Prospective longitudinal associations between persistent sleep problems in childhood and anxiety and depression disorders in adulthood.

Recent research suggests that childhood sleep problems may predict the development of subsequent internalizing problems in adolescence (Gregory & O'Connor, 2002). Despite growing support for this proposition, the longitudinal significance of childhood sleep problems for mental health remains unclear because studies to date have used relatively short follow-up periods. Furthermore, the possibility that sleep problems predict certain internalizing disorders and not others has not been fully investigated as existing studies tend to examine variation in general outcomes, rather than in specific psychiatric outcomes. This study seeks to clarify the longitudinal significance and specificity of persistent sleep problems in childhood on anxiety and depression diagnosed in adulthood.

Predictive Associations Between Sleep Problems and Internalizing Problems

Research examining predictive associations between sleep problems and internalizing problems has focused on adults. Such work has typically examined the association between insomnia and depression, and suggests that adults' sleep problems forecast depression (e.g. Breslau, Roth, Rosenthal, & Andreski, 1996; Chang, Ford, Mead, Cooper-Patrick, & Klag, 1997; Ford & Kamerow, 1989; Livingston, Blizard, & Mann, 1993; Weissman, Greenwald, Nino-Murcia, & Dement, 1997; for a review see Riemann & Voderholzer, 2003). Longitudinal research linking sleep problems and anxiety in adults also suggests a connection (e.g. Breslau et al., 1996; Ford & Kamerow, 1989). The biological mechanisms underlying these associations have yet to be fully elucidated.

Childhood risk indicators, such as parental loss, family conflict, and physical and sexual abuse, have been identified for the development of later internalizing problems (e.g. Birmaher et al., 1996; Fergusson. Horwood, & Lynskey, 1996). Less attention, however, has been paid to assessing the predictive associations between sleep problems and internalizing problems in children. One prospective longitudinal study of 490 children found that a composite measure of commonly reported sleep problems (e.g. nightmares, sleep-talking/walking, atypical sleep duration) at 4 years was associated with an increase in internalizing problems in mid-adolescence, after controlling for early symptoms of internalizing problems (Gregory & O'Connor, 2002). Other studies using different definitions of sleep problems and different age groups have found parallel associations (Gregory, Eley, O'Connor, & Plomin, 2004; Wong, Brower, Fitzgerald, & Zucker, 2004). Further studies examining clinically significant levels of internalizing problems have not found a longitudinal link (Johnson, Chilcoat, & Breslau, 2000; Stoleru, Nottelmann, Belmont, & Ronsaville, 1997). It is noteworthy that the prediction of later internalizing problems from sleep problems is more robust in adults than in children, suggesting that sleep problems may be a better predictor of internalizing problems in older than in younger participants. What no research has yet investigated is whether sleep problems in childhood predict adult outcomes. The hypothesis that sleep problems in childhood forecast internalizing psychopathology in adulthood is suggested by research demonstrating persistence and co-occurrence of sleep and internalizing problems over long time-frames (e.g. Gregory & O'Connor, 2002; Pine, Cohen, Gurley, Brook, & Ma, 1998).

Predictive Associations Between Sleep Problems and Internalizing Problems: Methodological and Conceptual Refinements

In contrast to the adult literature, which typically focuses on sleep problems predicting anxiety and depression disorders (e.g. Ford & Kamerow, 1989), much of the child literature focuses instead on individual differences incorporating both normal and pathological levels of disturbance. In particular, some of the strongest evidence to date for a link between sleep and later internalizing problems is based on nonclinical samples and variation within the normal range rather than on outcomes with obvious clinical relevance. Accordingly, in the current study we examine the extent to which early sleep problems in childhood predict clinically significant levels of disturbance in adulthood, namely, disorder.

A further problem in much of the previous research on children is the reliance on parent reports of both sleep problems and internalizing problems. This creates the possibility of rater bias artificially inflating the association between sleep and internalizing problems. The current study overcomes this limitation by directly assessing anxiety and depression according to standard clinical interview techniques.

Adult studies examining different types of internalizing problems have found sleep problems to forecast both anxiety and depression (e.g. Breslau et al., 1996; Ford & Kamerow, 1989). Child research tends not to examine different types of internalizing …