Results
As outlined in Figure 1, the study sample consisted of 305 participants of the 802 consecutively recruited help-seeking young people aged between 12 – 25 years (x̄age: 18.36, SD: 3.01, 75% female) of whom 64% (n = 194/305) reported moderate depression, 29% (n = 89/305) severe and 7% (n = 22/305) very severe depression (as defined by QIDS-C16 cut-offs). The average depression severity score was 14.93 (SD: 3.23, range 11 – 26) with no significant gender difference. In this sample, 34% (n = 104/305) were in some form of employment, 66% (n = 201/305) were in some form of education, with 21% (n = 65/305) not participating in either. Tobacco was the substance most frequently used on a daily basis with 36% (n = 110/305) of the sample identifying as daily smokers: 30% (n = 33/110) of these daily smokers were aged under 18 years. Amongst this sample, 6% (n = 18/305) reported daily alcohol consumption and 10% (n = 30/305) reported daily cannabis use. A total of 50% (n = 152/305) presented 'at risk' for tobacco misuse, 19% (n = 57/305) for alcohol and 31% (n = 93/305) for cannabis. There were no significant differences between gender and substance risk. Sleep disturbances were common: 72% (n = 219/305) reported sleep-onset insomnia, 33% (n = 100/305) experienced mid-nocturnal insomnia, 23% (n = 69/305) early-morning insomnia and 15% (n = 46/305) hypersomnia. A total of 70% (n = 213/305) reported poor sleep quality and 19% (n = 59/305) reported persistent fatigue.
Delayed sleep onset was reported among 18% (n = 56/305, Mdnbedtime 03:00, Mdnwaketime 09:30) of the primary sample. Normal sleep onset was reported by 82% (n = 249/305, Mdnbedtime 22:00, Mdnwaketime 07:00). For comparison among the non-depressed young people (e.g. QIDS-C16 range 0 – 10 and Clinical Stage < 2) in the overall help-seeking sample, the prevalence of delayed sleep was 11% (n = 42/417). Outlined in Table 1, a higher level of all types of substance misuse risk was seen in the delayed onset group. There were no significant differences in symptomatology, personality factors, disability or quality of life. Upon going to bed, delayed sleepers did not report taking longer to fall asleep than the normal phenotype; however, the delayed group reported shorter sleep duration, spending two hours less in bed (x̄: 6.71 vs. x̄: 8.75) and one hour less asleep (x̄: 5.89 vs. x̄: 6.89). The delayed onset sleepers did not report poorer sleep quality (75%, n = 42/56 vs. 69%, n = 171/249) or more persistent fatigue (16%, n = 9/56 vs. 20%, n = 50/249) than the normal sleep phenotype. There was no interaction of age group with these associations, however, in those aged between 12 – 17 years, alcohol (OR: 2.61, 95% CI: 0 .73 - 9.54) and cannabis misuse (OR: 2.07, 95% CI: 0.72 - 5.92) did not have statistically significant associations with delayed sleep onset due to fewer substance users, but had very similar effect sizes to the overall group.
Table 2 displays the logistic regression analysis evaluating the factors associated with delayed sleep onset. Age and fun-seeking behaviour were entered as continuous covariates; gender, education, and substance risk were dichotomous covariates. Step 1 confirmed that age and gender were not significantly associated with delayed sleep onset. The observed association of no education status and lower fun-seeking with delayed sleep phenotype was subsequently attenuated to non-significant after adjustment for substance misuse. In step 3, tobacco misuse risk emerged as the only factor significantly associated with delayed sleep onset (adjusted OR: 2.28, 95% CI: 1.04 - 5.01). The sensitivity analysis using the more restricted exposure categories of 'normal' (21:00 – 01:59) and 'delayed' (02:00 – 05:00) showed no differences in results.