Harnessing psychological processes to improve sleep

Recently Professor Allison Harvey published findings that support effectiveness of cognitive and transdiagnostic approaches to treating sleep disorders. The article was published by the American Psychological Association.

"Traditionally the development of psychosocial treatments has involved consensus among groups of skilled clinician researchers, while medication treatments have often been discovered by serendipity. There have recently been calls for increased use of scientific methods as the optimal path to derive efficient and effective treatments, and there are a number of examples of the use of science to derive psychosocial treatments. Indeed, while summarizing discouraging results from recent medication trials for treatment of severe mental illness, the director of the National Institute of Mental Health observed that “while psychosocial interventions have received much less marketing attention than pharmacological treatments, the results are arguably more encouraging” (p. 29) (Insel, 2009).

Following the treatment development approaches described by Clark (1999), Salkovskis (2002) and Onken, Carroll, Shoham, Cuthbert, & Riddle (2014), this article will describe two examples of treatment development processes that have used scientific methods. The two examples seek to improve sleep. Improving sleep is important because sleep problems are among the most common health problems. Insomnia, the most prevalent sleep disorder, is reported by around 10 percent of people. This rate markedly increases among individuals who have a psychiatric and/or medical disorder. Also, basic science has shown that sleep is critical for almost every basic human activity — including health, emotion regulation, learning, memory, creativity and appetite regulation. Hence, it is no surprise that, relative to good sleepers, individuals with insomnia report more psychological distress, more impairment in daytime functioning and more accidents, and they also take more frequent sick leave and utilize more health care resources (Sivertsen, Øverland, Bjorvatn, Mæland, & Mykletunb, 2009). Moreover, insomnia heightens the risk of developing depression, anxiety and substance-related problems (Baglioni et al., 2011; Breslau, Roth, Rosenthal, & Andreski, 1996).
Treating insomnia

Our program of research on chronic insomnia began more than a decade ago when we noticed that the cognitive level of explanation had been minimally studied. The small amount of work that had been published, along with the reports from our patients with insomnia, indicated a need to better understand the role of cognitive processes such as worry, rumination, attentional bias and unhelpful beliefs about sleep. So we began by scouring the existing literature for science related to cognitive processes in insomnia and in sleep and, following the pioneering work of Aaron T. Beck, we also examined the substantial literature on cognitive processes and cognitive therapy for other mental disorders. We summarized all that we found in a theoretical framework of the maintenance of insomnia at the cognitive level of explanation (Harvey, 2002). Our goal was to use this framework as a map for guiding the next phase of our research.

The central idea in the framework is that, regardless of the original trigger, chronic insomnia is maintained by a cascade of cognitive processes operating during the night and the day. The cognitive processes included in the framework are: worry/rumination, attentional bias toward sleep related threat, misperception of sleep and daytime functioning, unhelpful beliefs about sleep and the use of safety behaviors that prevent correction of unhelpful beliefs about sleep. We then set about empirically evaluating the framework both by conducting experiments that manipulate each cognitive process (see Kaplan, Talbot, & Harvey, 2009 for review) and by refining and testing a new cognitive treatment designed to reverse the cognitive processes specified in the framework.

The results from an initial open trial of this cognitive treatment suggested that reversing cognitive maintaining processes was indeed helpful for people with chronic insomnia. Insomnia severity was reduced and the cognitive processes we sought to reverse did indeed reverse (Harvey, Sharpley, Ree, Stinson, & Clark, 2007). But the results of an open trial are limited because we cannot rule out the possibility that they are simply due to the passage of time or to the positive effects of seeing a therapist.

Hence, in a next step, we began to collaborate with Charles Morin’s, MD, group at the Université Laval in Québec, Canada. We added the Laval cognitive approach targeting unhelpful beliefs about sleep to the cognitive treatment just described and then conducted a randomized controlled trial of 188 adult patients with chronic insomnia who were recruited and treated at UC Berkeley or at the Université Laval. Our main goal was to examine the unique contribution of behavior therapy (BT) and cognitive therapy (CT) relative to the full cognitive behavior therapy (full CBT). These treatments were individually delivered across eight weekly sessions.

There were significant improvements across all three conditions on measures of insomnia symptom severity, nighttime sleep disturbances and daytime functioning, and these improvements were generally sustained at six-month follow-up. The full CBT was associated with greatest improvements, the improvements associated with BT were faster but not as sustained and the improvements associated with CT were slower and sustained (Harvey et al., 2014). The latter result seems particularly interesting because the different trajectories of change may well provide unique insights into the process of behavior change via behavioral versus cognitive routes and they point to a need for future research to identify why an intervention targeting behavioral change generates faster improvement but is not as well sustained, while an intervention targeting cognitive processes generates slower but more sustained change.

These findings raise many questions. Is this pattern of findings specific to insomnia or do these findings replicate in BT versus CT for other conditions? Are the behavioral adjustments that are core to BT easier for a patient to implement when a therapist is available for “coaching?” Do we need more emphasis on establishing the behavioral recommendations as habits that the patient automatically reinitiates if/when insomnia recurs? Are there features of the procedures used in CT that are more conducive to habit formation? Does change to cognitive processes take longer but, once the skills are learned, are they more easily sustained?
Transdiagnostic approach

Another treatment development process at the core of our recent activities has been to pursue a relatively new approach in our field: to target research and treatment at transdiagnostic processes. This interest arose as we worked across a range of mental disorders and were struck by the similarities in the processes that maintain different disorders. The so-called “transdiagnostic processes” refer to processes that are common across more than one mental illness (Barlow, Allen, & Choate, 2004; Fairburn, Cooper, & Shafran, 2003; Harvey, Watkins, Mansell, & Shafran, 2004).

The potential advantages of studying, and intervening on, transdiagnostic processes are at least threefold. First, if a transdiagnostic process contributes to the maintenance of symptoms across multiple disorders, then one powerful approach is to focus treatment on that process rather than on the large number of discrete disorders currently listed in the DSM. Second, comorbidity among mental illness is the norm. Hence, a significant clinical dilemma is which disorder/s to prioritize for treatment. Targeting treatment at a transdiagnostic process provides one path forward (Harvey et al., 2004). Third, a transdiagnostic approach may reduce the heavy burden on clinicians, who must learn multiple disorder-focused protocols, with common theoretical underpinnings and interventions (Harvey et al., 2004). Indeed, this approach may solve the “too many empirically supported treatments problem” (p. 68) that impedes the dissemination and uptake of treatments (Weisz, Ng, & Bearman, 2014).

Promising progress has already been made toward developing transdiagnostic treatments that target transdiagnostic processes across the anxiety disorders and depression (Ellard, Fairholme, Boisseau, Farchione, & Barlow, 2010; Norton, 2008; Titov et al., 2011), eating disorders (Fairburn et al., 2009), schizophrenia (Bentall et al., 2009) and bipolar disorder (Ellard, Deckersbach, Sylvia, Nierenberg, & Barlow, 2012), and across anxiety, depression and conduct problems in youth (Weisz et al., 2012). Treatments targeting transdiagnostic processes, such as rumination (Nolen-Hoeksema & Watkins, 2011) and perfectionism (Egan, Wade, & Shafran, 2011), have also been effective (Riley, Lee, Cooper, Fairburn, & Shafran, 2007; Watkins et al., 2011).

Our team was an early contributor to discussions on transdiagnostic approaches (Harvey et al., 2004). In particular, we highlighted sleep and circadian dysfunction as a biologically (Harvey, Murray, Chandler, & Soehner, 2011) and theoretically (Harvey, 2008) plausible transdiagnostic contributor to mental illness and we proposed a transdiagnostic treatment for sleep and circadian disturbance (Harvey, 2009). In addition to noting that sleep problems were common across the disorders, we also noticed that the majority of studies within the sleep field were disorder-focused — that is, they tend to treat a specific sleep problem (e.g., insomnia). In our experience real life sleep and circadian problems are not so neatly categorized, particularly in severe mental disorders. Indeed, insomnia can overlap with hypersomnia (Kaplan & Harvey, 2009), delayed sleep phase (Giglio et al., 2010) and irregular sleep-wake schedules (Gruber et al., 2009). Hence, we wondered if it would be beneficial for traditional CBT for insomnia (CBT-I) to be modified to address this broader range of sleep dysfunctions.

Our initial test of this idea was with individuals who met diagnostic criteria for bipolar disorder who also had insomnia. We selected bipolar disorder for four reasons. First, sleep disturbances are prominent features of bipolar disorder. Second, the sleep disturbance escalates just before an episode and worsens during an episode. Third, there is empirical evidence indicating that sleep disturbance may be one causal pathway that leads to relapse in bipolar disorder. Fourth, sleep disturbance in bipolar disorder is complicated and often includes features of insomnia, delayed sleep phase and irregular sleep-wake schedules.

We again scoured the scientific literature to gain guidance on how to target these complicated features of sleep in bipolar disorder. This time we realized that we may be able to target them by supplementing CBT-I with elements from three existing evidence-based treatments: interpersonal and social rhythm therapy (Frank et al., 2005), chronotherapy (Wirz-Justice, Benedetti, & Terman, 2009) and motivational enhancement (Miller & Rollnick, 2002). In a pilot test of this transdiagnostic approach, alongside standard psychiatric care, individuals who met diagnostic criteria for bipolar disorder and who were also interepisode were randomly allocated to the new treatment — the bipolar disorder-specific modification of cognitive behavior therapy for insomnia (CBTI-BP; n = 30) — or psychoeducation (PE; n = 28) as a comparison condition.

Outcomes were assessed at baseline, the end of eight sessions of treatment and six months later. During the six-month follow-up, the CBTI-BP group had fewer days in a bipolar episode relative to PE (3.3 days vs. 25.5 days). The CBTI-BP group also experienced a significantly lower hypomania/mania relapse rate (4.6 percent vs. 31.6 percent) and a marginally lower overall mood episode relapse rate (13.6 percent vs. 42.1 percent) compared to the PE group. Also, relative to PE, CBTI-BP reduced insomnia severity and led to higher rates of insomnia remission at post-treatment and marginally greater at six months. Both CBTI-BP and PE showed statistically significant improvement on selected sleep and functional impairment measures. The effects of treatment were well sustained through follow-up for most outcomes, although some decline on secondary sleep benefits was observed (Harvey et al., in press). Hence, it appears that CBTI-BP was safe for individuals who meet diagnostic criteria for bipolar disorder (Kaplan & Harvey, 2013) and associated with reduced risk of mood episode relapse and improved sleep and functioning on certain outcomes.

So this is very encouraging. We have two demonstrations that a short (eight session) treatment is helpful for people with chronic insomnia and people with bipolar disorder who also experience significant sleep disturbance. The two new approaches tested were derived from science — cognitive therapy for insomnia was derived from more basic science on cognitive processes, and CBTI-BD was derived from existing evidence-based treatments. These results add to the evidence that we have powerful methods for improving sleep (Morin et al., 2006; Taylor & Pruiksma, 2014).

What next? Our team has become passionate about helping our field solve the dissemination problem. In particular, we are interested in “deployment treatment development” (Weisz et al., 2014) in which interventions are developed and tested with the clients and therapists who will receive and deliver the intervention, and within the organizations in which they will be delivered. This approach is in contrast to the usual method, which we have used in the past, in which the treatment is developed within a university or research setting and later tested for applicability to real world settings. As part of the deployment focused effort, we continue to develop, test and expand the transdiagnostic approach to sleep problems.

We are also interested in interweaving these sleep treatments into other evidence-based treatments. Given that mental illness and sleep problems appear to be mutually maintaining (Harvey, 2008), the interwoven treatment may represent a significant advance over the usual “sequential” approaches in which one of the comorbid mental illness or the sleep disturbance is the target of treatment, then the focus switches to the other disorder. Indeed, in forthcoming work conducted in collaboration with Greg Clarke, MD, of Kaiser Permanente, Oregon, we provide evidence of the efficacy of an interwoven treatment involving weekly delivery of CBT for depression and insomnia to youth, with improvements obtained in outcomes related to each condition.

By Allison G. Harvey, PhD

This research program has been supported by the National Institute of Mental Health (RO1 MH079188, R34 MH080958, R34 MH082034).

Allison Harvey, PhD, is a professor of clinical psychology at the University of California, Berkeley, where she is also the director of the clinical science program, the psychology clinic and the Golden Bear Sleep Research Clinic. After completing her training in Australia, Harvey moved to the University of Oxford as a postdoctoral fellow and then as a faculty member and fellow of St. Anne’s College. In 2004 she moved to UC Berkeley. Harvey's team aims to develop more effective treatments for psychiatric and health problems by (a) using basic science findings across levels of explanation as sources for deriving novel interventions and (b) using intervention research to develop and test hypotheses about mechanisms. Harvey’s research is funded by NIMH, NICHD and NIDA. Her research has been acknowledged with various awards including an Honorary Doctorate from the University of Orebro, Sweden. Author website."