Conclusions and Directions for Future Research 333sleep patterns relative to drug therapy, but these changes are
fairly durable over time.
Combined Psychological and
Pharmacological Treatments
Only “ve studies have directly evaluated the combined or
differential effects of behavioral and drug treatment modal-
ities. Three of those studies compared triazolam to relaxation
(McClusky, Milby, Switzer, Williams, & Wooten, 1991;
Milby et al., 1993) or sleep hygiene (Hauri, 1997), one com-
pared estazolam with and without relaxation (Rosen, Lewin,
Goldberg, & Woolfolk, 2000), and the other one (Morin,
Colecchi, Stone, Sood, & Brink, 1999) compared cognitive-
behavior therapy to temazepam. Collectively, these studies in-
dicate that both treatment modalities are effective in the short
term. Drug therapy produces quicker and slightly better
results in the acute phase (“rst week) of treatment, whereas
behavioral and drug therapies are equally effective in the
short-term interval (4 to 8 weeks). Combined interventions
appear to have a slight advantage over a single treatment
modality during the initial course of treatment. Furthermore,
long-term effects have been fairly consistent for the single
treatment modalities but more equivocal for the combined ap-
proach. For instance, sleep improvements are well sustained
after behavioral treatment while those obtained with hypnotic
drugs are quickly lost after discontinuation of the medication.
Combined biobehavioral interventions may yield a slightly
better outcome during initial treatment, but long-term effects
are more equivocal. Studies with short-term follow-ups (
1
month) indicate that a combined intervention (i.e., triazolam
plus relaxation) produces more sustained bene“ts than drug
therapy alone (McClusky et al., 1991; Milby et al., 1993). The
only two investigations with follow-ups exceeding six months
in duration report more variable long-term outcomes among
patients receiving a combined intervention relative to those
treated with behavioral treatment alone (Hauri, 1997; Morin
et al., 1999). Some of these patients retained their initial im-
provements whereas others returned to their baseline values.
Combined biobehavioral treatments should theoretically
optimize outcome by capitalizing on the more immediate and
potent effects of drug therapy and the more sustained effects
of psychological interventions. In practice, however, the
limited evidence is not clear as to whether a combined inter-
vention has an additive or subtractive effect on long-term
outcome (Kendall & Lipman, 1991; Morin, 1996). In light of
the mediating role of psychological factors in chronic insom-
nia, behavioral and attitudinal changes may be essential to
sustain improvements in sleep patterns. When combining
behavioral and drug therapies, patients• attributions of the
initial bene“ts may be critical in determining long-term out-
comes. Attribution of therapeutic bene“ts to the drug alone,
without integration of self-management skills, may place a
person at greater risk for relapse once the drug is discontin-
ued. Also, the literature on state-dependent learning suggests
that self-management skills learned while taking hypnotics
may not generalize after drug discontinuation. Thus, it is not
entirely clear when, how, and for whom it is indicated to
combine behavioral and drug treatments for insomnia.CONCLUSIONS AND DIRECTIONS FOR
FUTURE RESEARCHSleep is a critical component of health and, as such, insomnia
can either be a cause or a consequence of health problems.
Signi“cant advances have been made in the past two decades
in the treatment of insomnia and in our understanding of the
relationships between sleep and psychological and physical
health. Despite these advances, a great deal more research is
still needed to address critical issues regarding the nature,
epidemiology, and treatment of insomnia.
There is a need for more basic studies of psychological
and biological factors that are presumed to contribute to the
etiology of insomnia. For example, the role of cognitive fac-
tors (e.g., intrusive thoughts, faulty beliefs), attention, and in-
formation processing variables needs further investigation to
re“ne and validate our current conceptual model of insomnia.
New assessment technologies (e.g., spectral analysis) should
also be used to gain a better understanding of the etiological
mechanisms and phenomenological experience underlying
insomnia complaints.
Because we know very little about the natural history of
insomnia, longitudinal studies are needed to document the
course, evolution, early precursors, and risk factors of the
disorder. Likewise, since only a small proportion of individu-
als with insomnia actually seek treatment, it is important to
examine help-seeking determinants among this population.
This longitudinal line of research should also evaluate the
long-term consequences of insomnia on psychological (e.g.,
depression) and physical health (e.g., immune function). The
direct and indirect costs associated with insomnia should also
be more fully documented.
Although signi“cant progress has been made in the man-
agement of insomnia, only a small proportion of treated indi-
viduals achieve complete remission. Additional clinical trials
are warranted to examine what parameters could optimize the
outcome of psychological therapies. Research is also needed
to evaluate the effects of single and combined behavioral and
pharmacological treatments for insomnia and to examine