In this sample of help-seeking depressed young people, habitual delayed sleep onset was common: Nearly one in five participants with depression reported going to bed after 2:00 a.m. As the prevalence of delayed sleep onset in the general population of young people is less than 10%, the same proportion seen in the non-depressed help-seeking young people, these results suggest rates may be higher in young people experiencing moderate – severe depression Although the current participants were not clinically assessed for, or diagnosed with, DSPD, their sleeping patterns align with elements of the current diagnostic criteria
. In this sample, delayed sleep onset appeared to result in an average sleep duration of less than six hours and two hours less in bed. Despite this, the delayed sleepers did not report significantly poorer sleep quality or more persistent fatigue. This is consistent with the profile of DSPD found in previous studies of young people
[11, 16, 49]. Given that short sleep duration of less than six hours is associated with the chronicity of mental disorders in young people
, and a range of poor health outcomes, the delayed onset group may represent a particularly high risk group of depressed young people, though follow-up is needed. Delayed sleep onset accompanied by short sleep duration is consistent with the Clinical Staging Model for mood disorders and has been suggested to signify a circadian pathophysiological profile of depression that may have a different prognosis and response to treatment
[4, 38, 50]. In our studies of young people in the early stages of mood disorders, circadian disruption seems common and a potential avenue for targeted treatments.
In this sample, delayed sleepers did not have significantly different symptom profiles, and, importantly, no greater levels of depression symptom severity. This suggests that delayed sleep isn’t a marker for depression severity in a depressed sample but rather something qualitatively different about this group. Neither did they differ in age, self-reported functioning, employment and disability. Suggested differences in education and function did not withstand adjustment for multiple testing.
In this help-seeking group, delayed sleepers were significantly more likely to report alcohol, and tobacco misuse, with the latter emerging as the strongest, and only independent predictor of delayed sleep onset. Although this might suggest a predisposition to a hedonic or delinquent lifestyle among young people with delayed sleep
, no associated personality traits were found which undermines this line of thought. Instead, this association is consistent with other authors’ observations that people who smoke are at greater risk of experiencing difficulties falling asleep and mental health problems
. There is some biological plausibility too in that nicotine stimulates the cholinergic neurons of the basal forebrain component of the arousal system
 which are inhibited during both REM and non-REM sleep. Nicotine injections also increase awakenings in rodents
. In observed conditions using polysomnography, smokers exhibit an extended sleep latency
[55, 56] and spectral EEG changes reflecting this arousal
. Although cross-sectional, these results suggest that tobacco smoking is associated with delayed and short sleep in depressed youth, which can be helpful when considering treatment interventions. Furthermore, the prevalence of daily tobacco use (38%) in this group is exceptionally high when compared to the national prevalence of 16%
, particularly for those under 18 where the national prevalence is 4%. This finding is consistent with other help-seeking samples of young people
 and suggest that this group have an elevated risk of poor health.
Firstly, young people in the early stages of mental illness who present to care often do so in times of crisis, in which they present with a mixture of symptoms. At this early stage, the determination of a primary diagnosis based on conventional criteria, such as depression, may be somewhat unreliable using cross-sectional symptomatology. The high proportion of depressed females in this sample is not necessarily consistent with population prevalence
 but is consistent with help-seeking youth. Scott et al.
 found that males presenting to headspace services were more likely to have a primary diagnosis of behavioural or developmental disorders, whereas females were more likely to present with depressive symptoms. Secondly, this study utilised a cross-sectional measurement of sleep in which participants may have reported recent sleep as opposed to habitual
[60, 61]. As groups were classified based on sleep onset, it may be possible that different associations will emerge using other time points or sleep measures. The addition of objective sleep measures such as actigraphy, or, less practically, polysomnography would significantly improve the reliability and validity of the sleep data. Future studies would also benefit from the use of a chronotype questionnaire, such as the Morningness-Eveningness Questionnaire
 or Munich Chronotype Questionnaire
. Although the levels of depression experienced by this group may have altered their perceptions of sleep, there were no differences in depression (or other symptom scores) between the delayed and normal onset sleep groups which undermines any suggestion of response bias. However, the absence of a non help-seeking control group means that no comparisons can be made with those with no psychiatric disorder. When the substance use analyses were stratified by age, the number of high risk users of alcohol and cannabis in the younger group were substantially smaller in sample size, leading to a similar effect size of association becoming statistically non-significant. Although substance use risk was confirmed by clinical interview, there is a general consensus that some forms of substance use, such as alcohol, may be under-reported, particularly in those under 18 years
. As such, the non-significant finding between alcohol and cannabis for those aged less than 18 years should be interpreted with some caution.