Cognitive behavioral therapy for treatment of chronic primary insomnia: a randomized controlled trial, 2001, Edinger et al.

[ME/CFS Skeptic]

(Note that this is an old trial from 2001)

Abstract
Context: Use of nonpharmacological behavioral therapy has been suggested for treatment of chronic primary insomnia, but well-blinded, placebo-controlled trials demonstrating effective behavioral therapy for sleep-maintenance insomnia are lacking.

Objective: To test the efficacy of a hybrid cognitive behavioral therapy (CBT) compared with both a first-generation behavioral treatment and a placebo therapy for treating primary sleep-maintenance insomnia.

Design and setting: Randomized, double-blind, placebo-controlled clinical trial conducted at a single academic medical center, with recruitment from January 1995 to July 1997.

Patients: Seventy-five adults (n = 35 women; mean age, 55.3 years) with chronic primary sleep-maintenance insomnia (mean duration of symptoms, 13.6 years).

Interventions: Patients were randomly assigned to receive CBT (sleep education, stimulus control, and time-in-bed restrictions; n = 25), progressive muscle relaxation training (RT; n = 25), or a quasi-desensitization (placebo) treatment (n = 25). Outpatient treatment lasted 6 weeks, with follow-up conducted at 6 months.

Main outcome measures: Objective (polysomnography) and subjective (sleep log) measures of total sleep time, middle and terminal wake time after sleep onset (WASO), and sleep efficiency; questionnaire measures of global insomnia symptoms, sleep-related self-efficacy, and mood.

Results: Cognitive behavioral therapy produced larger improvements across the majority of outcome measures than did RT or placebo treatment. For example, sleep logs showed that CBT-treated patients achieved an average 54% reduction in their WASO whereas RT-treated and placebo-treated patients, respectively, achieved only 16% and 12% reductions in this measure. Recipients of CBT also showed a greater normalization of sleep and subjective symptoms than did the other groups with an average sleep time of more than 6 hours, middle WASO of 26.6 minutes, and sleep efficiency of 85.1%. In contrast, RT-treated patients continued to report a middle WASO of 43.3 minutes and sleep efficiency of 78.8%.

Conclusions: Our results suggest that CBT represents a viable intervention for primary sleep-maintenance insomnia. This treatment leads to clinically significant sleep improvements within 6 weeks and these improvements appear to endure through 6 months of follow-up.

Full text: https://jamanetwork.com/journals/jama/fullarticle/193729

 

[ME/CFS Skeptic]

Thought this was an interesting study because the authors tried to blind patients and therapist by providing a sham intervention to CBT.

There were three groups:

• CBT with sleep education and time-in-bed restrictions
• RT: progressive muscle relaxation train
  
• PR: a placebo/sham intervention that was based on eliminating presumed “conditioned arousal,” which prolongs nocturnal awakenings. If I understand correctly it consisted of identifying things patients did when they wake up and then trying to eliminate the association with not being able to fall asleep.

The latter was made up for the study but patients and therapists weren't aware of this. The study writes:

"Enrollees were blinded to hypotheses and the nature of the PT, but they were told they had a 1 in 3 chance of PT assignment. Therapists were blind to hypotheses and were uninformed that 1 of the treatments they administered was a placebo."
​

The authors also used an objective outcome (polysomnography) and it seems that CBT outperformed the other two groups for most outcomes.

 

[Trish]

The authors also used an objective outcome (polysomnography) and it seems that CBT outperformed the other two groups for most outcomes.

• CBT with sleep education and time-in-bed restrictions

I'm puzzled by this.

Was there any way to tell whether it was the CBT, or the sleep education and time in bed restrictions that led to the better outcome in that group? Did the other groups have sleep education and time in bed restrictions?

 

[Hutan]

There's lots that is good about this study. But, surely the main objective outcome of a treatment for insomnia is greater sleep times?

At baseline, the sleeplog (patient report, but multiple nights) and polysomnography (highly objective, but just one night) suggested that the participants were averaging slightly less that 6 hours sleep a night
sleep logs: CBT 5.8; RT 5.3; placebo 5.8; polysomn: CBT 6.0; RT 5.7; Placebo 5.9

Immediately post-treatment, the participants were still averaging around 6 hours sleep a night
sleep logs: CBT 6.0; RT 6.0; Placebo 6.0; polysomn: CBT 6.2; RT 5.6; placebo 5.6

It's not very impressive, considering the CBT participants had also had the benefit of a whole lot of advice on sleep hygiene. in fact, to me it looks like evidence that sleep hygiene advice doesn't do much at all, even for this sample of people who were carefully selected to have the sort of insomnia that you might have expected to respond to behavioural treatments.



At followup, six months later, we are told that the

the average CBT-treated participant could expect to achieve a mean subjective sleep time of slightly over 6 hours (Figure 4) which, given what is known about human sleep requirements,55 appears minimally normative/sufficient. Finally, CBT appeared superior to the other treatments in normalizing ISQ scores which reflect perceived sleep/wake functioning.

So, it seems around 6 hours of sleep (which is what the participants entered the study with) is now enough. The old, 'changing the definition of recovery' trick. And the CBT participants were better at filling out surveys positively.


It's important to note that the data for followup does not include the placebo treatment, apparently because the researchers felt these people had already done enough.

After their posttreatment assessments, CBT and RT recipients were asked to return for their final outcome assessment 6 months later. Given their time already invested in the study, the PT patients were not asked to complete the additional 6-month follow-up before receiving active treatment. Instead, they were debriefed and immediately offered active treatment with their previously assigned therapist. Those who accepted were randomized to 6-week courses of CBT (3 women, 6 men) or RT (4 women, 2 men) but their subsequent data were not considered in any of the statistical comparisons conducted. To maintain their blinding, therapists were told that PT recipients were offered a second, more tested treatment, because their initial treatment was a new therapy that had not yet received sufficient testing to justify its isolated use. They were also told that the PT was a promising treatment deserving of the scrutiny provided by this project.

Here's the relevant part of Figure 4:




That's not very easy to see, but if we just concentrate on the people who turned up for the follow-up assessment and provided a sleep log (15 or 14 out of the 25 who started, in each treatment arm) rather than those who declined followup (for whom there was some kind of imputation of outcomes):
CBT: 5.8 hours of sleep per night after treatment >> 6 hours of sleep at 6 months followup
Relaxation therapy: 5.9 hours of sleep per night after treatment >> 6.1 hours of sleep at 6 months followup

tldr:
So, basically, there has been a whole lot of effort, and all the participants, regardless of treatment, are likely to still be getting about 6 hours of sleep a night.

And so, yes, the researchers appear to have worked hard to eliminate bias, but then they still managed to not overcome it. They concluded that:

Given our results, CBT may have a deserving and important niche in the clinical management of PPI patients with sleep maintenance difficulties

PS. There's more that could be said about the problems with this study, but I thought the lack of change in sleep time combined with a claim of a positive outcome for CBT was probably the most fundamental.

 

[Hutan]

I've been quite annoyed with this study, with their use of minutes rather than hours of sleep time, so the fact that no one was getting a normal sleep time was difficult to see. And their charts with truncated axes, so you think tiny differences matter.

Here's a chart of the data, with the x axis being hours of sleep (0 to 8 hours). It's not a great chart, I can't be bothered making it good. But it's still way better than what this group of researchers produced, and they had months to think about how they could best present their data.



First, note that all of the sleep length times are from 5.3 to 6.2 hours.

Then look at the bottom three blue/orange lines (the CBT lines). See how the change from baseline to post-treatment is hardly anything, and that gain is gone by followup.

Now look at the middle three yellow/grey lines (the relaxation treatment lines). See how the baseline sleep log (grey) was quite low (5.3 hours). Look how the sleep log times (grey) actually increased a fair bit from baseline to followup.

Now look at the top two blue/green lines (the placebo treatment lines). Look at the blue sleep log data there (in light blue). See how it looks just the same as the sleep log data for CBT (in dark blue), except of course that the placebo arm has no followup data.

It would take quite a mind to decide that CBT had made a material positive difference to people's lives from that data, and that it is the best treatment. Especially considering the small sample size of this study, and the decrease in quality of life that probably results from having a regimented bed time, and a requirement to get out of bed immediately upon waking.

 

[Arvo]

Am I missing something? I fail to see where exactly the "C" from CBT comes into play, or which dysfunctional beliefs/cognitive distortions are being addressed here.

The whole core idea of CBT is that those drive the B, distorted behaviour.

People were getting sleep hygiëne instructions that they must follow, and some sort of sleep "training". The treatment was a package of behavioural things, but apart from mentioning that poor sleep hygiëne is considered a contributory to PPI (and therefore asking participants to report 1 or more sleep-disturbing practise at enrolment, which after that mention is completely ignored for the rest of the paper?), there isn't any confirmation anywhere that participants were specifically recruited for the study because of it, or that the study specifically revolved around changing obvious unhelpful behaviour. (So, in a way, there goes your B as well.)

The only "C"-ish part of the treatment was "The CBT recipients were first presented a standardized audio-cassette cognitive therapy module designed to correct misconceptions about sleep requirements and the effects of aging, circadian rhytms, and sleep loss on sleep/wake functioning."
So they got a tape with educational material. "Correcting misconceptions" can sound a bit CBT-ey, if you squint your ears, but this doesn't seem to be about changing faulty cognitions, but providing accurate information in case someone didn't know certain facts. (And there is no follow-up anywhere.)


It even makes me wonder a bit if there was funding budget for CBT, because it was "hot" at the time, and the authors wrote their study about adressing sleeping problems with a combination of education, sleep hygiene and "training "the body to sleep, in a way that they could get grant money under the CBT banner.*

*Edited to add: Which isn't necessarily a "bad" thing by them, as that's how these things go in practise, although it would have muddied the water regarding CBT

 

[ME/CFS Skeptic]

Then look at the bottom three blue/orange lines (the CBT lines). See how the change from baseline to post-treatment is hardly anything, and that gain is gone by followup.

Thanks for the clarification.

I was a bit confused by their data because they report the standard error instead of the standard deviation. The latter is 4-5 times larger than the former.

EDIT: I made an error here. I seem to have mistaken to data of sleep logs with baseline data. Apologies for the confusion, thanks to Hutan for pointing this out (see posts below).

So the 12.4 min increase in sleep time measured with polysomnography is indeed quite small. It seems that it mostly reached statistical significance compared to the other groups because those groups experienced a decrease of approximately 25 minutes.

 

[Hutan]

So the 12.4 min increase in sleep time measured with polysomnography is indeed quite small. It seems that it mostly reached statistical significance compared to the other groups because those groups experienced a decrease of approximately 25 minutes.

I'm not sure about your numbers? From Tables 2 and 3, for total sleep time as determined by polysomnography:
CBT: baseline 361.6, post-treatment 372.4 >> increase of nearly 11 minutes
Relaxation therapy: baseline 342.2, post-treatment 337.9 >> decrease of 4 minutes
Placebo treatment: baseline 352.5, post-treatment 334.0 >> decrease of 18.5 minutes
Have I mixed something up?

But yes, small differences, and the polysomnography was just for one night, whereas the sleep logs were for multiple nights. With the small numbers of patients and the very small sample of polysomnography data points (nights), trends in that data are probably not telling us very much, especially when the differences between treatments fall within the variability that was seen at baseline.

While I am here, the blinding of the therapists was probably not that good, and potentially of the later participants too. They had people with the placebo treatment getting to the 6 month mark, being told that they had had the placebo treatment (they were asked not to tell their therapist). These patients were not asked to provide followup data, and were offered the opportunity to do one of the two 'real' treatments - which were delivered by the same therapist that they had for the placebo treatment. So, therapists would have had an idea that the placebo treatment was regarded as 'not as good' while they were still providing it to later recruits.

'Sleep efficiency percentages' sound really important, but are just total sleep time as a percentage of total time in bed. The CBT participants were told to get out of bed if they had significant awakenings, whereas the other participants were not. So, the researchers effectively 'gamed the system' - making the sleep efficiency of the CBT group look better, when there was no real improvement in sleep.

There's still more...

 

[ME/CFS Skeptic]

I'm not sure about your numbers? From Tables 2 and 3, for total sleep time as determined by polysomnography:
CBT: baseline 361.6, post-treatment 372.4 >> increase of nearly 11 minutes
Relaxation therapy: baseline 342.2, post-treatment 337.9 >> decrease of 4 minutes
Placebo treatment: baseline 352.5, post-treatment 334.0 >> decrease of 18.5 minutes
Have I mixed something up?

Oops, my mistake. I seem to have mistaken to data of sleep logs with baseline data. Apologies for the confusion, thanks for your analysis of the paper.

 

[Jonathan Edwards]

The slightly disappointing thing is that it might seem that one could use studies like this to show that if you design your study badly enough you can get amazing results with CBT for anything, but sadly the results are rather short of amazing.