Is CBT-I Effective for Pain?

In 2006, the American Academy of Sleep Medicine Standards of Practice Committee¹ issued a call for more research studies on the behavioral treatment of insomnia to examine whether sleep intervention leads to an improvement in insomnia comorbid with other medical or psychiatric conditions. The study by Vitiello and colleagues is a significant contribution to this effort by demonstrating the sustained effectiveness of cognitive behavior therapy for insomnia (CBT-I) for sleep in patients with osteoarthritis. Vitiello et al.² found medium effects for changes in sleep-onset and sleep-continuity measures 1 year after the conclusion of CBT-I. This is one of the longest follow-up time periods employed in a study using state-of-the-art sleep assessments to examine the effect of CBT-I in patients with pain and insomnia.

Let's face it—CBT-I is a great treatment for sleep. The effects for CBT-I appear to translate from primary to comorbid insomnia in a variety of pain conditions. The results from this paper further support this premise.

And what about the pain outcomes reported in their study? Given the large effects of CBT-I on sleep, it was a disappointment to see the small effect that CBT-I had on pain. Despite statistical significance, the effect sizes for pain in this study ranged from 0.17 to 0.31, which is the range classified as a small effect by Cohen. Moreover, it should be emphasized that this effect is a within-group change (pretreatment to posttreatment), not a between-group change. We would expect between-group effects (CBT-I compared with a control group) to be even lower. Finally, the effect sizes reduced even more 1 year after treatment, such that the pain findings did not maintain statistical significance. Unfortunately, these weak findings are consistent with those of the parent randomized controlled trial, of which these data are a subset. They are also consistent with the findings from randomized controlled trials conducted by Edinger et al.³ in patients with fibromyalgia and Currie et al.⁴ in their heterogeneous groups of patients with chronic pain difficulties.

From a behavioral sleep medicine perspective, this outcome is hard to stomach. CBT-I should help more with pain—we know that sleep disturbance lowers the pain threshold, so, fixing the sleep should help pain intensity. Vitiello and colleagues provide an excellent theoretical model explaining these linkages. The small effect size does not appear to be due to weak sleep outcomes. It is true that CBT-I only had a small effect on total sleep time in this study. However, research suggests that sleep quality and time awake are excellent correlates of pain. The effect size for wake time after sleep onset in this study is a substantial 0.72.

A reasonable question is: Would an integrative therapy including both CBT-pain and CBT-I improve pain outcomes more?

A recent Cochrane System Review meta-analysis⁵ on CBT for pain suggests that the answer is “probably not.” The review showed that, when comparing CBT-pain to an active control condition, there is only a weak effect for a reduction in pain intensity. In other words, CBT for pain and CBT I seem to work about the same in terms of their outcomes on pain intensity— they help a little, but not much. Pain may be an outcome that has only a limited range to change. This is, after all, what we teach our patients with chronic pain—that it is time to switch to an acceptance model because the pain is not going to go away.

Since the literature base seems to indicate that effects on pain are small, it makes sense to try to target other pain problems that appear to be more treatable. Thanks to Vitiello and colleagues, it is clear that CBT-I does an excellent job of treating insomnia in patients with osteoarthritis. In addition, the work gives further evidence to the idea that CBT-I is an effective treatment for insomnia in patients with comorbid pain disorders.

DISCLOSURE STATEMENT

Dr. Haynes has indicated no financial conflicts of interest.