Cognitive-behavioral therapy for primary insomnia

Clinical Psychology Review 25 (2005) 539 – 558
Cognitive–behavioral therapy for primary insomnia
Jack D. EdingerT, Melanie K. Means
Department of Veterans Affairs and Duke University Medical Centers, Durham, NC, USA
Abstract
Primary insomnia (PI) is a prevalent form of sleep difficulty that impairs diurnal functioning, reduces quality of
life and enhances health care utilization/costs for millions worldwide. Whereas the underlying pathophysiology of
PI remains poorly understood, it is widely accepted that a host of cognitive and behavioral factors play important
roles in perpetuating this condition. As such, a multi-factorial, cognitive–behavioral therapy (CBT) has emerged as
a btreatment of choiceQ for managing the sleep/wake complaints of PI sufferers. This article considers the nature
and relative merits of CBT for treating PI patients. In addition, this article reviews studies supporting the general
efficacy and clinical effectiveness of CBT for treating PI complaints. Issues related to treatment implementation as
well as factors that mediate patients’ responses to CBT and predict treatment acceptance/outcome are also
considered. Finally, remaining questions regarding CBT’s application to PI are considered, and suggestions for
future research are provided.
Published by Elsevier Ltd.
Keywords: Cognitive–behavioral therapy; CBT; Primary insomnia
Primary Insomnia (PI) is a prevalent and potentially serious condition that adversely affects the
functioning, health status, and quality of lives of millions worldwide. Currently, both the nature and roles
of underlying neural mechanisms and hereditary factors involved in this form of sleep difficulty remain
poorly understood. However, it has long been recognized that a host of psychological and behavioral
factors play central roles in perpetuating this condition. As such, PI is particularly well suited for the
behavioral therapies that address these psychological and behavioral perpetuating mechanisms. Given
T Corresponding author. Psychology Service (116B), VA Medical Center, 508 Fulton Street, Durham, NC, USA. Fax: +1 919
416 5832.
E-mail address: jack.edinger@duke.edu (J.D. Edinger).
0272-7358/$ - see front matter. Published by Elsevier Ltd.
doi:10.1016/j.cpr.2005.04.003

---

540
J.D. Edinger, M.K. Means / Clinical Psychology Review 25 (2005) 539–558
this observation, it is no surprise that the bulk of the studies conducted to test and refine behavioral
insomnia therapies have been conducted with samples of PI patients.
The current article discusses the nature and significance of primary insomnia (PI) as well as the
behavioral therapies designed for its management. In the initial section of this article we review the
definition of PI and discuss its prevalence and associated morbidity. Subsequently, we consider PI’s
etiology and provide a conceptual rationale for use of behavioral treatments with PI by considering
those psychological and behavioral factors presumed to sustain this disorder. Following a brief
review of historically popular behavioral insomnia therapies, we describe current-day cognitive–
behavioral insomnia therapy (CBT) and consider both its general efficacy and its clinical
effectiveness with breal-worldQ patients. In addition, we discuss various issues related to treatment
implementation such as optimal treatment bdosingQ, cost-effectiveness, and methods of delivery, and
we consider those factors that mediate acceptance of and predict response to CBT. We conclude our
discussion by outlining the current limitations in the CBT insomnia literature and suggesting
directions for future research.
1. Primary insomnia: definition and epidemiology
PI is a diagnostic term specific to the American Psychiatric Association’s sleep disorder nosology
outlined in recent versions of its Diagnostic and Statistical Manuals. This diagnostic category first
appeared as a formal insomnia diagnosis in the revised, third edition of the Association’s Diagnostic
and Statistical Manual (American Psychiatric Association, 1987) and has been maintained through
subsequent revisions of this text (American Psychiatric Association, 1994, 2000). PI’s diagnostic
criteria listed in Table 1 highlight the primary or central role sleep/wake disturbance serves in
defining this condition. In fact, these criteria specify that a PI diagnosis is assigned when the
insomnia does not occur only during the course of another primary sleep or psychiatric disorder and
is not the direct result of a general medical disorder or substance use/abuse. As such, PI is perhaps
best conceptualized as a diagnosis established by exclusion of other primary and secondary forms of
sleep disturbance. Nonetheless, PI can usually be discerned from clinical interview, as expensive
and time-consuming laboratory tests are seldom needed for diagnosis of insomnia (Chesson et al.,
2000).
Table 1
Diagnostic criteria for primary insomnia
A. The predominant complaint is difficulty initiating or maintaining sleep or nonrestorative sleep for at least 1 month.
B. The sleep disturbance (or associated daytime fatigue) causes clinically significant distress or impairment in social,
occupational, or other important areas of functioning.
C. The sleep disturbance does not occur exclusively during the course of narcolepsy, breathing-related sleep disorder, circadian
rhythm sleep disorder, or a parasomnia.
D. The disturbance does not occur exclusively during the course of another mental disorder (e.g., major depressive disorder,
generalized anxiety disorder, delirium).
E. The disturbance is not due to the direct physiologic effects of a substance (e.g., a drug of abuse, a medication) or a general
medical condition.
Diagnostic criteria taken from the Diagnostic and Statistical Manual, Fourth Edition—Text Revision published by the American
Psychiatric Association.

---

J.D. Edinger, M.K. Means / Clinical Psychology Review 25 (2005) 539–558
541
Epidemiologic studies suggest that 9% to 15% of the general population suffer from some form of
chronic insomnia, whereas between 1% and 2% meet criteria for PI (Ohayon, 1997, 2002; Ohayon &
Partinen, 2002). Insomnia diagnoses in general are most common in middle-aged and older adults, but PI
is the most frequent insomnia diagnosis found in adolescents and young adults (Ohayon, 2002; Ohayon
& Roberts, 2001). As such, PI’s prevalence is relatively stable across age groups. Although many
insomnia sufferers go undetected (Ancoli-Israel & Roth, 1999), PI is common in primary care settings
and accounts for over 20% of all insomnia sufferers who present to specialty sleep disorders centers
(Coleman et al., 1982; Simon & VonKorff, 1997). Thus, PI appears sufficiently prevalent and disturbing
so as to frequently come to the attention of both sleep specialists and general medical practitioners.
Since PI is devoid of secondary causes, there is, perhaps, a temptation to view this problem as less
serious than those insomnias arising from medical, psychiatric, or substance abuse problems.
However, epidemiologic evidence suggests insomnia, uncomplicated by co-morbid psychiatric,
substance abuse, or medical disorders, substantially increases health care utilization/costs and
accounts for as many as 3.5 disability days per month among affected individuals (Hajak, 2001;
Simon & VonKorff, 1997; Weissman, Greenwald, Nino-Murcia, & Dement, 1997). Also, several
studies have shown that PI complaints reported during an initial interview predicted subsequent
depressive illnesses ascertained 1 to 45 years later even after other significant predictors were
statistically controlled (Breslau, Roth, Rosenthal, & Andreski, 1996; Chang, Ford, Mead, Cooper-
Patrick, & Klag, 1997; Ford & Kamerow, 1989; Livingston, Blizard, & Mann, 1993; Vollrath, Wicki,
& Angst, 1989). In addition, PI contributes to reduced productivity, work-related accidents, increased
alcohol consumption, serious falls among the elderly, and a general sense of being in poor health
(Brassington, King, & Bliwise, 2000; Gislason & Almqvist, 1987; Johnson, Roehrs, Roth, & Breslau,
1998; Johnson & Spinweber, 1983; Katz & McHorney, 1998). Thus, when encountered clinically, PI
warrants timely, effective, and enduring treatment.
2. Etiology
Currently, mechanisms underlying the pathophysiology of PI remain poorly understood.
Deficiencies in endogenous melatonin or benzodiazepine receptors and hyperactivity of corticotro-
phin releasing factor neurons have been cited as potentially important to the etiology of PI
(Richardson & Roth, 2001). However, the etiological importance of these factors has yet to be
empirically documented. Likewise, recent findings (Yves et al., 2003) showing a pronounced familial
propensity toward the development of PI imply that genetic or hereditary factors may play a role in
the etiology of this condition. To date, genetic aberrations specific to PI have not been identified, so
it is possible that familial trends toward this sleep difficulty are due to learning/modeling rather than
to genetic transmission.
Despite limited information concerning PI’s underlying pathophysiology and heritability, there is a
general agreement that this condition is perpetuated by the interplay of cognitive and behavioral
mechanisms as illustrated in Fig. 1. Setting the stage for sustained sleep difficulty is a host of
cognitive aberrations including misattributions about the causes of insomnia, misconceptions about
sleep needs and the effects of sleep loss, propensities toward catastrophizing about the consequences
of poor sleep, and dysfunctional beliefs about sleep promoting practices (Edinger et al., 2000; Morin,
1993; Morin, Stone, Trinkle, Mercer, & Remsberg, 1993). These cognitions, in turn, support and

---

542
J.D. Edinger, M.K. Means / Clinical Psychology Review 25 (2005) 539–558
Cognitive Factors
Dysfunctional Beliefs
Inhibitory Factors
Poor sleep hygiene
Homeostatic Dysregulation
Circadian Disruption
Conditioned arousal
Excessive TIB
Improper
Pre-bed & In-bed habits
Napping
Sleep Scheduling
Primary Insomnia
Fig. 1. Sleep-disruptive pre-bed habits include engaging in physically, psychologically, or emotionally arousing activities
shortly before bed. Sleep disruptive in-bed habits include watching TV, eating, reading, ruminating or spending long times
awake in bed at night.
sustain sleep-disruptive habits and conditioned emotional responses that either interfere with normal
sleep drive or timing mechanisms or serve as environmental/behavioral inhibitors to sleep. (Bootzin,
1977; Morin, 1993; Spielman, Saskin, & Thorpy, 1987; Webb, 1988). For example, daytime napping
or spending extra time in bed in pursuit of elusive, unpredictable sleep may only serve to interfere
with the body’s homeostatic mechanisms that operate automatically to increase sleep drive in the face
of increasing periods of wakefulness (i.e., sleep debt). Alternately, the habit of remaining in bed well
beyond the normal rising time following a poor night’s sleep may disrupt the body’s circadian or
bclockQ mechanisms that control the timing of sleep and wakefulness in the 24-h day. Additionally,
the repeated association of the bed and bedroom with unsuccessful sleep attempts may eventually
result in sleep-disruptive conditioned arousal in the home sleeping environment. Finally, failure to
discontinue mentally demanding work and allot sufficient bwind-downQ time before bed may serve as
a significant sleep inhibitor during the subsequent sleep period. In sum, all of these factors may
contribute to and perpetuate PI (Bootzin & Epstein, 2000; Edinger & Wohlgemuth, 1999; Hauri,
2000; Morin, Savard & Blais, 2000).
3. The behavioral therapies for primary insomnia
Given the important roles that the above-mentioned cognitive and behavioral factors play in
perpetuating PI, a variety of behavior therapies have been proposed for this condition. A detailed
overview of the most widely used therapies for PI is provided in Table 2. As shown in this table, a
host of formal relaxation therapies have been applied to PI since such therapies reduce the sleep-
related performance anxiety and bedtime arousal common to this condition (Borkovec & Fowles,

---

J.D. Edinger, M.K. Means / Clinical Psychology Review 25 (2005) 539–558
543
Table 2
Common behavioral therapies for primary insomnia
Type of treatment
Treatment description
Relaxation therapy n
Various techniques used to treat PI including progressive muscle relaxation, passive relaxation,
autogenic training, biofeedback, imagery training, meditation, and hypnosis.
n
The therapy goal is to reduce or eliminate sleep-disruptive physiological (e.g., muscle tension) and/or
cognitive (e.g., racing thoughts) arousal.
n
Regardless of the specific relaxation strategy employed, treatment typically entails conducting
specific treatment exercises and teaching relaxation skills over multiple treatment sessions.
Stimulus control
n
This approach is based on the assumption that both the timing (bedtime) and sleep setting (bed/
therapy
bedroom) are associated with repeated unsuccessful sleep attempts and, over time, become
conditioned cues for arousal that perpetuate insomnia. As a result, the goal of this treatment is that
of re-associating the bed and bedroom with successful sleep attempts.
n
In practice, this therapy requires instructing the patient to: (a) go to bed only when sleepy; (b)
establish a standard wake-up time; (c) get out of bed whenever awake for long periods; (d) avoid
reading, watching TV, eating, worrying and other sleep-incompatible behaviors in the bed/bedroom;
and (e) refrain from daytime napping.
n
Stimulus control instructions usually can be administered in one visit, but follow-up visits to facilitate
compliance are beneficial.
Sleep restriction
n
Sleep restriction therapy reduces nocturnal sleep disturbance primarily by restricting the time allotted
therapy
for sleep each night so that, eventually, the time spent in bed closely matches the individual’s
presumed sleep requirement.
n
This treatment typically begins by calculating the individual’s average total sleep time (ATST) from a
sleep log that is kept for 1 to 2 weeks.
n
An initial time-in-bed (TIB) prescription may either be set at the ATST or at a value equal to the ATST
plus an amount of time that is deemed to represent normal nocturnal wakefulness (e.g., ATST + 30
min). The initial TIB prescription is seldom set below 5 h per night.
n
On subsequent visits TIB may be adjusted up or down in 15 to 30 min increments dependent upon the
patient’s sleep performance and waking function.
n
Therapy typically entails an initial visit to introduce treatment instructions and follow-up visits to alter
TIB prescriptions.
Sleep hygiene
n
Patients are educated about healthy sleep behaviors and sleep-conducive environmental conditions.
Typically they are encouraged to exercise daily, eliminate the use of caffeine, alcohol, and nicotine,
eat a light snack at bedtime, and ensure that the sleeping environment is quiet, dark, and comfortable.
n
Sleep hygiene is seldom used as a primary intervention, but is often included with other interventions.
1973; Jacobson, 1964; Nicassio & Bootzin, 1974; Schultz & Luthe, 1959). Stimulus control therapy
(SCT: Bootzin, 1972, 1977), a simple, straightforward classical conditioning approach, also has been
popular since this treatment is well suited to address the conditioned timing (bedtime) and setting
(home bedroom) cues for arousal that may perpetuate sleep difficulty. In contrast, more global
lifestyle and environmental factors that serve as sleep inhibitors are targeted in sleep hygiene therapy
(Hauri, 1982). Finally, sleep restriction therapy (SRT) designed primarily to restore normal
[homeostatic] sleep drive has enjoyed wide popularity as well.
Over the years, each of these first generation behavioral therapies has shown some promise for
treating PI, although SCT and SRT have proven to be the most efficacious (Morin, Culbert, &
Schwartz, 1994; Morin, Hauri et al., 1999; Murtagh & Greenwood, 1995). Nonetheless, as
illustrated in Fig. 2, none of these interventions addresses all PI perpetuating mechanisms.
Relaxation approaches target conditioned arousal and sleep-related anxiety but largely ignore

---

544
J.D. Edinger, M.K. Means / Clinical Psychology Review 25 (2005) 539–558
Fig. 2. Sleep-disruptive pre-bed habits include engaging in physically, psychologically, or emotionally arousing activities
shortly before bed. Sleep disruptive in-bed habits include watching TV, eating, reading, ruminating or spending long times
awake in bed at night. Note that arrows drawn from the treatment bboxesQ with solid lines indicate a primary effect whereas
dotted lines indicate a possible incidental or secondary effect of the treatment.
cognitive, homeostatic, and circadian aberrations that sustain sleep difficulty. Sleep hygiene directly
addresses some inhibitory mechanisms and indirectly may alter some dysfunctional sleep-related
beliefs through the sleep education it typically includes. SCT directly addresses both circadian and
inhibitory mechanisms (e.g., conditioned arousal) involved in PI but has less direct effects on
homeostatic sleep drive since it places no limits on time spent in bed. Furthermore, this therapy
does not include specific interventions for cognitive factors that sustain many sleep-disruptive
attitudes and behaviors. Finally, SRT addresses the homeostatic and circadian mechanisms that
sustain PI but ignores many inhibitory factors and maladaptive beliefs and attitudes about sleep.
Due to the limitations of these first generation approaches, a more omnibus hybrid approach
currently called cognitive–behavioral therapy (CBT) has become popular for the management of PI
(Edinger, Hoelscher, Marsh, Lipper, & Ionescue-Pioggia, 1992; Hoelscher & Edinger, 1988; Morin,
1993; Morin, Kowatch, Barry, & Walton, 1993). Although various renditions of this approach have
been described, included in all of these are treatment components that address the array of cognitive,
homeostatic, circadian, and sleep-inhibitory factors that sustain PI. Common to all forms of CBT is a
form of cognitive therapy such as formal cognitive restructuring (Morin, 1993) or a standardized
sleep education package (Edinger et al., 1992; Hoelscher & Edinger, 1988) to correct dysfunctional
beliefs and attitudes about sleep common to PI. Moreover, in order to address the host of
homeostatic, circadian, and conditioning factors that may perpetuate PI, CBT combines SCT and
SRT resulting in a modified set of behavioral prescriptions. With the modified regimen, patients are
instructed to: (a) adhere to a standard rising time; (b) avoid extended periods in bed awake; (c)
eliminate sleep-incompatible behaviors in the bed and bedroom; (d) avoid daytime napping; and (e)
limit total time in bed (TIB) each night to a specific, individually prescribed amount. The initial TIB
prescription is set to approximate the patient’s average sleep time (estimated from a pre-treatment

---

J.D. Edinger, M.K. Means / Clinical Psychology Review 25 (2005) 539–558
545
sleep log). TIB subsequently is adjusted up or down (in 15- to 30-min increments) on a weekly basis
contingent upon observed sleep improvements or decrements. When sleep becomes well consolidated
and waking function appears improved or at least unimpaired, no further TIB adjustments are
effected. Thus, CBT provides a comprehensive treatment for PI by eliminating common sleep-
disruptive behaviors and correcting the beliefs/attitudes that support such practices.
4. Efficacy of CBT for treating PI
The usefulness of a treatment for a particular disorder is typically first established by efficacy studies
in which the treatment is tested with well-characterized and carefully selected patient samples using a
rigorous experimental design. The initial studies designed to test CBT for treating PI utilized single
subject multiple baseline designs to test the efficacy of CBT with middle-aged and older patients who
presented sleep-maintenance difficulties (Edinger et al., 1992; Hoelscher & Edinger, 1988). In the earlier
of these studies, four PI sufferers were given 4 weeks of CBT after varying baseline periods. In the later
study, seven PI patients completed varying baseline periods and 4 weeks of progressive relaxation
treatment prior to completing 4 weeks of CBT. In both studies the majority of patients showed clinically
significant subjective and objective improvements in common measures of sleep consolidation (e.g.,
sleep efficiency, nocturnal wake time) only after the introduction of CBT. Furthermore, these patients
maintained these improvements through 3- to 6-month follow-up periods.
Since the time of these early reports, four well controlled randomized clinical trials have been
conducted specifically to test CBT’s efficacy with PI patients. Table 3 provides a listing and brief
description of these studies. The findings from these studies collectively suggest that CBT produces
significantly greater subjective and objective sleep improvements than do no treatment, pharmacologic
and non-pharmacologic placebo interventions, and progressive relaxation therapy. These studies also
demonstrated that sleep improvements resulting from CBT tend to endure across post-treatment follow-
up periods varying from 3 to 24 months. In addition, findings from one of these studies (Morin,
Colecchi, Stone, Sood, & Brink, 1999) suggest that combined CBT and pharmacotherapy (temazepam)
may produce slightly greater short-term sleep improvements than does CBT alone. However, the
advantages of this combined approach over the long-term seems less clear since many patients receiving
this treatment regress significantly over long-term follow-up. In fact, on average, those receiving CBT
alone showed better maintenance of their sleep improvements at the end of a 2-year follow-up than did
those receiving combined CBT and pharmacotherapy.
It also should be noted that findings from these and related studies show CBT leads to
improvement in clinically important sleep-related metrics in addition to subjective/objective measures
of sleep. Compared to PI patients receiving relaxation therapy, placebo treatment, or no treatment,
CBT-treated PI patients show larger improvements on questionnaire measures of their sleep/wake
insomnia symptoms, personal self-efficacy in regard to sleep, and dysfunctional sleep-related
cognitions (Edinger, Wohlgemuth, Radtke, Marsh, & Quillian, 2001a; Edinger, Wohlgemuth, Radtke,
Marsh, & Quillian, 2001b; Morin, Blais & Savard, 2002; Morin, Colecchi et al., 1999). Likewise,
when improvement rates are used as a metric, CBT appears to fair well against other treatments.
Edinger et al. (2001a), for example, found that a significantly greater percentage of their CBT-treated
patients with sleep-maintenance PI met predetermined improvement criteria (i.e., a 50% pre- to post-
treatment reduction in wake time after sleep onset) than did patients treated with either relaxation

---

546
J.D. Edinger, M.K. Means / Clinical Psychology Review 25 (2005) 539–558
Table 3
CBT treatment efficacy studies conducted with primary insomnia patients
Citation
N
Mean
Treatment
Treatment Treatment
Follow-up Findings
age
groups
duration
format
period
(years)
(weeks)
(months)
Morin et al.,
24 67.1
CBT
8
Group
12
CBT significantly better than WL
1993
WL
across most measures.
For CBT, percentage reductions in
WASO were 53.7% (diaries) and 51.3%
(PSG). For WL, percentage reductions
in WASO were 12.3% (diaries) and
10.4% (PSG).
Improvements from CBT persisted over
12-month follow-up.
Morin et al.,
78 65.0
CBT
8
Group
24
3 active treatments improved sleep better
1999
PCT
than placebo drug by post-treatment.
CBT + PCT
Combined therapy slightly better than
Placebo drug
single treatments in the short-term.
Percentage reductions in WASO were
63.5% for CBT + PCT, 55% for CBT,
46.5% for PCT, and 16.9% for placebo.
As a group, those treated with CBT alone
maintained improvements across 2-year
follow-up better than those receiving
combined therapy although considerable
variability was noted in the latter group.
Mimeault &
54 50.8
SHCBT
6
Self-help
3
Both active treatments improved SE%
Morin,
SHCBT + P
manual F Phone
and TWT whereas WL did not.
1999
WL
support from
Phone assistance provided slight
therapist
short-term benefit but this advantage did
not persist through follow-up. By
follow-up the SHCBT group showed
36% and 19.1% improvements in TWT
and SE%, respectively. SHCBT + P
showed 32.5% and 13.4%
improvements in TWT and SE%.
Edinger et al., 75 55.8
CBT
6
Individual
6
CBT produced a 54% reduction in
2001a,
PMR
therapy
subjective WASO compared to 16%
2001b
Placebo
for PMR and 12% for placebo.
PSG differences smaller but supported
CBT over other treatments.
CBT significantly better than other
treatments for reducing subjective
insomnia symptoms and dysfunctional
beliefs about sleep.
Note: CBT = cognitive–behavioral therapy; PMR = progressive muscle relaxation therapy; WL = waiting list condition;
PSG = polysomnography; PCT = pharmacotherapy; SHCBT = self-help CBT treatment; P = phone therapy; SE% = sleep
efficiency; TWT = total wake time; WASO = wake time after sleep onset.

---

J.D. Edinger, M.K. Means / Clinical Psychology Review 25 (2005) 539–558
547
therapy or a non-drug placebo therapy. Thus, the efficacy studies show that CBT produces marked
changes across a range of clinically important treatment outcome measures.
5. Clinical effectiveness of CBT with PI
The various well-designed efficacy studies reviewed above provide strong support for CBT’s
usefulness with PI. However, one shortcoming of these studies is their use of highly screened and
selected research volunteers that may not represent the typical PI patient encountered in clinical practice.
As a result, it is questionable whether the very promising findings of these studies are likely to generalize
to the population of PI sufferers as a whole. To make this determination, additional clinical effectiveness
studies conducted with less carefully screened breal-worldQ patients are necessary.
Fortunately, a number of studies have evaluated the clinical effectiveness of CBT with clinic samples
of PI patients. Table 4 provides an overview of such studies. Included in this table are only those studies
that evaluated CBT’s effectiveness specifically and exclusively with PI sufferers. Excluded from the
table are effectiveness studies conducted with multicomponent behavioral treatments other than CBT as
well as studies of mixed insomnia subtypes that lacked separate evaluation of CBT’s effectiveness within
the PI subgroup. The studies listed all included cohorts composed largely of unrecruited and/or physician
referred patients treated in primary or tertiary treatment settings.
Considered collectively, these studies confirm impressions derived from the efficacy studies and
suggest that CBT is highly effective with breal-worldQ patients across a range of important outcome
measures. Specifically, these studies confirm that CBT results in important improvements in subjective
measures of sleep including sleep onset latency, wake time after sleep onset, sleep efficiency, and total
sleep time. A subset of these studies (Edinger & Sampson, 2003; Espie, Inglis, Tessier, & Harvey, 2001)
show that clinic patients treated with CBT manifest larger improvement in these sleep measures than do
either untreated patients or patients treated with an alternate behavioral approach. The cumulative
findings of the studies listed also suggest that CBT enhances sleep-related self-efficacy, reduces
depressive and anxiety symptoms, corrects dysfunctional beliefs about sleep, reduces use of sleep
medications, and leads to improvements in insomnia-related sleep/wake symptoms. Given such findings,
CBT can be considered a highly effective treatment for the sleep and related complaints of the PI patients
encountered in clinical practice.
6. Treatment implementation
Although current data support both the efficacy and effectiveness of CBT for PI, a number of factors
impact the successful implementation of this treatment. The following sections discuss these factors,
which include methods of delivering CBT, treatment dosage, cost-effectiveness, and accessibility.
6.1. Treatment delivery
CBT for insomnia was developed to be provided to patients in individual sessions (Morin, 1993). In
efforts to improve cost-effectiveness and increase accessibility, however, a number of alternative
delivery methods have been investigated. Studies employing a group CBT format of 6 to 8 sessions (4 to

---

548
Table 4
CBT treatment effectiveness studies conducted with primary insomnia patients
Citation
N
Mean
Study design/
Treatment Treatment Follow-up Findings
age
Patient types
duration
format
period
(years)
(weeks)
(months)
Morin, Stone,
31
42.7
Case series/PI
14
Individual 24
On average, PI patients showed N 50% reductions in SOL and
McDonald &
patients from
WASO by treatment end whereas SE% increased from 70.2% to
J.D.
Jones, 1994
sleep center;
84%.
Edinger
67.7% used
Patients significantly reduced medication use from the beginning to
sleep aid at
the end of treatment.
study entry.
Follow-up showed treatment gains maintained.
,
M.K.
Perlis et al.,
47
39.2
Case series/PI
4 to 9
Individual N/A
61% completed at least 4 sessions and showed global improvement
2000
patients from sleep
in symptoms.
Means
center;
Patients showed a 65% reduction in SOL, a 46% reduction in
53% comorbid
number of nocturnal awakenings, a 48% reduction in WASO, and a
/
Psychiatric or
13% increase in TST.
Clinical
medical disorders.
Backhaus et al., 20
43.0
Single group/PI
6
Group
36
Measures of SOL, TST, and SE% all improved.
2001
patients from
By the 3-year follow-up SOL, TST, and SE% were improved by
Psychology
sleep clinic;
26%, 28%, and 33%, respectively.
60% on hypnotics
CBT reduced dysfunctional cognitions, depression, and anxiety.
Espie et al.,
139 51.4
CBT vs. WL/PI
6
Group
12
CBT produced significant reductions in SOL and WASO but WL
2001
patients from medical
did not. Within the CBT group, 64.2% has SOL values and 63.3%
Review
clinics; N50% took
has WASO values b30 min by 12-month follow-up.
sleep medications.
Significantly increased TST shown at follow-up for CBT group.
25
84% of hypnotic users were drug free at follow-up.
Edinger &
20
51.0
CBT vs. SH/PI
2
Individual 3
Abbreviated 2-session CBT superior to SH for improving
(2005)
Sampson,
patients in
subjective sleep measures and measures of sleep-related
2003
primary care at
self-efficacy, insomnia symptoms, and dysfunctional beliefs about
539–558
VA hospital; 65%
sleep.
had comorbid medical
52% of those receiving CBT reported at least a 50% reduction in
or psychiatric
their WASO.
diagnoses.
55.6% of CBT-treated patients who entered the study with
pathological scores on an Insomnia Symptom Questionnaire (ISQ),
achieved normal ISQ scores by their final outcome assessment.
Note: CBT = cognitive–behavioral therapy; SH = sleep hygiene therapy; WL= waiting list condition; SE% = sleep efficiency; SOL = sleep onset latency;
WASO = wake time after sleep onset; TST = total sleep time. The Morin et al. (1994) study included 100 mixed insomnia patients but only findings for the 31
primary insomnia (PI) patients are considered here.

---

Document Outline

• Cognitive-behavioral therapy for primary insomnia
  • Primary insomnia: definition and epidemiology
  • Etiology
  • The behavioral therapies for primary insomnia
  • Efficacy of CBT for treating PI
  • Clinical effectiveness of CBT with PI
  • Treatment implementation
    • Treatment delivery
    • Treatment dosage
    • Cost-effectiveness
    • Treatment accessibility
  • Treatment acceptability and adherence
    • Acceptance of CBT
    • Adherence to CBT
  • Predictors and mediators of treatment outcome
    • Demographic variables
    • Severity of insomnia
    • Medical and psychological factors
    • Cognitive mediators
  • Conclusions
  • References

---