Predicting Insomnia in Young Adults: The Role of Autism Symptom Severity, Sensory Atypicality and Intolerance of Uncertainty.

Background: Poor sleep, primarily insomnia symptoms, is highly prevalent in autism across the lifespan and has a significant negative impact on all aspects of functioning, often over and above core autism symptoms. Therefore, insomnia constitutes an important treatment target, however, factors underlying insomnia problems in autism are currently poorly understood. Both autism symptom severity and atypical sensory behaviour are related to poor sleep. Intolerance of uncertainty is a trans-diagnostic risk factor predisposing individuals with and without autism to experience higher levels of stress and anxiety, and might therefore be related to insomnia problems. However, its role in ASD has not been previously explored.

Objectives: To examine the role of autism symptom severity (ASS), sensory seeking behaviour (SSB) and intolerance of uncertainty (IU) in predicting insomnia symptoms in young adults with and without autism.

Methods: Participants enrolled in the Autism CRC School Leavers longitudinal study completed online questionnaires on ASS (Autism Quotient [AQ]-28), SSB (Repetitive Behaviour Questionnaire [RBQ]-2A-SSB), depression (Patient Health Questionnaire-9), anxiety (DSM-5- Dimensional Anxiety Scales: Generalized Anxiety), IU (Intolerance of Uncertainty Scale [IUS]-12) and sleep (Pittsburgh Sleep Quality Questionnaire [PSQI]). There were 53 young adults (M_age=18.39 years, SD=2.44; 56.6% male) with 65.4% (N=52) reporting an autism spectrum diagnosis; those with a diagnosis (M_AQ-28 = 76.79, SD=10.92; 73.5% male) had a significantly higher AQ-28 score (P<.001) than the non-autism group (M_AQ-28= 58.83, SD=14.15).

Results: 56.6% (N=53) of individuals had a sleep problem (PSQI>5), including 55.9% of those with, and 61.1% of those without autism (χ² (1, N=52) = .005, P=.946). Neither age nor gender were significantly associated with SSB (P=.707; P=.788), IU (P=.785, P=.743), PSQI (P=.721, P=.205), anxiety (P=.396, P=.127) or depression (P=.855, P=.633), respectively. Consistent with more males with autism, the AQ-28 score was significantly associated with gender (P=.039), but not age (P=.673). Depression and anxiety scores were all moderately or strongly and significantly associated with ASS, SSB and IU. A hierarchical regression predicting insomnia (PSQI score) was conducted, with scores for AQ-28 entered at Step 1, RBQ-2A-SSB entered at Step 2, and IUS-12 entered at Step 3. The final model was significant F (3,49) = 7.059, P<.001, R²=30.2 % of variance in sleep accounted for (adjusted R²=25.9%). At Step 1 (AQ-28) ΔR²=12.8% (P=.009), at Step 2 (RBQ2A-SSB) ΔR²=9.1% (P=.019), and at Step 3 (IU-12) ΔR²=8.3% (P=.020); IU was a significant unique predictor of insomnia (β=.349; P=.020) in the final model.

Conclusions: Autism symptom severity, sensory atypicality and intolerance of uncertainty together accounted for 30% of variance in insomnia scores, with intolerance of uncertainty being an independent predictor. This supports increased autism symptoms and intolerance of uncertainty as risk factors for insomnia. Our study thus provides the first evidence pointing to the trans-diagnostic factor, intolerance of uncertainty, which underlies susceptibility to anxiety and depression, as a potential key factor in the development of insomnia in autism. Further exploration, including models with anxiety and depression, is indicated to understand how insomnia develops in autism, thus informing the development of both prevention and treatment approaches for insomnia across the lifespan.