The primary goal of this study was to examine predictors of adherence to Iyengar yoga classes for breast cancer survivors. Overall, adherence to Iyengar yoga classes in this study was modest (63.9%) but comparable to other studies of yoga in breast cancer survivors.[33
] In a randomized controlled trial, Moadel et al
] reported participants attended on average 7 out of 12 classes (58.3%). Similarly, Danhauer et al
] reported a 58% adherence rate from a pilot randomized study of women with breast cancer participating in 10 weekly, restorative yoga classes. Vadiraja et al
] compared the impact of an integrative yoga program to brief supportive therapy among breast cancer patients receiving adjuvant radiation therapy and reported that, out of a possible 36 sessions, 29.7% of participants’ attended 10 to 20, 56.7% attended 20 to 25 sessions, and 13.7% attended >25 over 6 weeks.
A major difference between the studies discussed and our own is the treatment status of participants. A portion of participants in the Moadel et al
] and Danhauer et al
] studies were concurrently undergoing treatment for breast cancer (48% and 34% respectively), which may account for their lower adherence rates. Adherence in our study is also comparable to studies in healthy populations.[36
Although several studies have examined medical, demographic and psychosocial predictors of adherence to yoga.[33
] to the best of our knowledge, this is the first study to examine predictors of yoga adherence within a theory of planned behavior framework. Key findings from our study are that several motivational variables predicted adherence to the Iyengar yoga program. We found that women intending to attend more classes had significantly greater adherence than women intending to attend fewer classes. Similarly, Courneya et al
] found intention to be an independent predictor of adherence in breast cancer survivors training for a dragon race competition.
We also found that self-efficacy and instrumental attitude (benefits) were significant predictors of adherence. Women who selected 7 “extremely” to indicate their confidence that they could complete the exercises in the classes and for the perceived benefits from participation achieved significantly greater adherence than women selecting ≤6. Likewise, in a home-based exercise trial for breast cancer survivors, Pinto et al
] found that baseline self-efficacy significantly predicted adherence. Moreover, in a study by Duncan et al
] cancer patients participating in Iyengar yoga classes reported that the individual level instruction and use of props in the classes allowed them to perform the exercise and thus increased their confidence. These preliminary data suggest that future interventions should consider strategies to improve intention, confidence, and the perceived benefits of yoga to maximize adherence and potentially improve outcomes.
In our study, no participant demographic variables were significant predictors of adherence. Likewise, Pinto et al
] reported that demographic and medical variables were not significant predictors of adherence for breast cancer survivors participating in their home-based trial. A possible explanation for the lack of prediction from the demographic variables in our study is that our sample was a small, relatively homogenous sample of non-Hispanic Caucasian middle age women. In their larger multi-ethnic sample, Moadel et al
] reported that younger age predicted lower adherence. Moreover, a larger proportion of Hispanic women, as compared to African American and white women did not attend classes suggesting the potential importance of considering ethnicity. Future research with a larger more representative sample of breast cancer survivors is needed to substantiate our findings.
With respect to medical variables and health behaviors, stage of disease at diagnosis, diagnosis of a second cancer, and yoga experience in the past year significantly predicted adherence. Participants diagnosed with more advanced disease had greater adherence compared to those diagnosed with earlier stage disease. Concomitantly, participants reporting diagnosis of a second cancer attended significantly more classes than participants with a sole diagnosis of breast cancer. A possible explanation is that participants diagnosed with more advanced disease and or another type of cancer may be more concerned about their disease and overall health and therefore more motivated to adhere to methods to improve their health. In other studies, the role of disease stage has been mixed. Irwin et al
] reported a positive association between disease stage and exercise levels, and likewise in the START trial, Courneya et al
] found that breast cancer patients with more advanced disease had better adherence. In contrast, Hong et al
] reported a negative association, and Milne et al
] reported no association.
In our study, treatment regime was not predictive of adherence. In contrast, in their randomized controlled trial of yoga in breast cancer patients, Moadel et al
] reported significant associations between radiation therapy, non anti-estrogen therapy, and lower adherence. A possible reason for the difference in our findings is that a proportion of participants in the Moadel et al
] study were receiving radiation at the time of the study, and this may have made adherence more challenging than for participants post treatment. The small number of women receiving anti-estrogen therapy in our study (n
=10) may have obscured any meaningful findings in relation to adherence and warrants further investigation with a larger sample.
It is well established that past exercise is one of the best predictors of future exercise,[40
] and therefore it seems intuitive that participants who had previous yoga experience would adhere better to the yoga program. In support of this, we found that participants who reported participating in yoga within the last year attended significantly more classes than those who reported no prior yoga experience. A possible explanation for this finding is that previous experience with yoga led to greater self-efficacy that they would be able to complete the exercises in the yoga classes. The possible mediating effect of self-efficacy on adherence warrants investigation in future research.
With respect to participant-reported outcomes, less fatigue and greater happiness significantly predicted greater adherence. Similarly, Moadel et al
] reported that lower adherence was associated with increased fatigue. Danhauer et al
] reported that participants with greater health related QoL and lower fatigue attended more classes. It is apparent that breast cancer survivors reporting higher levels of fatigue are at risk for lower levels of adherence to exercise programs and therefore are important targets for intervention.
The overall strengths and limitations of our trial have been reported elsewhere.[15
] Briefly, limitations include the small, homogenous convenience sample, and the use of multiple comparisons. Because of the small sample size, it is not clear whether some of the findings were truly non significant or due to inadequate power to detect a difference. Conversely, the issue of multiple comparisons is another limitation; it is possible that some of the significant findings were due to chance. In this study, participants were already enrolled in the yoga class and were therefore a selected group that could limit the generalizability of the findings. Moreover, limited variability in several of the theory of planned behavior variables at baseline is likely attributable to the selected and highly motivated sample of participants already participating in the Iyengar yoga program. We focused on predictors of adherence to supervised Iyengar yoga among women post treatment, and it is likely that predictors of adherence will be different for participants concurrently undergoing treatment for breast cancer. Likewise, we were not able to examine other potential predictors of adherence to the IY classes that may have been important such as transportation, distance to the facility, and work and/or family obligations.
Several important strengths warrant discussion. To the best of our knowledge, no study to date has conducted a comprehensive examination of the medical, demographic, and social cognitive correlates of adherence to an Iyengar yoga program in this population. The findings suggest that several motivational variables are important predictors of yoga adherence with this population providing support for the model in future research. The theory of planned behaviour appears to be useful for examining predictors of adherence to a yoga program in this population and appears to be an appropriate theoretical model for future research and interventions.