Tai Chi practice and change in exercise habits in survivors of acute coronary syndromes: is mindfulness a possible mechanism?

Abstract

Background.

Mechanisms by which mind-body practice may improve exercise behaviors in patients with coronary heart disease (CHD) are unknown.

Purpose.

To identify predictors of exercise behaviors developed in response to a tai chi intervention.

Methods.

We conducted an exploratory analysis using data from a pilot study of tai chi for physically inactive patients with CHD. We used Latent Class Analysis to identify patterns of moderate to vigorous physical activity (MVPA) (accelerometry-assessed, average min/week) over the 9-month study period. Logistic regression was used to evaluate associations between MVPA patterns and changes in exercise self-efficacy (Self-efficacy for exercise scale), social support (Multidimensional Scale of Perceived Social Support), resilience (Brief Resilience Survey), and mindfulness (Five Facets of Mindfulness) at 6 months.

Results.

Data supported a 3-class model over time: 31% of participants were in class 1 (high and sustained MVPA, > 200 min/week), 21% were in class 2 (low and decreasing MVPA, < 100 > 50 min/week), and 48% in class 3 (low and stable MVPA, < 50 min/week). Higher mindfulness scores at 6 months were associated with a high and sustained MVPA pattern (b=1.89, SE=0.15, p=.01). Social support was inversely associated with low and decreasing MVPA (b = −.72, SE = .34, p = .04). No associations were noted with the other factors.

Conclusions.

Improvements in mindfulness were associated with a distinct high and sustained MVPA pattern. Although only hypothesis-generating, these findings suggest that improvements in mindfulness skills could be a mechanism for the possible effect of Tai Chi in the promotion of MVPA.

Introduction

Physical activity (PA) is a key target for the secondary prevention of coronary heart disease (CHD), yet only a minority of patients with CHD meets current PA recommendations.¹ Referral to cardiac rehabilitation programs has been used to promote PA in these patients, but unfortunately less than 30% of eligible patients enroll in rehabilitation programs and attrition is high.² This has prompted initiatives to promote the study of alternative options to improve exercise habits in this population.² One such option is tai chi, a Chinese martial art integrating slow, gentle body movements with the cultivation of. awareness of the breath and body movements. Despite its gentle approach, energy expenditures during tai chi can reach levels consistent with moderate-intensity aerobic activity and practice has been associated with improvements in functional capacity.³

Two studies, including our own,^4,5 have shown that tai chi can improve physical activity. Specifically, our study showed for the first time that a 6-month tai chi program can improve objectively assessed (accelerometry) moderate to vigorous PA (MVPA) in physically inactive CHD patients.⁵ However, the mechanisms by which mind-body exercise may improve exercise behaviors in patients with CHD are unknown. For the present analysis we were interested in studying the possible mechanisms underlying the effect of tai chi training on MVPA. We used Latent Class Modeling, a novel data-driven approach, to identify patterns of MVPA developed in response to tai chi training and to examine factors that may affect these patterns.

Methods

This exploratory analysis was conducted using data from the Gentle Cardiac Rehab Study (NCT02165254), a pilot clinical trial designed to examine the feasibility and possible effects of tai chi on PA among patients with CHD. .⁵ A detailed description of the parent study has been provided elsewhere.⁶ Briefly, participants were recruited via flyers placed in public venues, online resources, and advertisements on local media outlets. Inclusion criteria were a history of CHD, being able to understand and speak English, age > 21, and being physically inactive. Participants were excluded if they were cognitively impaired, reported severe depressive symptoms, were practicing tai chi, had contraindications to PA or if they were enrolled in cardiac rehabilitation.

Participants were randomized (via a computer-generated schedule) to a “LITE” (n=13; 2 times/week for 12 weeks) or to a “PLUS” tai chi intervention (n=16; 3 times/week for 12 weeks, 2 times/week for 13–16 weeks, 1 time/week for 17–24 weeks). Classes (1-hour duration) were offered in a dedicated room at the rehabilitation center at our institution and delivered by certified tai chi instructors. The type and sequence of tai chi poses was identical in the two groups and involved a series of Yang style exercises adapted for our study population. All participants were instructed to practice tai chi at home at least three times per week with the guidance of a 30-minute video. All participants provided informed consent and the study protocol was approved by the Institutional Review Board at our institution.

Measures

Assessments were conducted at baseline, 3-, 6, and 9 months since baseline. Exercise self-efficacy was measured using the self-efficacy for exercise scale, a measure of the patients’ confidence in their ability to exercise (0=not confident to 10=very confident),⁷ with scores ranging from 0 to 90 and higher scores indicating higher exercise self-efficacy. Social support was evaluated using the Multidimensional Scale of Perceived Social Support,⁸ a 12-item tool to measure how one perceives their social support system. Resilience was measured using the Brief Resilience Survey, an instrument validated in cardiac patients with scores ranging from 1 (low resilience) to 5 (high resilience).⁹ Mindfulness skills were assessed with the Five Facets of Mindfulness questionnaire (short form) a 15-item questionnaire that measures five identified components of mindfulness practice: observing, acting with awareness, non-judging of inner experience, and non-reactivity to inner experience. Higher scores indicate higher mindfulness skills.¹⁰ PA was measured using accelerometers (Actigraph, LLC, Fort Walton Beach, FL model wGT3X-BT, worn during non-class time). Data were processed using ActiLife software with minimum wear time set at 10-hours/day and non-wear time set at 60 minutes.

Descriptive measures included socio-demographic characteristics (self-reported using standard surveys), medical history (abstracted from medical records); body mass index (BMI - weight (Kg)/height (meters) squared) calculated from physical measurements collected at the baseline visit; and depression (measured with the Hospital Anxiety and Depression scale,¹¹ a self-administered questionnaire with two sub-scales measuring anxiety and depression and higher scores (range: 0–21) indicating greater psychological morbidity.

Analysis.

A hybrid data driven approach was used for this analysis. Latent Class Analysis (LCA), a data-driven approach, was used to identify patterns (classes) of MVPA over the study period (9 months). This technique reduces participant level data from a vector of up to 9 months of data to a single class, corresponding to their most likely pattern of PA behavior. As the number of classes in LCA is assumed to be fixed, we tested a series of models ranging from 2 to 4 classes (with optimal model fit defined by the minimum Bayesian Information Criteria (BIC)). In this case, the optimal solution was a 3-class model (significant model fit and significantly lowest BIC value). Differences in baseline characteristics by MVPA class were assessed using ANOVA, chi-squared tests, and non-parametric tests as appropriate. Finally, a series of multivariate logistic regression models were used to test whether changes in mindfulness, exercise self-efficacy, social support, and resilience (baseline to 6 months) were associated with MVPA class at 9 months. Measurement of the mediators at 6 months to predict outcomes over 9 months was chosen to ensure the assumption of temporal precedence was met. Models adjusted for baseline differences in MVPA and intervention allocation in the parent study. Analyses were carried out in SAS 9.3 with significance level set at alpha = .05 apriori. Unstandardized regression coefficients and standard errors are reported, allowing effects to be interpreted in the scale of the original data.

Results

Participants (n=29) were on average 67.9 years old (SD=10.3). The majority was male (63%), white (93%), and reported at least some college education (75%). Almost 50% of the study participants were diabetic, 70% had high cholesterol levels or hypertension, more than 50% were obese, and about 30% were still active smokers.

Baseline differences by MVPA class

Data supported a 3-class model of MVPA (average minutes/week) over time. Overall, 31% of participants had high and sustained MVPA over time (class 1, >200 min/of MVPA/week), 21% had low and decreasing MVPA, (class 2, <100 >50 min/week), and 48% had low and stable MVPA (class 3, <50 min/week) (Figure 1).

Figure 1. Patterns of MVPA over Time.

Pattern of MVPA (Moderate to vigorous physical activity) over 9-month time period. (Legend: Class 1: high and sustained; Class 2: Low and Decreasing; Class 3: low and stable)

As shown in Table 1, there were significant differences in marital status and education by MVPA class, such that those with high and sustained MVPA over time (class 1) were more likely to be in a relationship vs. those with low MVPA (classes 2 and 3) (p <.05). Class 3 participants were less likely to have at least some college level education (p <.05). There were no other significant baseline differences across classes.

Table 1:

Baseline characteristics by pattern of MVPA

	Class 1a	Class 2	Class 3
Age	65.22 (10.73)	66.67 (7.47)	70.14 (11.10)
Marital Status*
Single/Widowed/Divorced	22.2%	50%	64.3%
Partnered	77.8%	33.3%	35.7%
Separated	0	16.7%	0
At least some college education*	77.8%	100%	42.9%
Retired	66.7%	66.7%	64.3%
Female	12.5%	16.7%	41.7%
Weight	80.42 (9.63)	85.36 (11.48)	94.55 (32.27)
FFM (mean (SD))	54.33 (8.54)	55.17 (3.97)	54.29 (6.98)
SEE (mean (SD)	61.33 (14.59)	59.67 (16.06)	63.14 (18.90)
Social Support (mean (SD)	5.73 (0.73)	5.07 (0.92)	5.95 (1.29)
BRS (mean (SD)	3.78 (0.78)	3.58 (0.71)	3.40 (0.52)
HADS Depression subscale (mean (SD)	4.22 (3.42)	4.83 (3.60)	4.64 (3.81)
Smoking	11.1%	50.0%	28.6%
Diabetes	33.3%	33.3%	64.3%
High Cholesterol	66.7%	66.7%	85.7%

^*p < .05 for between-class difference.

^aClass 1 = high and sustained; Class 2 = low and decreasing; Class 3 = low and stable

FFM = Five Facets of Mindfulness questionnaire

SEE = Self-efficacy for Exercise scale

BRS = Brief Resilience Survey Scores

HADS = Hospital Anxiety and Depression scale

Associations between psychosocial factors and MVPA patterns

We found an association between changes in mindfulness scores and class, such that significantly greater improvements in mindfulness skills at 6 months (b=1.89, SE=0.15, p=.01) were seen participants with high and sustained MVPA over time (class 1), controlling for baseline differences in MVPA and intervention allocation in the parent study. Resilience, exercise self-efficacy, and social support were not associated with MVPA patterns. Participants with low and decreasing MVPA (class 2) had significantly lower social support scores (total score: b = −.72, SE=.34, p =.04; family scores: b=−1.26, SE=.46, p=.01; friends scores: b=−.94, SE=.32, p=.01) compared to class 3 participants.

Discussion

We used a novel data-driven analytical approach to explore whether changes in resilience, exercise self-efficacy, perceived social support, and mindfulness skills may play a role in modifying PA patterns in a group of physically inactive patients with CHD s who received tai chi training. We found that improvements in mindfulness skills were significantly associated with high and sustained MVPA patterns. While these findings are only hypothesis-generating given the small sample size, our explanation for these results, to be tested in future larger studies, is that improvements in mindfulness skills could be a key mechanism by which tai chi may improve PA.

Although the literature on this topic is still in its infancy, several studies suggest that our explanation may be plausible. In a small qualitative report of tai chi exercise for patients with chronic heart failure participants enrolled in tai chi reported increased mindfulness/self-awareness and decreased stress reactivity compared to controls.¹² Preliminary studies also suggest that interventions designed to develop and cultivate mindfulness (i.e., “the ability to “pay attention on purpose, in the present moment, and nonjudgmentally”), may buffer seasonal decreases in PA in both adult and adolescent samples.^13,14 Although tai chi training does not involve “formal” mindfulness meditation training, awareness of breath and body movements is encouraged throughout the training. These practices are also key components of mindfulness training and may explain why increased mindfulness skills may develop in response to tai chi training.¹⁵

In this study, social support, resilience, and exercise self-efficacy were not associated with PA patterns. While a former RCT of tai chi for heart failure has reported improved cardiac self-efficacy in tai chi compared to controls,¹⁶ our study sample was different in that we enrolled a different population and we also used a different measure of exercise self-efficacy. It is also possible that self-efficacy may play different roles in different populations.

Strengths of this analysis include the use of a data-driven analytical approach and of objective measurements of PA. Several limitations should also be noted, including the small sample size, the post-hoc nature of this analysis, the use of a self-reported measure of mindfulness skills. While we acknowledge such limitations, these hypothesis-generating findings (to be replicated in a larger studies, different populations, and different types of mind-body exercise) suggest that promoting mindfulness skills via mind-body exercise could produce the greatest improvements in PA and, possibly, cardiovascular outcomes in this population.

Highlights.

• Tai Chi (TC) training may improve physical activity (PA) but mechanisms are unknown

• We used a data-driven approach to identify patterns and predictors of PA after TC

• Mindfulness scores at the end of training predicted higher PA at follow-up

• Improvements in mindfulness skills could be a key mechanism by which TC promotes PA

• These preliminary findings need to be confirmed in larger studies