Clinical efficacy of individual cognitive behavior therapy for psychophysiological insomnia in 20 outpatients

Authors


*Miki Sato, MD, Department of Psychiatry, Jikei University School of Medicine, 105-8461, 3-25-8 Nishi-Shinbashi, Minato-Ku, Tokyo, Japan. Email: satomiki@jikei.ac.jp

Abstract

Aim:  Twenty patients (14 of them women) suffering from psychophysiological insomnia (PPI) were enrolled for cognitive behavior therapy (CBT). The mean age of the patients was 56.9 years, and the mean duration of insomnia morbidity was 8.9 years. Each received individual combined CBT treatments consisting of stimulus control, sleep reduction, cognitive therapy and sleep hygiene education over a period of 1 month.

Methods:  Just before the CBT and after its completion, sleep measurements were conducted that involved (i) sleep logs, Dysfunctional Beliefs and Attitudes about Sleep Scale (DBAS), and the Pittsburgh Sleep Quality Index (PSQI); (ii) actigraphy measurement; (iii) dissociation between subjective and objective evaluation of sleep calculated from sleep logs and actigraphy results; and (iv) correlation between DBAS and the aforementioned sleep parameters. Because the intention was to focus on patients' incorrect cognition about sleep, the definition ‘changes in dissociation between the sleep log and actigraphically measured sleep’ was used as the primary outcome and ‘changes in DBAS score’ as the secondary outcome.

Results:  After the CBT the following was found: (i) underestimation by PPI patients of the objective evaluation of sleep; (ii) a decrease in the dissociation between the subjective and objective evaluation of sleep; (iii) improvement of the DBAS; and (iv) improvement of sleep logs and actigraphy measurements. Moreover, there was a correlation between the improvement of PSQI, sleep logs and DBAS.

Conclusion:  CBT for insomnia is able to redress incorrect cognition about sleep, leading to improvement of the disorder.

THE DISEASE CONCEPT of psychophysiological insomnia (PPI), as defined in the International Classification of Sleep Disorders Second Edition,1 (ICSD-II) is almost the same as primary insomnia as defined in DSM-IV-TR.2 PPI is characterized by morbid fear of insomnia, mental arousal, and heightened somatic tension in bed. It is well known that cognitive behavior therapy (CBT) for insomniacs is effective for PPI patients.3–6 Details of the mechanism of action of CBT for the disorder, however, have not been well elucidated.

There is reportedly an underestimation of objective evaluation of sleep and a dissociation between subjective and objective evaluation of sleep in PPI,7,8 and it is suspected that dissociation, that is, sleep state misperception, might play a pathologic role in the mechanism of the disorder.9 Considering this, we compared subjective with objective sleep evaluation, and also conducted CBT on the consecutive outpatients with PPI, to investigate its clinical efficacy, focusing on the changes in dissociation, as an indicator of the improvement in patients' cognition about their sleep. On the basis of the results obtained, we then considered the relationship between the pathological mechanism of PPI and sleep state misperception.

METHODS

The study protocol and therapy regimen were approved by the Jikei University School of Medicine Ethics Committee (Acceptance no: 15-97(4122)), and the study was started after obtaining full informed consent from each subject.

Study participants

The eligible subjects were consecutive PPI patients attending Jikei University Hospital as outpatients between 2004 and 2005, wishing to receive CBT for insomnia. They were considered for inclusion if they met the ICSD-II criteria for PPI,1 but were excluded if they (i) met the DSM-IV-TR criteria for an axis I psychiatric disorder and/or substance abuse; (ii) required psychotropic medication for psychiatric symptoms; (iii) had possible sleep apnea syndrome or narcolepsy as judged from clinical interviews and daytime polysomnography (PSG);10 or (iv) had symptoms suggestive of restless legs syndrome.

During the treatment, patients continued to take any medication already prescribed before enrollment, so as to avoid any effects of medication withdrawal.

A total of 24 patients gave written informed consent to take part in the study. During CBT, however, four patients dropped out at their own request or at the recommendation of their attending physicians. Data for these patients were excluded from the statistical analysis. Finally, 20 PPI patients (14 female) were included. Their mean age was 56.9 ± 12.6 years (range, 27–76 years), and mean duration of insomnia morbidity, that is, the subjectively reported period from initial appearance of insomnia to the time of receiving CBT, was 8.9 ± 6.2 years (range, 1–22 years; Table 1).

Table 1.  Subject characteristics (n = 20)
Age (years)SexDuration of insomnia morbidity (years)Hypnotics jbefore CBT equivalent to FNP 1 mg = 1Contents of incorrect beliefs and attitudes about sleep
  1. FNP, flunitrazepam.

67F51I feel insomnia is ruining my ability to enjoy life and prevents me from doing what I want.
57F173Without an adequate night's sleep, I can hardly function the next day.
71F102I am worried that if I go for a few nights without sleep, I may have dementia.
62M41Inadequate sleep hygiene
59M153I need over 7 h of sleep to feel refreshed and function well during the day.
60M12By spending more time in bed, I usually get more sleep and feel better the next day.
33M51Without an adequate night's sleep, I can hardly function the next day.
62F90.5I need 8 h of sleep to feel refreshed and function well during the day.
67F122I am worried that I may lose control over my abilities to sleep.
54F10When I feel depressed, or anxious during the day, it is mostly because of poor sleep.
27M54I am worried that I may lose control over my abilities to sleep.
56F94Because my husband falls asleep soon and stays asleep through the night, I should be able to do this.
62F220I am worried that if I go for a few nights without sleep, I may have dementia.
70F191I believe insomnia is running my ability to enjoy life and prevents me from doing what I want.
45F171I can't ever predict whether I'll have a good or poor night's sleep.
64F71I am worried that if I go for a few nights without sleep, I may have dementia.
48M82.5Without an adequate night's sleep, I can hardly function the next day.
76F50.5Because my husband falls asleep soon and stays asleep through the night, I should be able to do this.
41F30Inadequate sleep hygiene
57F32I believe insomnia is ruining my ability to enjoy life and prevents me from doing what I want.

The past use average daily dosage of hypnotics was 1.58 mg (flunitrazepam 1 mg being equivalent for 1; Table 1),11 and the pharmacological therapy that subjects received before CBT was not sufficiently effective for improving the symptoms of insomnia. We were able to observe various forms of incorrect cognition about sleep in each subject before CBT (Table 1).

Cognitive behavior therapy

Because none of the authors had received any training in CBT, we designed our own therapy regimen with reference to the method described by Morin et al.5 and Edinger et al.3 The first author gave instruction to each subject, and the other authors supervised the content of the CBT. The final format for CBT was then chosen, and every subject received the same treatment content. We defined the 7 days prior to the initial CBT session as the pre-CBT period, and the following 4 weeks as the treatment period. The post-CBT period was also defined as the first 7 days after the treatment period (5–6 weeks from the initial CBT session).

In each session we identified incorrect cognition and the attitude of the patients that had been responsible for causing their chronic insomnia (Table 1), and to also promote the patients' understanding regarding their fear of insomnia. The CBT consisted of the following.

Stimulus control therapy

Stimulus control attempts to break the association between the sleep environment and wakefulness by teaching the patients not to be engaged in activities that might disturb their sleep.4,12,13 The instructions we gave were as follows: (i) go to bed only when becoming sleepy; (ii) do not use the bedroom for anything except sleep or sex; (iii) get out of bed and go to another room whenever unable to fall asleep over a period of 30 min, and return to bed only when becoming sleepy again.

Sleep reduction therapy

This treatment seeks to increase homeostatic sleep drive through partial sleep deprivation and thereby improve sleep ability.4,14 A bedtime and rising time schedule was prescribed in an attempt to improve sleep quality and decrease time spent awake during the night. Time in bed was reduced in accordance with the total sleep time, as recorded in the sleep log, and rising time was always fixed. The time the patient went to bed was adjusted on the basis of sleep efficiency.

Cognitive therapy

As mentioned in the following section, we calculated the dissociation between patients' subjective sleep evaluation from their sleep logs and their objective sleep data measured on actigraphy during the pre-CBT period. To facilitate better understanding by the patients, we showed them the results of dissociation between the two parameters as an indicator of sleep state misperception. Subsequently, cognitive therapy12 was carried out to identify patient-specific incorrect cognition about sleep so that we could correct any dysfunctional in this regard.

Sleep hygiene education

Sleep hygiene education included instruction about health practices and environmental factors that can be beneficial for maintaining sufficient sleep, and also details regarding homeostatic drive for sleep, circadian factors, and the effects of drugs and habits prior to sleep.15

Cognitive behavioral treatment programs

In the first 60–90-min sessions, after the introduction of combined CBT, the therapy was started for each individual patient by the first author. Thereafter, patients underwent sessions of 15 min once every 2 weeks during a 1-month period. A total of three sessions was given to each patient in the present study.

During the treatment period we redressed the patients' bedtimes, rising times, incorrect habits before sleep, incorrect understanding of sleep, and any underestimation of their sleep quality.

Measurements

During the pre- and post-CBT periods we conducted measurements including the Pittsburgh Sleep Quality Index (PSQI),16,17 the Dysfunctional Beliefs and Attitudes about Sleep Scale (DBAS),18,19 sleep log and actigraphic recording (mini motionlogger actigraph; Ambulatory Monitoring, Ardsley, NY, USA). The reliability and validity of these instruments have been verified.

Pittsburgh Sleep Quality Index

We used the Japanese version of the PSQI.17 With this index, we investigated subjective sleep quality, sleep onset latency (SOL), total sleep time (TST), sleep efficiency (SE) and difficulty in falling asleep.

Dysfunctional Beliefs and Attitudes about Sleep Scale

Sleep-related cognition plays an important role in perpetuating insomnia,20 and it is reported that reduction of DBAS is correlated with improvement of sleep parameters.21 Therefore in the present study we used the Japanese version of the DBAS19 to measure patients' faulty cognition about sleep and to evaluate the clinical efficacy of CBT.

The DBAS is a self-recorded questionnaire developed by Morin et al. It consists of a 28-item scale that extracts various beliefs and attitudes about sleep, focusing on the following five themes: (i) consequences of insomnia; (ii) control and predictability of sleep; (iii) sleep requirement expectations; (iv) causal attributes of insomnia; and (v) sleep-promoting practices. A higher score indicates a more dysfunctional belief.

The average scores of 19 patients (one was unable to complete the DBAS) for the five themes were compared between the pre- and post-CBT situations.

Sleep logs

During the pre- and post-CBT periods, patients were asked to complete sleep logs, just after getting up in the morning, for 7 days. We averaged bedtime, rising time, sleep-onset time (SONT), sleep-offset time (SOFT), SOL, TST and total time in bed (TIB). In principle, bedtime and rising time on the sleep logs were recorded by each patient's family members to improve the objectivity of the data.

Actigraphy

During the pre- and post-CBT measurement periods, patients were required to wear an actigraph on their non-dominant wrist at all times for 7 days. Based on these rest/activity data, estimation of their sleep was made using the algorithm devised by Cole et al.,22 which has a >90% agreement with nocturnal PSG.23 From this result, we obtained the 7-day averaged data for objective SONT, SOFT, SOL, and the number of awakening episodes lasting >5 min (NOA), awakening time after sleep onset (WASO), TST, SE, and moving time during sleep (MT). SE was also calculated as the percentage of objective TST for each patient's actigraphy chart per TIB, recorded objectively on the sleep log by family members.

Because we were focusing on patients' incorrect cognition about sleep, we defined ‘changes in dissociation between the sleep log and actigraphically measured sleep’ as the primary outcome. We defined ‘changes in DBAS score’ as the secondary outcome.

With regard to the primary outcome, the differences in SONT, SOFT, SOL and TST measured on the sleep logs were compared with those obtained on actigraphy on the same night. A positive difference in SOL and SONT, and conversely a negative difference in TST and SOFT, indicated underestimation of the objective evaluation of sleep.

Thus the numerical changes in differences among the aforementioned four items were investigated. We compared the average differences before CBT with those after CBT. Moreover, the variances of the differences in the aforementioned four items before CBT were compared with those after CBT. We considered these differences and their variances to represent the dissociation between subjective and objective evaluation of sleep.

In addition, we examined the correlations between the changes in DBAS and the changes in the other parameters.

Statistical analysis

We examined the changes in the aforementioned parameters, for example sleep logs and actigraphy measurements, PSQI, DBAS, and the differences between the sleep logs and actigraphy, using the Wilcoxon signed-rank test. The variances of the differences were examined using F-test.

The correlations between the parameters were also analyzed using the Spearman rank correlation test. Differences at P < 0.05 were considered to be statistically significant.

RESULTS

Changes in insomnia parameters from the pre- to post-CBT period

On the PSQI, not only the total score, but also the scores for subjective sleep quality, SOL, and TST were decreased significantly at post-CBT evaluation (Table 2). As for DBAS, the scores for all of the five themes were decreased significantly at post-CBT evaluation (Table 2).

Table 2.  PSQI and DBAS before and after CBT (n = 20)
PSQIBefore CBT
(mean ± SD)
After CBT
(mean ± SD)
  • *

    P < 0.05,

  • **

    P < 0.01,

  • ***

    P < 0.001 (Wilcoxon signed-rank test).

  • CBT, cognitive behavioral therapy; DBAS, Dysfunctional Beliefs and Attitudes about Sleep Scale; PSIQ, Pittsburgh Sleep Quality Index; SE, sleep efficiency; SOL, sleep onset latency; TST, total sleep time.

Subjective sleep quality2.3 ± 0.61.2 ± 0.4***
Subjective SOL2.4 ± 11.6 ± 1**
Subjective TST2.3 ± 0.81.4 ± 0.9**
Subjective SE1.6 ± 1.11 ± 0.9
Difficulty in falling asleep1 ± 0.21 ± 0.2
Total score12.7 ± 38.9 ± 3***
DBAS  
Consequences of insomnia54.6 ± 32.929.5 ± 32***
Control and predictability of sleep49.2 ± 3326.3 ± 31.1***
Sleep requirement expectations41.1 ± 31.422.3 ± 30.4**
Causal attributions of insomnia33.8 ± 30.821.8 ± 31.1*
Sleep-promoting practices34.8 ± 34.617.1 ± 28.1**

Changes in the results of sleep log and actigraphy measurements

As shown in Table 3, evaluation of the sleep logs indicated that the reduction of SOL in the post-CBT period compared with that in the pre-CBT period was significant. Bedtime also became significantly later after CBT. Objective results also showed that both SOL and WASO measured on actigraphy became shorter after CBT. The SE was increased to >90%. MT was decreased significantly. As shown in Table 3, CBT improved both the subjective and objective sleep evaluations.

Table 3.  Sleep logs and actigraphy before and after CBT (n = 20)
 Sleep log Actigraphy
Before CBT
(mean ± SD)
After CBT
(mean ± SD)
Before CBT
(mean ± SD)
After CBT
(mean ± SD)
  • *

    P < 0.05,

  • **

    P < 0.01,

  • ***

    P < 0.001 (Wilcoxon signed-rank test)

  • CBT, cognitive behavioral therapy; MT, moving time during sleeping; NOA, number of awakening episodes lasting more than 5 min; SE, sleep efficiency (actigraphy: TST; sleep log: TIB); SOFT, sleep offset time; SOL, sleep onset latency; SONT, sleep onset time; TIB, total time in bed; TST, total sleep time; WASO, awakening time after sleep onset.

SONT (h)24.3 ± 1.623.8 ± 2.9*SONT (h)23.7 ± 1.423.9 ± 1.2
SOFT (h)5.9 ± 1.96.4 ± 1.3SOFT (h)6.8 ± 1.66.7 ± 1.2
SOL (min)68.3 ± 59.726.2 ± 20***SOL (min)30.8 ± 43.27.2 ± 9.9***
TST (min)330 ± 113.5352 ± 72.7*TST (min)396.6 ± 90.3391.4 ± 64.2
Bedtime (h)23.1 ± 1.423.7 ± 1.1*NOA (times)3 ± 3.12.9 ± 3.4
Rising time (h)7.1 ± 1.56.8 ± 1.1*WASO (min)24.8 ± 26.715.8 ± 17.2*
TIB (min)475.4 ± 111425.7 ± 64.4**SE (%)84.4 ± 12.792.1 ± 5.9***
   MT (counts/min)10.4 ± 6.38.4 ± 4.8*

Dissociation between subjective and objective evaluation

The patients realized subjectively that SOL was longer and TST shorter than the objective values at the pre-CBT evaluation. In the same way, SONT was subjectively later and SOFT earlier compared with the objective evaluations (Table 3). Thus, dissociation between the subjective and objective evaluations of sleep was demonstrated.

In Fig. 1 the differences are presented between sleep logs and actigraphy measurements for the 20 patients on the same night. After CBT we confirmed that the post-CBT data for SONT, SOL and TST converged and were closer to zero.

Figure 1.

Differences between subjective (sleep log) and objective (actigraphy) measurements between the (▴) pre- and (○) post-cognitive behavioral therapy (CBT) periods (n = 20). SOFT, sleep offset time; SOL, sleep onset latency; SONT, sleep onset time; TST, total sleep time. **P < 0.01; ***P < 0.001 (F-test).

We then validated the changes indicated in Fig. 1 statistically. As a result, after CBT, the differences between the sleep log and actigraphy for SONT and SOL were significantly decreased (P < 0.05), whereas those for SOFT and TST were significantly increased (P < 0.05). The pre-CBT and post-CBT SONT were 38.2 ± 58.2 min (mean ± SD) and 19 ± 20.7 min, respectively, and those for SOFT were −52.8 ± 74.6 min and −35.7 ± 53.2 min, respectively. The corresponding data for SOL were 37.4 ± 51.9 min and 19.1 ± 20.4 min, and those for TST were −66.5 ± 106 min and −36.6 ± 68.7 min, respectively.

After CBT, the variance of the aforementioned differences in SONT (P < 0.001, F = 5.4), SOL (P < 0.01, F = 4.2), and TST (P < 0.01, F = 4.1) were significantly lower than those before CBT.

Correlation between changes in DBAS items and those of sleep measurement parameters

The improvements in the ‘consequences of insomnia’ score on DBAS were positively correlated with the earlier objective rising times based on the sleep logs (P < 0.05, ρ = 0.48), and improvements in the ‘control and predictability of sleep’ score were positively correlated with reduction in subjective SOL based on the PSQI (P < 0.05, ρ = 0.51).

DBAS was not correlated with either actigraphy measurements or the dissociation between the sleep log data and actigraphy measurements.

DISCUSSION

Cognitive behavior therapy

It has been reported that combined CBT is more effective than each CBT component alone.3–5 We therefore chose combined CBT and every subject received the same treatment.

During the treatment process we showed patients their actigraphy data as a tool for assisting cognitive therapy. As a result, the patients understood that they could sleep better than they thought before CBT, and this understanding undoubtedly contributed to their achievement of an ideal bedtime, probably leading to both a reduction of TIB by 30 min and an increase of SE to >90% after CBT.

Changes in sleep measurement parameters after CBT

Subjective sleep evaluation based on the sleep logs, and the PSQI score, were improved after CBT. For PPI patients, a subjective rather than an objective evaluation seems to be much more important.9 Therefore improvements in subjective evaluation with CBT have clinical significance.

Dissociation between subjective and objective evaluation, and correlation between DBAS and sleep measurement parameters

If the differences between sleep logs and actigraphy measurements are close to zero, and if the variance of these differences decreases, this would indicate that most of the patients in the present study had appropriate subjective evaluation of sleep, that is, neither underestimation nor overestimation. Therefore we compared the values and the variance of these differences before and after CBT.

Comparison of the results of sleep logs with those of actigraphy during the pre-CBT period indicated underestimation of the objective evaluation of sleep.

Although before CBT there was wide scattering of the dissociation between sleep log data and actigraphy measurements among the patients, after CBT the values converged towards zero. Indeed the value and variance of the differences after CBT indicated statistically a shift towards modification of the patients' inappropriate subjective evaluation of sleep.

Thus, CBT effectively allowed most of the patients to modify their inappropriate subjective evaluation of sleep that was evident before CBT.

Through CBT, patients were able to realize that their sleep was better objectively than subjectively. Sequentially, they became more confident in terms of their sleep, and they were able to modify their underestimation. We also tried to identify incorrect cognition and attitude about why chronic insomnia had occurred, and to promote the understanding of patients about their fear of insomnia. We considered that these were the reasons why CBT allowed patients to cope with various incorrect cognitions about sleep.

Before CBT, subjects had complained about symptoms of insomnia, even though they had received full medication. On the basis of the present results, we consider that CBT was effective for subjects with pharmacologically intractable PPI.

We were also able to confirm the clinical efficacy of CBT for PPI using DBAS. Through CBT, the patients acquired correct knowledge and habits related to sleep. Consequently, the unrealistic cognition about sleep that patients held before CBT, for example ‘I must get 8 h of sleep to feel refreshed’, were revised.

We had assumed before the study that modifying incorrect cognition about sleep could improve evaluation of sleep. Therefore, we investigated the correlation between DBAS and sleep measurements. This showed that improvements in DBAS were correlated with an earlier rising time based on sleep logs and a reduction in SOL based on PSQI. These results suggested that modifying incorrect cognition about sleep was able to improve inadequate sleep hygiene and symptoms of insomnia. We found no evidence, however, for any correlation between actigraphy measurements or the aforementioned dissociation and DBAS in this study. Further study is needed to investigate the correlation between objective measurements, underestimation of objective sleep evaluation and incorrect cognition about sleep.

Study limitations

Not all of the present results can be considered to indicate the efficacy of CBT because this was not a randomized study. All of the subjects wished to receive individual CBT, and therefore might have been very motivated. It will be necessary to investigate the clinical efficacy of CBT in a further randomized study.

Usually, the PSQI is used as a measure of subjective sleep quality over 1 month, but in the presents study subjects were asked about sleep quality for 7 nights as post-CBT evaluation, because there was only a 5-week interval between before and after CBT.

Conclusion

The authors conducted individual CBT for 20 outpatients with PPI. We were able to observe the patients' underestimation of their objective sleep evaluation, and were also able to confirm the clinical efficacy of CBT, especially in terms of changes in incorrect cognition about sleep as represented by the decrease of dissociation between subjective and objective evaluation of sleep, that is, a decrease in the underestimation of objective evaluation of sleep leading to appropriate subjective evaluation of sleep, and the improvement of DBAS. Through CBT, patients were able to realize that their sleep was better objectively than subjectively, and acquired correct knowledge of sleep and related habits.

ACKNOWLEDGMENT

The DBAS was translated into Japanese by the first author by permission of Morin CM.

Ancillary