332 Insomnia
(Hauri, 1991). Although these factors are rarely of suf“cient
severity to be the primary cause of insomnia, they may
potentiate sleep dif“culties caused by other factors. Sleep hy-
giene is typically incorporated with other interventions to
minimize interference from poor sleep hygiene practices.
Basic recommendations involve avoidance of stimulants
(e.g., caffeine, nicotine) and alcohol, exercising regularly,
and minimizing noise, light, and excessive temperature. It
may also include advice about maintaining a regular sleep
schedule and avoiding napping, although these instructions
are part of the standard stimulus control therapy.
Additional nondrug interventions are available for treat-
ing insomnia including paradoxical intention, hypnosis,
acupuncture, ocular relaxation, and electro-sleep therapy.
Those methods have not yet received adequate empirical val-
idation in controlled studies. Psychotherapy may also be use-
ful to address predisposing factors to insomnia, but there has
been no controlled evaluation of its ef“cacy.
Summary of Outcome Evidence
Evidence for Efficacy
Two meta-analyses recently summarized the “ndings of more
than 50 clinical studies (involving over 2,000 patients) of non-
pharmacological interventions for insomnia (Morin, Culbert,
& Schwartz, 1994; Murtagh & Greenwood, 1995). The data
indicate that behavioral treatment (lasting an average of 4 to
6 weeks) produces reliable changes in several sleep parame-
ters of individuals with primary insomnia. Almost identical
effect sizes, 0.87 and 0.88, have been reported in both meta-
analyses for sleep-onset latency, the main target symptom in
studies of sleep-onset insomnia. An effect size of this magni-
tude indicates that, on average, insomnia patients are better off
(fall asleep faster) after treatment than about 80% of untreated
control subjects. Reliable effect sizes, falling in what is con-
ventionally de“ned as moderate to lar ge, have also been re-
ported for other sleep parameters, including total sleep time
(0.42...0.49), number of awakenings (0.53...0.63), duration of
awakenings (0.65), and sleep quality ratings (0.94). These
effect sizes are comparable to those reported with benzodi-
azepines and zolpidem (Nowell et al., 1997). In terms of
absolute changes, sleep-onset latency is reduced from an av-
erage of 60 to 65 minutes at baseline to about 35 minutes at
posttreatment. The duration of awakenings is similarly de-
creased from an average of 70 minutes at baseline to about
38 minutes following treatment. Total sleep time is increased
by a modest 30 minutes, from 6 hours to 6.5 hours after treat-
ment, but perceived sleep quality is signi“cantly enhanced
with treatment. Overall, the magnitude of these changes indi-
cate that between 70% to 80% of treated patients bene“t from
treatment. These results represent conservative estimates of
ef“cacy because they are based on average ef fect sizes com-
puted across all treatment modalities.
Comparative studies of different psychological treatments
have generally, but not always, shown stimulus control
therapy and sleep restriction to be the most effective single
treatment modalities. As psychological interventions are not
incompatible with each other, they can be effectively com-
bined. Multifaceted interventions that incorporate behav-
ioral, educational, and cognitive components often produce
the best outcome.Durability and Generalizability of ChangesCognitive-behavior therapy for insomnia produces stable
changes over time. Improvements of sleep parameters and
satisfaction with those changes are well maintained up to
24 months after treatment. While increases in total sleep time
are fairly modest during the initial treatment period, these
gains are typically enhanced at follow-up, with total sleep
time often exceeding 6.5 hours. Although promising, these
data must be interpreted cautiously because less than 50% of
studies report long-term follow-up and, among those that do,
attrition rates increase substantially over time.
The large majority of behavioral and pharmacological
treatment studies have focused on primary insomnia in other-
wise healthy and medication-free patients. Thus, an impor-
tant question is whether the “ndings obtained in these
research studies generalize to patients typically seen in clini-
cal practice, patients who often present with comorbid
medical and psychiatric disorders. Preliminary “ndings from
uncontrolled clinical case series (Chambers & Alexander,
1992; Dashevsky & Kramer, 1998; Jacobs, Benson, &
Friedman, 1996; Morin, Stone et al., 1994) have yielded
promising results suggesting that patients with medical and
psychiatric conditions, or even those using hypnotic medica-
tions can bene“t from behavioral treatment for sleep distur-
bances. Because these studies have a more naturalistic
focus and are not as rigorously controlled as randomized con-
trolled trials, these conclusions are only tentative at this time
(Currie, Wilson, & Pontefract, 2000).
In summary, behavioral treatment produces reliable and
durable sleep improvements in primary insomnia. The major-
ity (70% to 80%) of treated patients bene“t from treatment,
but only a minority become good sleepers and a small pro-
portion of patients do not respond at all to treatment. Behav-
ioral treatment often leads to a greater sense of personal
control over sleep and reduces the need for hypnotic medica-
tions. Behavioral interventions require more time to improve