More than a thousand articles on mental disorders are published in medical journals each month! Also, clinicians have limited training, time, and inclination to keep up with reading research articles critically on a regular basis. Thus, a disturbing disconnect (for which there are no easy solutions) exists between clinical research and usual clinical practice.
Only a very small fraction of research articles report findings with immediate clinical utility. In this new column, I will present succinct descriptions of selected research articles and discuss their clinical implications. I will also use discussion of these articles to provide tips about interpreting and applying research findings. Please e-mail me at firstname.lastname@example.org with your questions and comments.
This month’s column focuses on a recent JAMA article1 on the treatment of chronic insomnia—an important condition because it affects millions of people in this country.2
Summary of the study
Both cognitive-behavioral therapy (CBT) and hypnotics are effective in chronic insomnia, but many patients have incomplete improvement with either treatment alone. Also, many patients may need continued treatment after the acute phase.
Because of these 2 issues, this study addressed the following questions:
• Is there any benefit from adding a hypnotic to CBT in the acute phase?
• Does adding a hypnotic to CBT in the acute phase improve the longer-term outcome over 6 months?
• For patients treated in the acute phase with a combination of CBT and a hypnotic, should the hypnotic be discontinued after the acute phase is over?
• For patients treated with CBT alone in the acute phase, are further intermittent “maintenance” CBT sessions over 6 months helpful?
This was a randomized, controlled trial involving 160 patients (mean age, 50 years) with chronic insomnia (ie, insomnia lasting 6 months or more). Patients whose insomnia was secondary to another specific illness (eg, progressive medical illness, a medication adverse effect, current major depressive disorder, sleep apnea, restless legs syndrome) were excluded. For the 6-week acute phase, patients were randomized to CBT alone (1 group session per week) or combined treatment (CBT plus zolpidem(Drug information on zolpidem), 10 mg at bedtime). Details of the type of CBT used have been described elsewhere.3
After the acute phase, patients were again randomized to different treatment groups for the next 6 months. Those who had received CBT alone in the acute phase received either 1 maintenance group CBT session per month or no further treatment.
Patients who had received combined treatment in the acute phase received either combined treatment (CBT plus zolpidem, with zolpidem now being used intermittently) or CBT alone.
To summarize: on the basis of treatment received during the acute and maintenance phases, patients were divided into 4 categories: CBT-CBT, CBT–no treatment, combined-combined, and combined-CBT.
TIP: In any clinical trial, look for the completion rate overall and in each treatment group, because “dropouts” can significantly bias a study unless this can be effectively controlled for statistically.
About 80% of the patients completed the study—a rate similar across the treatment groups.
Is there any benefit to adding a hypnotic to CBT in the acute phase?
In the acute phase, both the CBT alone and the combined treatment groups showed equal mean improvement in mean sleep latency, time awake after sleep onset, and sleep efficiency (from daily diaries).
TIP: Mean values of variables and mean differences between groups are usually reported in studies but are not enough because they combine patients who did have a substantial response with those who did not. We need to also consider what percentage of patients showed a clinically useful response.
In the acute phase, patients were categorized as “responders” and “remitters” based on predefined levels of improvement in the Insomnia Severity Index (a patient-rated scale). The proportion of responders (about 60%) and remitters (about 40%) was similar in the CBT alone and combined treatment groups. There were no statistically significant polysomnographic differences between the CBT and combined treatment groups either. The only difference was that combined treatment was slightly more effective than CBT alone in increasing total sleep time.
Does adding a hypnotic to CBT in the acute phase improve the longer-term outcome (ie, over 6 months)?
Patients who received combined treatment in the acute phase seemed to have a higher remission rate at 6 months (56%) than those who received CBT alone (43%). However, this difference was not statistically significant.
TIP: If a finding is not “statistically significant,” it means that on the basis of what this study actually found, it is not ruled out as statistically improbable that the 2 treatment groups had similar outcomes. So, this apparent difference (56% vs 43%) could have simply been due to chance. Therefore, we must assume that there is no real difference between the 2 groups.
Thus, this study did not find that the addition of a hypnotic to CBT in the acute phase improves outcome over the subsequent 6 months.
For patients treated in the acute phase with a combination of CBT and a hypnotic, is it better to continue the hypnotic after the acute phase?
Patients in the combined-CBT group had a higher remission rate (68%) at the end of 6 months than those in the combined-combined group (42%)—a statistically significant difference. So, patients who received both CBT and a hypnotic in the acute phase and who then continued (intermittent) use of the hypnotic after the acute phase was over had worse outcomes than those who discontinued the hypnotic completely.
For patients treated with CBT alone in the acute phase, are further monthly maintenance CBT sessions over the next 6 months helpful?
At the end of 6 months, 85% of patients in the CBT-CBT group and 88% in the CBT–no treatment group were either moderately or markedly improved. This indicates that there was no additional benefit with once a month maintenance group CBT sessions after the acute phase. (These percentages should not be compared with the percentages of responders above; they were based on different criteria.)
Putting the results into clinical context
TIP: In understanding whether and how to apply the results of any study to your patients, ask: (1) How convincing are the findings? (2) Are the findings consistent with those of other studies? (3) Are the findings generalizable to the kind of patients I see and to possible modifications of the treatment?
Although this column is mainly limited to discussion of the JAMA study, it is important to note that its findings are generally consistent with those of previous studies.
For the acute phase, although this study did not address this specific question, a meta-analysis of 21 small studies comparing CBT alone with a hypnotic alone in chronic insomnia found no difference on any measure except greater improvement in latency of onset of sleep in the behavior therapy group.4 A crucial fact to remember is that hypnotics work only as long as they are taken, while the effects of CBT seem to be sustained even after the few initial sessions are over.4-6
The addition of a hypnotic to CBT had only limited benefit (increased total sleep time but no advantage on any of several other clinical and polysomnographic measures). This is despite the fact that (because there was no pill-placebo group) patients knew that they were definitely receiving an active medication, which tends to exaggerate the effects of any drug. Other studies have also reported limited or no benefit to adding medication to the CBT.4-6
Nevertheless, I think that this conclusion needs nuanced interpretation for several reasons. First, patients who participate in such studies are recruited by referrals and advertising, and are probably motivated to accept CBT alone or with a hypnotic. Second, we do not know whether many patients agreed to participate because they had responded poorly to a hypnotic in the past. Third, the percentage of remitters at 6 months was numerically higher in the CBT plus hypnotic group. Although this could have been due to chance (ie, not “statistically significant”), an alternative explanation could be that this is a real difference, but this particular study did not have enough patients to demonstrate it with statistical confidence. Fourth, is it possible that higher dosages or a different hypnotic may be more effective?
While CBT alone should be used as first-line treatment for most patients with chronic insomnia, probably we should consider adding a hypnotic for some patients. Even if a hypnotic is added in the acute phase though, it seems best to not continue its intermittent use after the acute phase. This worsens the outcome, but speculation about the potential reasons for this is beyond the purview of this column.
For patients who received CBT alone in the acute phase, maintenance CBT sessions did not provide any additional benefit. However, more frequent CBT sessions could be tried for patients with residual insomnia after the acute phase.
Only 15% of patients in this study had a comorbid psychiatric diagnosis and none were currently ill. The principles of CBT for insomnia used in this study were general, however, and were not focused on issues particular to specific comorbid disorders. We could tentatively act on the assumption that these results largely apply to many patients with psychiatric disorders and chronic insomnia.
Because of the high prevalence of chronic insomnia, all clinicians should become skilled at providing CBT for patients with this disorder. An example of a cognitive intervention provided was to address myths about sleep—eg, that there is an absolute need to sleep 8 hours every night or that all daytime impairments are the result of poor sleep. Behavioral “sleep hygiene” measures should be discussed with the patient on an ongoing basis. Please click here for a patient handout of sleep hygiene measures.
Despite convincing data, CBT for insomnia is used very infrequently—even though group CBT is less expensive than medication because its effects last even after the initial few sessions (6 in this study), it has no adverse effects, and its effects are longer- lasting. Physicians could devote parts of some sessions to providing CBT or collaborate with another clinician who provides individual or group CBT as an alternative to groups.