism traits), COMPASS (somatic symptoms), DSM-5 dimensional anxiety scale, PHQ-9 (depression), Glasgow Sensory questionnaire (GSQ), and Levels of Emotional Awareness – short form (LEAS) and the Pittsburgh Sleep Quality Index (PSQI) as a measure of sleep quality. 

Results: Full data for 32 participants are available to date; square root transformations were used for non-normal variables.  Correlational analyses showed that all variables except autism traits, r = .31, p > .05, and emotional awareness, r = -.10, p > .05, were significantly (all p < .001) and strongly (all r > .50) associated with sleep quality. Next, a hierarchical multiple regression with sleep quality as the dependent variable was conducted. ASD traits were entered at step 1, and explained 10% of the variance, F (1, 30) = 3.34, p = .078. Sensory sensitivity was entered at Step 2, explaining an additional 19.2% of the variance, F change (1, 29) = 7.87, p = .009. The addition of anxiety and depression at step 3 explained an additional 23.8% of variance, R squared change = .24, F change (2, 26) = 6.84, p = .004.  Somatic complaints was entered at step 4 but did not explain any additional variance, F change (1, 26) = 0.02, p = .891. The final model explained 53.1% of the variance in sleep quality, F (5, 26) = 5.88, p < .001, with depression being a unique significant predictor, beta = .43, p = .047.  

Conclusions:   Preliminary results indicate that sensory sensitivity, anxiety and depression are strongly contributing to the maintenance of poor sleep quality in autism. This is in line with research in non-autistic populations suggesting that psychopathology and sleep are reciprocally related, while sensitivity to environmental stimuli can have a negative effect on sleep. These findings indicate potential future avenues for management of sleep problems in the autism population.