Summary of the literature

The strategy employed in reviewing research on MBI effects on physical health outcomes is to add to other comprehensive reviews (Carlson 2012; Piet and Hougaard 2011) by highlighting more recent, well-designed studies that may inform further understanding of efficacy and mechanisms of action. The focus is on quantitative studies and primarily randomized controlled trials (RCTs). While there is also an interesting and growing body of qualitative work on MBIs in many disease conditions (Mackenzie et al. 2007), space limitations make it impossible to cover that work here. The review is organized by disease type, rather than type of intervention or outcome studies, with diseases with a greater bulk of research summarized first.


Cancer is a leading cause of morbidity and mortality worldwide, with 13 million new cases and 7.6 million deaths recorded in 2008 alone (Boyle and Levin

2008). Due to improvements in treatments and increasing incidence rates, growing numbers of people are living longer after having been treated for cancer. In the USA, an estimated 13.7 million Americans with a history of cancer were alive on January 1, 2012 (American Cancer Society 2012). Cancer treatments and disease processes often leave survivors with symptoms and side effects such as lingering fatigue, sleep difficulties, pain, anxiety, depression, and worries about cancer recurrence (Carlson et al. 2004a). Hence, there is a significant need for psychosocial interventions to help people cope with the difficulties of cancer diagnosis and treatment.

As a response to this need, clinicians and researchers have been interested in applying MBIs to the vagaries of the cancer experience, and there is now a large body of work investigating their efficacy for patients with various types of cancer. This literature has been reviewed on several occasions since 2006 (Lamanque and Daneault 2006; Ledesma and Kumano 2009; Matchim and Armer 2007; Matchim et al. 2011; Musial et al. 2011; Shennan et al. 2011), with four reviews and meta-analyses just in 2012 and 2013 (Carlson 2012; Cramer et al. 2012; Piet et al. 2012; Zainal et al. 2013). A 2011 meta-analysis of 19 studies reported medium effect sizes on mood (d = 0.42) and distress (d = 0.48) (Musial et al. 2011) outcomes, consistent with earlier reviews. Two other meta-analyses focused only on breast cancer patients, reporting large effect sizes on stress (d = 0.71) and anxiety (d = 0.73) across nine studies with various designs (Zainal et al. 2013) and medium effects on depression and anxiety in three RCTs (Cramer et al. 2012). Piet et al. examined 22 randomized and non-randomized studies of cancer patients and reported moderate effect sizes on anxiety and depression in non-randomized studies (0.6 and 0.42, respectively), and slightly smaller effects for RCTs (Piet et al. 2012). Finally, Carlson reviewed all cancer- related studies applying levels of evidence criteria and concluded that there is Level 1 (highest) evidence for the efficacy of mindfulness-based interventions in oncology (Carlson 2012).

The first MBI study in a cancer population was published by our group in 2000, on the MBCR program. It was an RCT that assigned 89 patients with a variety of cancer diagnoses to either MBCR or a wait-list control (Speca et al. 2000). Patients in the program improved significantly more on mood states including anxiety, anger, and depression and on symptoms of stress such as physical tension, cardiopulmonary, gastrointestinal (GI), and cognitive symptoms than controls, with large improvements of approximately 65% on mood and 35% on stress symptoms. In a six-month follow-up, these improvements were maintained (Carlson et al. 2001). More home practice was associated with greater decreases in overall mood disturbance and the greatest improvements were seen on anxiety, depression, and irritability. Since that time many pre/post observational studies without comparison groups, and RCTs with usual care or wait-list control groups, have been published, citing improvements in a range of outcomes including quality of life (QL) domains such as emotional, social, role, and physical functioning, and psychological improvements on measures including stress symptoms, anxiety, depression, fear, and avoidance (for reviews see Ledesma and Kumano 2009; Matchim et al. 2011; Musial et al. 2011; Shennan et al. 2011).

While a substantial number of RCTs compare MBIs to wait-list or usual care controls, some with quite large sample sizes (Branstrom et al. 2012; Foley et al. 2010; Hoffman et al. 2012; Lengacher et al. 2009), still few studies have included randomization to active comparison groups. One exception is a threearmed trial in which Henderson et al. (2012) randomized 172 early-stage breast cancer patients into MBSR, a nutrition education program matched on contact time, or a usual care control condition. They also included follow-up assessments post-program (four months) and one and two years later. The MBSR group improved more than the other two conditions on a wide range of measures at the four-month post-program assessment, including quality of life, active behavioral and cognitive coping, avoidance, and spirituality, as well as depression, hostility, anxiety, unhappiness, meaningfulness, and several measures of emotional control. These group differences eroded somewhat over time, however, as participants in the other two groups continued to improve more slowly, so that by 24 months the only group differences apparent were on measures of anxiety, unhappiness, and emotional control, still favoring MBSR over usual care, but not the other active intervention. From this study it appears that MBSR participation may help speed up the natural course of cancer recovery across many domains, and also add a shift in perspective and skills in emotion regulation that are lasting.

More recently, we conducted two comparative effectiveness RCTs of the MBCR program: MINDSET and I-CAN Sleep. The MINDSET study directly compared MBCR to another active group intervention for cancer support, supportive expressive therapy (SET; Classen et al. 2008), in 271 distressed breast cancer patients and a minimal intervention control condition (a one-day stress management seminar) (Carlson et al. 2013). Similarities between interventions are the group format, size, structure, and contact hours. However, the two treatment modalities are quite distinct in their content, focus, and theoretical underpinnings, with the focus of SET on group support and emotional expression. We also included only distressed breast cancer survivors, in order to avoid floor or ceiling effects on outcomes, and included both psychological and biological outcomes. Participants in the control group were re-randomized to one of the active interventions, then everyone was followed after either MBSR or SET for a full year post-program to determine long-term outcomes. The study was also powered for moderator analyses to determine if outcomes differed across groups for people with different baseline characteristics, personalities, and preferences.

Pre-to-post program, women in MBCR improved more on stress symptoms compared with women in both the SET and control groups, on quality of life compared with the control group, and in social support compared with the SET group (Carlson et al. 2013), but cortisol slopes (a marker of stress responding) in both active intervention groups were maintained over time relative to the control group, whose cortisol slopes became flatter, where steeper slopes are generally considered to be healthier. The two intervention groups also maintained their telomere length, a potentially important marker of cell ageing, over time compared to controls (Carlson et al. 2014a). Over the longer-term followup of one year, the MBCR participants maintained all the benefits received in the group, while the SET participants still had higher levels of stress and mood disturbance and lower quality of life (Carlson et al. 2015, under review). This suggests that the MBCR group provided longer-term protection from distress for these women compared to those who did not participate.

The most preferred treatment (by over half of participants) was MBCR, and those who got their preferred treatment (regardless of what it was) improved more on quality of life and spirituality over time (Carlson et al. 2014b). Preference seemed to be a more powerful predictor of outcome than individual personality traits. This begins to tell us something beyond what can be learned from classic RCT designs: Preference matters. Treatment credibility and expectancy for benefit are likely important components in harnessing the power of the individual to produce meaningful change.

In another head-to-head comparative effectiveness trial, we also tried something novel for behavioral interventions by blinding participants to treatment. We did this by advertising the study simply as “I-CAN Sleep: Non-drug treatments for insomnia in cancer survivors,” and didn’t tell participants what the treatments were until they were already enrolled. Even then they only knew about the treatment they received, and didn’t know what the other treatment was (S. N. Garland et al. 2011). They were randomly assigned to either MBCR or cognitive behavioral therapy for insomnia (CBT-I), the gold-standard treatment for insomnia. This is a very tough test of efficacy for MBCR, and was designed as a non-inferiority trial to test whether the novel treatment for sleep (MBCR) was as good as the gold-standard.

In total, 111 patients with a variety of cancer types were randomly assigned to CBT-I (n = 47) or MBCR (n = 64) (S. N. Garland et al. 2014). Immediately postprogram, MBCR was inferior to CBT-I for improving the primary outcome of insomnia severity, but MBCR was non-inferior three months later. The time taken to fall asleep was reduced by 22 minutes in the CBT-I group and by 14 minutes in the MBCR group at follow-up. Similar reductions in wake time after sleep onset were observed for both groups. Total sleep time increased by 0.60 hours for CBT-I and 0.75 hours for MBCR. CBT-I improved sleep quality and dysfunctional sleep beliefs, and both groups resulted in reduced stress and mood disturbance. This indicated that while MBCR was slower to take effect, it could be as effective as the gold-standard treatment for insomnia in cancer survivors over time, and is a viable option for patients preferring that type of approach.

In another application, Monti et al. (2013) conducted a large RCT of a novel MBI called mindfulness-based art therapy (MBAT), also in distressed women with breast cancer. A total of 191 women were randomized to either an eight- week MBAT intervention or a breast cancer educational support program. Both groups improved on psychosocial stress and quality of life, but women with high stress levels at baseline improved only in the MBAT group. This shows that this type of intervention can be superior to other active interventions for those with higher need.

One final example of stretching the boundaries of traditional service delivery is a study we conducted of an online adaptation of MBCR for people living in rural and remote areas who didn’t have access to face-to-face MBCR groups (Zernicke et al. 2013). We conducted a wait-list RCT comparing those in the online program to a group randomly assigned to wait for the next online program. Our primary interest was in whether people would sign up, and if they would complete the program and get any benefit. The participants attended each week at a set time like an in-person group, and could see and interact with the instructor and the other participants in the online classroom using webcams and headsets. We enrolled 62 people and 83% of those completed the program (similar to in-person programs). All participants said the program either met (40%) or exceeded (60%) their expectations and all said they would recommend the program to other cancer patients. There were significant improvements and medium effect sizes in the online MBCR group relative to controls on total mood disturbance, stress symptoms, and spirituality (Zer- nicke et al. 2014).

In summary, evidence has accumulated through a number of high quality RCTs comparing MBIs to other active interventions showing superiority across a range of outcomes, in large sample sizes over substantial follow-up periods, but still mostly in breast cancer survivors. More work is needed for other types of cancer, late stage patients, and patients undergoing active treatments.

So far I have reviewed the clinical trials looking at psychological outcomes of MBIs in cancer patients, but a number of studies have also assessed their impact on biomarkers such as salivary cortisol and measures of immune functioning relevant to stress and cancer progression. For example, Carlson et al. (2003, 2004, 2007) measured immune and endocrine function pre/post MBSR in 59 breast and prostate cancer survivors, showing increased T cell production of IL-4 (an anti-inflammatory cytokine) and decreased interferon gamma (IFN-X) and NK cell production of IL-10. Patterns of change in cytokines over one year of follow-up also supported a continued reduction in pro-inflammatory cytokines (Carlson et al. 2007). In that group of patients, salivary cortisol profiles also shifted pre- to post-intervention, with fewer evening cortisol elevations found post-MBCR, and some normalization of abnormal diurnal salivary cortisol profiles (Carlson et al. 2004b). Over a year of follow-up, continuing decreases in overall cortisol levels were seen, mostly due to further decreases in evening cortisol (Carlson et al. 2007), recently replicated in a large RCT comparing MBCR to SET and control (Carlson et al. 2013). This is significant as higher cortisol levels, particularly in the evening, are considered to be a potential marker of dysregulated Hypothalamic-Pituitary-Adrenal axis functioning and poorer clinical outcomes, such as shorter survival times in metastatic breast cancer patients (Sephton et al. 2000).

Another group also showed decreased late-afternoon cortisol in combination with re-establishment of NK cell activity and cytokine production toward normal levels post-MBSR compared to controls (Witek-Janusek et al. 2006, 2008). Lengacher and colleagues (2011) reported increased response of T cells to antigen stimulation and an improved ratio of Th1/Th2 cytokines in early-stage breast cancer patients after a six-week MBSR program, and also found decreases in both cortisol and IL-6 pre-to-post session, and decreases in baseline levels across sessions in patients with advanced cancer and their caregivers (Len- gacher et al. 2012). The interpretation of these results is complex, but in general an anti-inflammatory environment is thought to be more favorable to cancer outcomes than one with elevated Th2 (pro-inflammatory) cytokines (Armaiz- Pena et al. 2009).

Measures of autonomic system function have also been conducted, since cancer survivors are at high risk for cardiovascular disease due to the toxicity of many cancer treatments. In a group of 72 women with various forms of cancer, weekly home blood pressure (BP) monitoring showed significant decreases in systolic blood pressure over the course of the program for women with higher premorbid BP in MBSR, compared to those in a comparable naturalistic waitlist group (Campbell et al. 2012).

In summary, the literature on cancer and MBIs is substantial and continues to grow, with improving quality of research design through the application of active control groups, larger samples, more diverse patient groups, longer follow-up periods, and a wide range of outcomes. Outcomes consistently favor MBSR over usual care and other active interventions immediately post-program across a range of psychological and QL outcomes. Its superiority over other active interventions over a longer period of time has yet to be definitively shown, though some specificity of effect is emerging. Its value in improving cancer-related biomarkers also still requires further investigation.

< Prev   CONTENTS   Source   Next >