Mindfulness is Associated with Treatment Response from Non- Pharmacological Exercise Interventions in Knee Osteoarthritis

Abstract

Objective

To examine the association between baseline mindfulness and response from exercise interventions in knee osteoarthritis (OA).

Design

Cohort Study; Responder analysis of a clinical trial subset

Setting

Urban tertiary care academic hospital

Participants

86 participants with symptomatic, radiographic knee OA (mean age, 60 years; 74% female; 48% white)

Interventions

12 weeks (twice per week) of Tai Chi or Physical Therapy exercise

Main Outcome Measure(s)

Treatment response was defined using Osteoarthritis Research Society International criteria indicating meaningful improvements in the Western Ontario and McMaster Osteoarthritis Index (WOMAC) pain, function, or Patient Global Assessment scores. At baseline, participants completed the Five Facet Mindfulness Questionnaire (FFMQ; mean total score ±SD: 142±17) and were grouped into three categories of total mindfulness: higher, medium, or lower. Relative risk (RR) ratios were used to compare treatment response across groups.

Results

Participants with higher total mindfulness appeared to be 38% (95% CI: 1.05, 1.83) more likely to meet responder criteria than those with lower mindfulness. We found no significant difference between medium and lower mindfulness groups (RR=1.0; 95% CI: 0.69, 1.44). Among the 5 mindfulness facets, medium Acting-with–Awareness was 46% (95% CI: 1.09, 1.96) more likely to respond than lower Acting-with–Awareness, and higher Acting-with–Awareness was 34% more likely to respond, but this did not reach significance (95% CI: 0.97, 1.86).

Conclusions

In this study, higher mindfulness, primarily driven by its Acting-with-Awareness facet, was significantly associated with a greater likelihood of response to non-pharmacological exercise interventions in knee OA. This suggests that mindfulness-cultivating interventions may increase the likelihood of response from exercise.

Knee osteoarthritis (OA) affects more than 20 million Americans and is a leading cause of physical disability and chronic pain^1,2. Current clinical guidelines emphasize rehabilitative exercise treatment as the foundation of conservative symptom management³. However, the manifestations of knee OA and response to treatment varies^4–7. This heterogeneity of clinical presentation and treatment response confounds interpretation of therapeutic clinical trials and complicates the care decisions for clinicians^4,5. Thus, identifying clinical factors to define patient profiles most likely to benefit from treatment would help therapeutic decision-making^6,8. Prior studies have explored several potential physical and biomechanical predictive factors of non-pharmacological treatment response, including muscle strength, inflammation, and tibiofemoral joint alignment⁹. However, there are considerably fewer studies on the psychological predictors of response.

Psychological health has a well-established role in perpetuating the cycle of chronic pain¹⁰. Because exercise interventions require active patient participation, psychological factors may influence treatment efficacy through its behavioral effect on adherence^11,12. One prior study found that absence of depressive symptoms predicted symptomatic response from exercise rehabilitation in OA⁷. While this study examined traditional (i.e. depression and anxiety) and general (e.g. mental health) constructs of psychological health, there is growing interest in the beneficial role of mindfulness, a relatively newer psychological construct used in biomedical research.

Mindfulness is a psychological state achieved by focusing one’s awareness on the present moment, while acknowledging and non-judgmentally accepting one’s feelings, thoughts, and bodily sensations¹³. It is also a dispositional characteristic or trait that varies from person to person and has an approximately normal distribution among the general population¹⁴. Rapidly accumulating evidence shows that mindfulness interventions enhance physical and psychological health among patients with chronic pain^15–17. While these studies have primarily focused on the efficacy of mindfulness treatments, the evidence that mindfulness, as a characteristic, may influence treatment outcomes is scarce.

Research on the characteristic of mindfulness can examine the role of this trait on a variety of commonly-used treatments. For exercise interventions in chronic pain, mindfulness may be associated with improved outcomes because of its link with better pain coping skills^18,19 and greater treatment adherence^20,21. Among the general population, mindfulness is associated with greater physical activity and greater weight loss from exercise^21–24. Investigating the association between mindfulness and exercise treatment outcomes is warranted because poor adherence and low pain-coping skills are both major barriers to exercise in OA^11,12. To date, no study has examined longitudinal associations of mindfulness on treatment outcomes among these patients.

The purpose of this study was to examine the association between baseline mindfulness and treatment response from Tai Chi mind-body exercise or exercise-based Physical Therapy among adults with symptomatic, radiographic knee OA. The hypothesis is that greater mindfulness would be associated with increased treatment response.

Methods

Study Design

This study was a secondary analysis of a 52-week National Institutes of Health-funded, single-blind, randomized trial comparing Tai Chi and Physical Therapy among adults meeting the American College of Rheumatology (ACR) Criteria for symptomatic knee OA. Both interventions had similar response rates and pain/function/global health improvements. It was decided to combine data from participants receiving either intervention. A detailed description of the methodology and primary study results for this trial has been previously published^25,26.

The trial was conducted at a University Medical Center in Boston, Massachusetts, between October 2010 and September 2014. Participants were recruited through multimodal strategies, including print advertisements, online media, a booth at a senior exposition, and a clinical patient database. We obtained informed consent before baseline assessments for eligibility. Study inclusion criteria included: 1)Age ≥40 years, 2)Fulfillment of the ACR criteria for knee OA, and 3)Western Ontario and McMaster Osteoarthritis Index (WOMAC) pain score ≥40. Exclusion criteria included: 1)Prior experience with complementary medicine or Physical Therapy programs for knee OA within the past year, 2)Severe medical limitations precluding full participation, 3)Intra-articular steroid injections or surgery in the past three months, 4)Intra-articular hyaluronic acid injections in the past six months, 5)Mini-Mental Status examination score <24, or 6)inability to walk without an assistive device (Figure 1). Participants were randomly assigned in a 1:1 ratio to Tai Chi or Physical Therapy. The physical function assessors were blinded to the treatment assignments during evaluations and did not have access to the data until data collection was complete. All participants signed an informed consent form before enrollment, and the study was approved by the Institutional Review Board. This study analyzed participants who had mindfulness data at baseline. The Five Facet Mindfulness Questionnaire (FFMQ) was formally implemented in the study after subject recruitment and data collection had already begun. After the FFMQ was included in the parent trial, all subsequent participants had the opportunity to complete the FFMQ at baseline. Therefore, this investigation only reports data from participants who completed the FFMQ at baseline, which was a fixed available sample. Of 204 participants from the parent trial, 86 completed the FFMQ at baseline and, therefore, were analyzed. Seventy-six participants were assessed after 12-week intervention; 10 were lost to follow-up (Figure 2).

Figure 1.

Inclusion/Exclusion Criteria

Figure 2. Sub-study Flow Diagram.

FFMQ = Five Facet Mindfulness Questionnaire.

Interventions

Tai Chi is a traditional Chinese mind-body exercise characterized by lower intensity physical activities, including slow, flowing movements that incorporate muscle strengthening, balance, postural training, meditation, and deep breath training. Accumulating evidence supports Tai Chi as an effective therapeutic exercise option for patients with chronic disease^27,28. In this study, each Tai Chi class lasted 60 minutes and occurred twice weekly over 12 weeks. One of 3 Tai Chi instructors (20 combined years of experience) administered a standardized 10-form Yang style Tai Chi protocol for knee OA at each class²⁵. Classes typically involved an initial warm-up and a review of Tai Chi principles and movement, breathing techniques, and relaxation methods.

The Physical Therapy intervention followed national guidelines for knee OA treatment²⁹, and consisted of joint mobilizations, lower limb stretching, active and passive range-of-motion exercises, and strengthening, aerobic, and balance exercises. One of 3 licensed physical therapists (over 20 combined years of experience) offered two 30-minute one-to-one sessions per week for the first 6 weeks. For the second 6 weeks, participants completed four 30-minute sessions per week at home with weekly telephone monitoring. After an initial 1-hour musculoskeletal examination, physical therapists tailored the Physical Therapy regimen toward specific individualized-participant treatment goals. Throughout both interventions, participants were asked to practice for at least 20 minutes daily at home.

Outcome Measures

The FFMQ(39-items; higher scores reflect greater mindfulness) measures total mindfulness and its facets^30,31. The 5 Facets: 1)Observing: noticing internal and external stimuli. 2)Describing: mentally labeling observed stimuli with words. 3)Acting-with-Awareness: attending to current actions, as opposed to behaving automatically/absentmindedly. 4)Non-judging: refraining from evaluation of one’s sensations/cognitions/emotions as negative or intolerable. 5)Non-reacting: allowing thoughts and feelings to come and go, without getting carried away by them.(Supplementary Table S1)

The WOMAC, v.3.1, has subscales measuring pain intensity (5 items, 0–500) and physical function (17 items, 0–1700) in knee osteoarthritis³². Higher scores indicate greater pain or poorer physical function, respectively. A 0–10 visual analog scale, the Patient Global Assessment (1-item; higher scores reflect greater disease activity) measures the patient’s perception of disease activity³³. The Short Form-36 (36-items; higher scores reflect greater health status) measures health-related quality of life³⁴. Scores were combined to obtain two aggregate scores, ranging from 0 to 100: the Physical Component Summary score and the Mental Component Summary score. The Six-minute Walk measures functional exercise capacity;³⁵ participants walk as far as possible within a six-minute period.

The Beck Depression Inventory (21-items; higher scores reflect greater depressive symptoms) measures the severity of depressive symptoms³⁶. The Perceived Stress Scale (10-items; higher scores reflect a greater stress) measures the degree to which situations over the past month are appraised as uncontrollable/overwhelming³⁷. The Arthritis Self-Efficacy Scale-8 (10-point Likert scale; higher scores indicate higher self-efficacy)³⁸ measures self-efficacy, the belief that one can successfully take action to produce a desired outcome. All measures used in this study have been validated.

Responder Criteria

Treatment response was defined according to the Outcome Measures in Rheumatology- Osteoarthritis Research Society International criteria for OA: 1. ≥50% improvement in pain or function and change of ≥20 points (0 to 100 scale) in WOMAC pain or function; or 2. ≥2 of the following: ≥20% improvement and change >10 points in WOMAC pain, ≥20% improvement and change >10 points in WOMAC function, or ≥20% improvement in Patient Global Assessment and change >10 points (0 to 100 scale)³⁹.

Statistical Analysis

Missing FFMQ items were imputed from the mean of the non-missing items of the respective facet⁴⁰. All data were examined visually and statistically for normality of distribution and presented as mean ±SD unless otherwise stated. Due to skewed distribution of scores and in the absence of established cut-points for the FFMQ³¹, participants were stratified into 3 tertiles based on baseline total score: higher, medium, and lower. Differences in characteristics among the three mindfulness groups were tested with analysis of variance or chi-square tests. Where necessary, Fisher’s exact tests compared categorical variables due to small sizes of some subgroups. Psychological variables (i.e. depression, stress, self-efficacy, and quality of life) were compared at baseline to illustrate previously reported associations between mindfulness and greater psychological health among the mindfulness tertiles ⁴¹. Significance level was p≤0.05. To examine whether this study sub-sample differed from the parent trial, baseline characteristics were compared using t-tests or chi-square tests.

To check for confounding, logistic regression was performed when baseline demographic variable or treatment response-determining variable (i.e. WOMAC Pain, Function, or Patient Global Assessment) differed significantly among tertiles. Any identified confounding factors were subsequently included in a multivariable logistic model. Because improvements in psychological health may be in the causal pathway, baseline psychological variables were not tested for confounding. It was planned to compare the likelihood of achieving responder criteria among the three mindfulness groups using chi-square tests and either risk ratios or odds ratios (95% CI), depending on whether confounding variables were found. Data were analyzed using SAS statistical software (Version 9.4).

Results

For missing FFMQ items at baseline, 5 people had 1 missing item and 1 person had 2 missing items. Participants were predominantly female (74%), well-educated (85%), and white (48%); mean age: 60 years; and body mass index (BMI): 33 kg/m² (Table 1). Demographic and clinical variables did not significantly differ between the participant sub-sample and the parent clinical trial (not shown but available from the authors). Significant differences for psychological characteristics were found as expected; wherein greater mindfulness had greater health. There were no significant differences among mindfulness groups for treatment assignment or demographics, except BMI, and no significant baseline differences were found among mindfulness groups for the response-determining variables (i.e. WOMAC Pain, Function, or Patient Global Assessment). Including BMI in a logistic regression model did not change the results. Therefore, no multivariable models were performed, and all results are reported as risk ratios.

Table 1.

Demographic and Clinical Characteristics of Participants by Mindfulness Levels (n = 86)

Variable		Lower Mindfulness (n = 23)	Medium Mindfulness (n = 26)	Higher Mindfulness (n = 27)	p-value
Age, years		63.1 (10.5)	57.4 (9.5)	60.5 (10.8)	0.11
BMI		29.9 (6.3)	33.0 (6.3)	35.3 (7.4)	0.01
Sex, n (%)	Female	19 (65.5%)	20 (71.4%)	25 (86.2%)	0.18
	Male	10 (34.5%)	8 (28.6%)	4 (13.8%)
Pain duration, years		11.2 (16.0)	8.2 (8.4)	11.5 (15.1)	0.61
Kellgren	0-1	2 (7.1%)	5 (17.9%)	1 (3.6%)	0.38
Lawrence	2	10 (35.7)	9 (32.1%)	12 (42.9%)
Grade†, n (%)	3	14 (50.0%)	11 (39.3%)	9 (32.1%)
	4	2 (7.1%)	3 (10.7%)	6 (21.4%)
Race, n (%)	White	11 (37.9%)	14 (50.0%)	16 (55.2%)	0.37
	Black	11 (37.9%)	8 (28.6%)	11 (37.9%)
	Other	7 (24.1%)	6 (21.4%)	2 (6.9%)
Education†, n (%)	High school	7 (24.1%)	3 (10.7%)	3 (10.3%)	0.38
	Some College	9 (31.0%)	11 (39.3%)	12 (41.4%)
	College Grad	5 (17.2%)	9 (32.1%)	4 (13.8%)
	Graduate School	8 (27.6%)	5 (17.9%)	10 (34.5%)
Intervention, n (%)	Tai Chi	17 (58.6%)	13 (46.4%)	16 (55.2%)	0.64
	Physical Therapy	12 (41.4%)	15 (53.6%)	13 (44.8%)
Total Mindfulness [min: max:]		124 (8.4)	141 (3.6)	161 (9.5)	<0.0001
FFMQ Score Range, 39–195
WOMAC Pain* Score Range, 0–500mm		289.0 (111.6)	253.7 (93.1)	253.8 (95.7)	0.31
WOMAC Function* Score Range, 0-1700mm		946.4 (387.2)	930.8 (335.1)	921.1 (360.6)	0.96
Patient Global Assessment* Score Range, 0–10cm		5.2 (2.2)	5.1 (2.0)	5.1 (2.2)	0.98
Perceived Stress Score Range, 0–40		17.03(6.72)	13.43(5.93)	8.97(5.17)	<0.0001
Self-Efficacy Score Range, 0-10		5.73(1.84)	5.85(2.21)	7.35(1.62)	0.002
6-Minute Walk Test meters		381.47(94.47)	399.90(94.75)	408.94(85.06)	0.52
SF MCS Score Range, 0–100		47.11(10.30)	54.43(6.55)	56.85(6.54)	<0.0001
SF PCS Score Range, 0–100		33.34(7.43)	35.60(10.11)	40.34(8.45)	0.01
Beck Depression Score Range, 0–63		12.76(11.28)	6.32(6.26)	3.71(4.21)	0.0001

All values are mean (SD), unless otherwise stated. FFMQ = Five Facet Mindfulness Questionnaire, higher scores = higher mindfulness; SF = Short Form-36; MCS = Mental Component Summary; PCS = Physical Component Summary.

^*Higher scores = more pain, worse function, or more global disease.

^†Fisher’s exact tests were used in place of chi-square tests due to small sample sizes in some sub-groups.

For those with higher mindfulness, 96% responded from Tai Chi or Physical Therapy (Table 2). However, for those with medium or lower mindfulness, only 69% responded. Table 3 summarizes the likelihood of response by total and facet mindfulness. Higher total mindfulness was 38% (95% CI:1.05, 1.83) more likely to respond than lower, but no significant difference was found between medium and lower total mindfulness (1.0; 95% CI:0.69, 1.44). Among the 5 facets, there was a greater magnitude of effect in the Acting-with-Awareness facet than in other facets. Medium Acting-with-Awareness was 46% (95% CI:1.09, 1.96) more likely to respond than lower. Participants with higher Acting-with-Awareness were 34% (95% CI:0.97, 1.86) more likely to respond than lower, which trended toward significance.

Table 2.

Distribution of Treatment Response by Total Mindfulness Levels

Mindfulness Tertile*	No Treatment Response	Treatment Response†	Row Total
Lower, N (%) [Score Min: 98; Max: 134]	7 (30.4%)	16 (69.6%)	23
Medium, N (%) [Score Min: 135; Max: 148]	8 (30.8%)	18 (69.2%)	26
Higher, N (%) [Score Min: 149; Max: 181]	1 (3.7%)	26 (96.3%)	27
Column Total	16	60	76

N = number of participants.

^*Mindfulness as measured by the total Five Facet Mindfulness Questionnaire score; Score Range, 39–195.

^†Treatment response as defined by the Outcome Measures in Rheumatology- Osteoarthritis Research Society International (OMERACT-OARSI) Responder Criteria.

Table 3.

Relative Risk of Treatment Response by Total and Facet Mindfulness Levels (n = 76)

FFMQ Variable	Mindfulness Tertiles	Relative Risk (95% Confidence Interval)	p-value (Chi-square)
Total Score Range, 39–195	Higher	1.38 (1.05, 1.83)	0.01
	Medium	1.00 (0.69, 1.44)	0.98
	Lower	Reference
Observing Score Range, 8–40	Higher	0.91 (0.69, 1.20)	0.51
	Medium	0.97 (0.74, 1.29)	0.85
	Lower	Reference
Describing Score Range, 8–40	Higher	1.04 (0.76, 1.41)	0.82
	Medium	1.12 (0.84, 1.49)	0.43
	Lower	Reference
Acting-with-Awareness Score Range, 8–40	Higher	1.34 (0.97, 1.86)	0.09
	Medium	1.46 (1.09, 1.96)	0.01
	Lower	Reference
Non-Judging Score Range, 8–40	Higher	1.18 (0.88, 1.58)	0.24
	Medium	1.10 (0.80, 1.51)	0.56
	Lower	Reference
Non-Reacting Score Range, 7–35	Higher	1.12 (0.85, 1.47)	0.42
	Medium	1.04 (0.77, 1.41)	0.78
	Lower	Reference

FFMQ = Five Facet Mindfulness Questionnaire.

Discussion

This study examined the association between baseline mindfulness and treatment response from non-pharmacological exercise interventions among adults with symptomatic knee OA. Participants with higher total mindfulness were 38% more likely to respond to Tai Chi exercise or exercise-based Physical Therapy. Among the 5 mindfulness facets, Acting-with –Awareness facet had the greatest magnitude of effect. Importantly, these results provide a preliminary suggestion that adults with higher mindfulness may represent a novel patient sub-group more likely to respond to these exercise treatments.

These results are consistent with a previous study wherein mindfulness predicted treatment response to exercise intervention for weight loss⁴², and extend prior observational research showing that mindfulness was associated with reduced stress and disability⁴³, lower pain catastrophizing⁴⁴, and increased physical activity²⁴. Overall, these results contribute to the evidence indicating that as a characteristic, mindfulness may influence better health outcomes from exercise. Although the rationale underlying this association is not known, it may be that higher mindfulness facilitates greater adherence^7,9,23.

Preliminary evidence suggests that cognitive mechanisms of mindfulness involve a better ability to extinguish unhealthy behavioral habits and emotionally adapt to stress⁴⁵. Indeed, psychological factors did not predict response from pharmacological interventions, which involve a relatively more passive amount of patient participation^9,46. However, the overall evidence supporting the link between mindfulness or other psychological factors and exercise response has been inconsistent, primarily due to heterogeneity in study design^7,9,47,48. As a complex construct, mindfulness must be disentangled from its associations with other psychosocial variables to better understand how it functions as a predictor.

Baseline mindfulness facets did not predict treatment response, except Acting-with-Awareness: medium levels were significant and higher levels trended toward significance. Importantly, Acting-with-Awareness had a relatively larger magnitude of effect than the other facets, which suggests it may be the most sensitive facet in predicting response. This may be because attending to current actions, as opposed to behaving absentmindedly, facilitates better exercise efficiency or pain coping. Notably, in a cross-sectional investigation, the Acting-with-Awareness facet attenuated the influence of pain on stress in knee OA⁴¹. Taken together, these preliminary findings indicate that total mindfulness and its Acting-with-Awareness facet appear most relevant to improved outcomes from exercise interventions in OA. However, further study is needed to distinguish the medical relevance of mindfulness facets⁴⁹, and the length of the FFMQ limits applicability to patient care settings. Thus, the development of short forms or computer-adaptive mindfulness instruments would enhance the feasibility of its assessment in routine care.

This study has implications for research in OA and exercise science, and responds to the call for further research to identify psychological factors that predict response in OA^4,8,9,50,51. Identifying factors that influence treatment response may enhance stratified-randomization strategies that attempt to balance comparator groups of exercise trials to reduce confounding^8,52. Current guidelines recommend researchers designing OA exercise trials collect psychosocial variables that may influence the effectiveness of intervention, including outcome expectations and depression⁵³. These findings provide initial support for mindfulness as a confounding variable of exercise trials in OA. If confirmed, mindfulness-cultivating interventions might be used as a preparative adjunct or combined with exercise to enhance the likelihood of treatment response. Discordant from initial hypothesis, however, participants with medium total mindfulness were not significantly more likely to respond to intervention than lower mindfulness. While it cannot be ruled out that sample size influenced this result, it is possible that the association between mindfulness and response is not linear. However, because medium Acting-with-Awareness was significant while higher Acting-with-Awareness was not, further evidence is necessary to clarify the shape of this association and whether it differs by total or facet mindfulness.

In the clinical setting, these results help to advance the understanding of the heterogeneity of patient response to exercise. This may inform clinicians who seek to understand why some patients do not respond, and subsequently consider alternative options. The dearth of evidence explaining this heterogeneity has impeded the use of individualized exercise prescriptions by physicians to promote greater physical activity among these patients^8,51,54,55. By expanding the evidence to clarify the heterogeneity of response, these findings help advance the integration of prescriptive exercise by physicians into routine medical care for OA⁵⁵.

Limitations

This study is a secondary analysis of clinical trial data, and as such the results require confirmation in a further study primarily designed to examine mindfulness as a predictor of response. The sample was not representative of all persons with knee OA because they were symptomatic and knowingly participated in a trial. Although this may influence the study generalizability, the participants had characteristics of patients seen routinely in clinical settings. Furthermore, while the mindfulness level of this sample was skewed towards higher mindfulness, they were comparable to those of highly educated samples³⁰, which is concordant with the educational background of these participants. The second limitation of this study was that the sample size was not large. Some of the non-significant associations could be explained by a lack of power. However, lacking optimal statistical power does not diminish the importance of the significant findings. Namely, that higher total mindfulness was associated with higher response, and that Acting-with-Awareness had the greatest magnitudes of effect among facets. Thirdly, because the parent trial did not include a control group, this study could not disentangle the contributions of placebo or ‘regression-to-the-mean’, which occur in any treatment intervention. Nevertheless, based on recent platinum-quality evidence showing that various exercise-based therapies directly cause benefits in pain and function in OA⁵⁶, the implicit assumption that treatment response was attributable to the interventions was reasonable. Furthermore, several steps were taken to reduce bias, including the validated, widely-accepted definition of response, the radiographic, symptomatic confirmation of diagnosis, the representative characteristics of the subsample with the parent trial, and the proportionally balanced, randomized intervention assignments among the mindfulness groups. Fourthly, due to sample size, this study did not compare the association magnitudes according to treatment type. As a commonly-cited underlying mechanism of Tai Chi⁵⁷, mindfulness may have a stronger association with response in Tai Chi than Physical Therapy. However, the parent trial showed that the response rates did not statistically or meaningfully differ by treatment²⁶. If mindfulness therefore had a stronger association with response in Tai Chi, this difference must have been equally counter-balanced by a different factor in Physical Therapy, which is unlikely. Further study in a larger trial is warranted. Finally, the use of tertiles to categorize participants according to a continuous variable can reduce sampling power and is limited as a primarily data-driven approach. It is therefore important to empirically identify clinically relevant cut-points for the FFMQ that can be used to further examine the prognostic value of mindfulness levels among those with chronic diseases.

Despite limitations of this study, the main implication remains that mindfulness may be useful to better understand patient heterogeneity of response from exercise in knee OA. This is the first study known to empirically show this in OA or other chronic musculoskeletal pain conditions. Future studies involving a clinical trial design, larger samples of participants and different types of exercise interventions are needed. Additional exploration of causal pathways underlying this association may further inform how mindfulness can be best examined in research and applied to clinical settings.

Conclusions

In conclusion, higher baseline mindfulness was associated with treatment response from Tai Chi or Physical Therapy exercise interventions among adults with symptomatic knee OA. In addition, Acting-with-Awareness may be the facet driving this association. This investigation underscores the role of psychological health in the heterogeneity of treatment response in OA, and further studies in this area may guide efficient clinical decision-making for non-pharmacological therapy of patients with knee OA.

Supplementary Material

supplement

Table S1. Items of the Five Facet Mindfulness Questionnaire (FFMQ)^*

Highlights.

• Higher baseline mindfulness was associated with treatment response in knee osteoarthritis.

• This finding was primarily driven by the Acting-with-Awareness mindfulness facet.

• Mindfulness may be a prognostic indicator of treatment response among these patients.

• Mindfulness-cultivating interventions might be help as a preparatory adjunct to exercise.

List of Abbreviations

ACR

American College of Rheumatology

BMI

Body Mass Index

FFMQ

Five Facet Mindfulness Questionnaire

OA

Osteoarthritis

WOMAC

Western Ontario and McMaster Osteoarthritis Index