Physical Activity and Functioning Following Tai Chi and a Wellness Comparison Intervention in Veterans with Gulf War Illness

Abstract

Gulf War Illness (GWI) is a highly prevalent chronic, multisymptomatic condition associated with worsening physical health and quality of life among veterans. Specifically, GWI symptoms may disrupt physical functioning that negatively impacts overall health. Complementary and integrative health interventions that promote physical activity, such as Tai Chi, may be critical to addressing physical functioning in veterans with GWI. The present study is a secondary analysis of data from a randomized controlled trial comparing Tai Chi to a Wellness intervention on physical functioning outcomes in a sample of 53 veterans with GWI and pain. Gait speed, fast walking speed, lower body strength, and self-reported physical activity were assessed across four timepoints: baseline, posttreatment, 3-month follow-up, and 9-month follow-up. Treatments were delivered in a 60-minute group format twice a week for 12 weeks (24 sessions total). Results from multilevel regression models revealed that Tai Chi was related to greater increases in time spent on flexibility exercise whereas improvements in gait speed were associated with Wellness, though between-treatment differences were small. These findings are promising because they provide preliminary evidence for the utility of two treatments that may improve specific physical functioning outcomes in veterans with GWI, especially in accommodating physical difficulties associated with the condition that few treatments have addressed.

Introduction

Gulf War Illness (GWI) is a chronic, multisymptomatic condition that impacts nearly a third of the 700,000 US military veterans deployed to the 1991 Persian Gulf War.^1,2 This disorder of unknown etiology is characterized by distressing symptoms, including fatigue, musculoskeletal pain, cognitive and memory problems, sleep difficulties, skin abnormalities, gastrointestinal challenges, mood concerns, and respiratory complaints.^3,4 GWI is particularly debilitating because symptoms persist or even worsen over time.^5–7 Furthermore, greater GWI severity is related to substantially worse physical and mental health-related quality of life.⁸

GWI symptoms often disrupt the ability to engage in physical activity, which could contribute to the condition’s negative impact on quality of life.⁹ Veterans with GWI experience greater levels of pain, fatigue, and malaise following physical exertion relative to veterans without GWI.¹⁰ As a result, many veterans with GWI demonstrate low rates of exercise and these rates continue to decline even after engagement in exercise-promoting interventions.¹¹ Waning exercise in this population is problematic given the many health benefits associated with physical activity, including the potential for improvements in such GWI symptoms as cognitive functioning, sleep quality, and mood.^12–14

Complementary and integrative health (CIH) approaches that promote physical activity, while accommodating the physical and psychological difficulties of GWI, could be critical to addressing this condition in veterans.^4,9,15 The Veterans Affairs’ (VA) Whole Health initiative may provide routes for helping veterans enhance their physical functioning as it offers movement- and education-based CIH interventions that promote individually tailored physical activity goals and/or incorporate exercise marked by gentle body movements and/or stretching.^16,17 Indeed, these interventions (e.g., yoga, wellness coaching) are related to improvements in physical functioning, including greater exercise capacity and health goal attainment.^18,19 In addition, CIH interventions (e.g., mindfulness-based treatments, yoga) are efficacious at reducing GWI symptoms.^18,20,21 Thus, CIH approaches that promote physical activity may be viable options for improving physical functioning in veterans with GWI.¹⁵

Tai Chi is a whole-person, mind-body practice with roots in ancient China involving slow and gentle body movements paired with mindful breathing and awareness.²² It is one of eight CIH approaches currently approved by the VA.²³ Tai Chi is theorized to hold promise in improving physical functioning in veterans with GWI due to its ability to robustly target symptoms characteristic of this illness.²⁴ This CIH approach is also associated with benefits (e.g., greater self-efficacy, functional mobility, well-being) for chronic conditions similar to GWI marked by co-occurring physical and psychological problems, such as fibromyalgia and chronic pain.²⁵ Moreover, Tai Chi may enhance physical functioning by bolstering exercise capacity, mobility, and strength, as well as reducing disability and stiffness.²⁶ Given Tai Chi’s efficacy in treating chronic conditions with GWI symptom overlap, this intervention could be particularly useful for enhancing physical health in veterans with GWI.^27,28

The present study evaluated physical functioning outcomes of Tai Chi relative to a Wellness intervention comparison condition in a sample of veterans with GWI. This secondary analysis of data from a randomized controlled trial (RCT) compared these two group interventions on measures of physical activity, gait speed, fast walking speed, and lower body strength. We hypothesized that Tai Chi will be associated with greater improvements over time on physical functioning outcomes compared to Wellness. Findings inform the utilization of CIH approaches in the VA healthcare system where these interventions are increasing in popularity and GWI is highly prevalent.^29,30

Methods

Participants

Full details on participant recruitment and eligibility criteria are reported elsewhere.^31,32 Veterans were recruited for a single-blind RCT (Date of Trial Registration: 1/14/2016) via flyers, therapy referrals, and direct telephone contact. Inclusion criteria were: (1) served in the 1991 Persian Gulf War (August 1990 to July 1991); (2) met criteria for GWI as defined as one or more symptom from at least two of the following clusters: musculoskeletal pain, fatigue, and mood/cognition; (3) one symptom cluster included experiencing pain (musculoskeletal, joint, or stiffness) for at least 6 months; (4) lived in the same general location for the next 3 months (e.g., was not planning to move to a different state); (5) spoke and comprehended English; and (6) was able to attend pre-scheduled group treatment appointments.

Exclusion criteria were: (1) did not possess capacity for informed consent; (2) was experiencing a major medical, psychiatric, or neurological condition or moderate or severe traumatic brain injury that would impede safe engagement in study procedures; (3) changed psychotropic or pain medication within the past month; (4) was currently practicing Tai Chi, yoga, or mindfulness mediation in a routine and/or formalized manner (i.e., at least 3 hours a week for more than 3 months); (5) had difficulty standing on feet for approximately 60 minutes (the typical duration of a Tai Chi session during this study); (6) exhibited disruptive, disrespectful, or threatening behavior toward staff and/or other participants; and (7) concurrent participation in a different study investigating treatment for GWI or pain.

The total sample of the parent RCT was 53 veterans and all participants were retained in this secondary analysis. The sample was 88.7% (n = 47) male and 11.3% (n = 6) female, and the mean age was 54.89 years (SD = 8.04). The sample was 66.0% (n = 35) White, 22.6% (n = 12) Black, 3.8% (n = 2) Other, and 7.5% (n = 4) reported multiple races. There were no significant differences between treatment conditions in sex, age, or race.³² Regarding attrition rates, 88.7% (n = 47) of the sample attended at least one treatment session and 52.8% (n = 28) attended most (at least 75%) of the sessions.³² Further details about the demographics and attrition of the sample are reported in Niles et al.³²

Measures

Physical Activity Recall (PAR)^33,34

The PAR is a semi-structured interview that indexes time in hours devoted to physical activity over the past 7 days. Participants were asked to recall their physical activity across occupational, household, and sporting domains during the morning, afternoon, and evening. Participants reported the time engaged in each activity and classified their perceived physical exertion as moderate, hard, or very hard based on the examples provided in the instructions.³⁴ The PAR is a reliable and valid measure of physical activity.^35–37 Because few participants in the current study rated perceived exertion at the hard or very hard levels, we decided to collapse together moderate, hard, and very hard physical exertion to create an outcome variable for time devoted to exercising with at least moderate effort (PAR-Exercise).³⁸ We also calculated outcome variables for time engaged in flexibility exercises (PAR-Flexibility) and time spent on strength training (PAR-Strength) to collect specific data on time devoted to different types of health-promoting activities, especially because Tai Chi is a practice that cultivates flexibility.

Short Physical Performance Battery (SPPB)³⁹

The SPPB is a multifaceted assessment battery that evaluates lower-body physical performance. It encompasses three subtests: 4-Meter Walk Test (Walk 4M), Chair Stand Test (CST), and Standing Balance Test (SBT). The Walk 4M is a measure of gait speed in which study staff timed how many seconds it took participants to walk 4 meters. Participants were instructed to walk “at [their] normal walking pace.” Each participant completed two trials of the Walk 4M. The times from these two trials were then averaged to calculate participants’ final gait speed score, measured in seconds. The CST is a timed evaluation of the strength and power of lower extremities. For the CST, after displaying proficiency in standing once, participants were instructed to complete five chair stands from a sitting position with their arms folded across their chests as fast as possible while study staff timed them. Participants were scored based on how long it took them in seconds to complete the task with lower scores indicating faster completion time. The SBT is a measure of static balance in which participants are asked to maintain balance for up to 10 seconds in three different positions: side-by-side (both feet next to each other), semi-tandem (toes of one foot positioned next to the midpoint of the other foot), and tandem (heel of one foot placed in front of the toes on the other foot) while study staff timed them. Participants were scored on their ability to hold their balance in each posture for 10 seconds without stepping out of position or grabbing for support. The SPPB possesses good validity and reliability.^40,41

50-Foot Walk Test (Walk 50FT)⁴²

The Walk 50FT is a widely used timed measure of fast walking speed. This test involves a participant walking 50 feet without any use of an aid while study staff monitor and record their time in seconds. Participants were instructed to walk “as fast as possible.” Two trials of the Walk 50FT were conducted and we calculated an average score of the two trials. A lower Walk 50FT score is indicative of higher (i.e., faster) walking speed. The Walk 50FT is empirically supported, possesses strong validity and reliability, and demonstrates sensitivity to change over time among individuals with chronic disorders.^42–44

Interventions

Tai Chi

The Tai Chi group intervention was conducted in person and based on the classical Yang Tai Chi 108 postures.⁴⁵ These postures and their accompanying movements were selected for the Tai Chi intervention because they (1) are of moderate intensity, (2) are easy to understand, (3) involve meditative components for psychological stress, (4) address postural stability, and (5) can be beneficial for physical functioning without severe strain.^31,46 Prior RCTs evaluated similar Tai Chi interventions and documented salutary effects for chronic conditions.^47,48 Participants were provided with a Tai Chi manual at the start of the intervention that included information about session content, practice techniques, and safety precautions. Tai Chi sessions were led by instructors trained in Yang-style Tai Chi with over 20 years of experience. During sessions, instructors provided education about Tai Chi principles, focusing on how these principles can help one’s mind and body. Sessions involved a warm-up, review of Tai Chi principles, meditation with Tai Chi movement, relaxation, and breathing techniques. Instructors tailored sessions to the needs of participants to the extent that they accounted for differences in endurance and flexibility with individualized attention as indicated. Participants were asked to practice Tai Chi at home between sessions with guidance from supplemental materials found in their manuals and online video recordings made accessible to them. They were also encouraged to continue to practice after completing the Tai Chi intervention and were provided with online home practice materials (e.g., links to videos) and information on where they could access community-based Tai Chi groups.

Wellness

The Wellness group intervention was conducted in person and utilized as an active treatment comparison condition to Tai Chi.³¹ Education-based interventions emphasizing patient-centered goal-setting in areas of self-care pertinent to the VA’s Whole Health initiative can positively impact physical functioning by bolstering mobility, health competency, and goal achievement.^19,49 Moreover, wellness education has been employed as an active treatment comparison condition for Tai Chi in prior research on chronic health conditions.^50,51 In the present study, Wellness groups were facilitated by study staff with advanced training in psychology, and supervised by the principal investigators who are licensed and credentialed clinical psychologists. Facilitators provided psychoeducation via written materials and brief video clips about areas of self-care covered in the VA Whole Health initiative’s Circle of Health domains, such as physical activity, good sleeping habits, adaptive coping, and healthy eating.¹⁷ Facilitators also conducted a brief mindfulness activity during each session. The Wellness intervention began with information about the treatment rationale and structure, and an overview of the Circle of Health. During the third session, participants were provided with information about goal setting, and on setting Specific, Measurable, Action-oriented, Realistic, and Timed (SMART) goals. Participants were asked to generate up to three self-care goals each week in areas of their choosing with a focus on enhancing their physical or emotional health. In the subsequent sessions, each of the seven domains in the Circle of Health were covered over one to three sessions and facilitators led discussions and encouraged participants to consider the information in the context of their own life. The last two sessions consisted of reviewing the information covered over the course of the group and discussing ways to continue making progress after the group ended. Referrals were discussed as needed with participants who expressed interest.

Procedure

Full details on the parent RCT’s procedures are reported elsewhere.³² Procedures were approved by the Institutional Review Board at the VA Boston Healthcare System (Approval Number: 1577652). Veterans interested in participating were initially evaluated based on the eligibility criteria with a phone screen. If potentially eligible, veterans attended a baseline assessment where they provided informed consent and final eligibility was confirmed using subjective and objective health measures. Once officially eligible to participate, they were randomized (using an online randomizer) to receive either Tai Chi (n = 27) or Wellness (n = 26). Group treatment sessions lasted approximately 60 minutes and took place twice a week for 12 weeks (24 sessions total). Study staff and group leaders assessed treatment adherence using fidelity checklists. Measures were administered at the following timepoints: baseline, posttreatment, 3-month follow-up, and 9-month follow-up. Trained assessors were blinded to treatment condition. Most of the study measures were completed in person but, to maximize data capture, the PAR was completed via telephone when participants could not attend assessment sessions. The SPPB and Walk 50FT were never administered via telephone.

Data analyses

We employed an intent-to-treat approach to data analyses that retained all participants across timepoints regardless of treatment completion. We conducted analyses in R, specifically using the nlme package for multilevel regression models and the sjPlot package to create graphs.^52,53 There was a lack of variability in the data for the SBT subtest of the SPPB as the vast majority of participants successfully completed this subtest. To avoid a potential ceiling effect introduced by this subtest, we did not combine the three SPPB subtests into a summary score. Instead, we analyzed the other two SPPB subtests, Walk 4 M and CST, as separate outcome variables.

All variables were transformed using a log10 transformation except for PAR-Strength, which was transformed with a cube root transformation as it yielded better skewness and kurtosis. After transforming variables, the Walk 4M (skewness = 1.30, kurtosis = 2.27), Walk 50FT (skewness = 1.58, kurtosis = 3.54), and PAR-Strength (skewness = 1.67, kurtosis = 2.87) still demonstrated slightly elevated skewness and kurtosis. Given that the transformations substantially improved normality across outcome variables and that multilevel models are robust to non-normal data, we elected to use the transformed variables in all subsequent analyses.

We calculated raw means, standard deviations, medians, and interquartile ranges for each outcome variable across timepoint by treatment condition (coded: 0 = Wellness, 1 = Tai Chi). We evaluated baseline differences between treatments in the outcomes using independent-samples t-tests. Subsequently, we performed multilevel regression models to examine differences between treatments in the outcome variables from baseline to 9-month follow-up. First, we conducted an initial model for each outcome variable specifying time (coded: 1 = Baseline, 2 = Posttreatment, 3 = 3-Month Follow-Up, 4 = 9-Month Follow-Up) as linear and included it as a Level 1 predictor, treatment as a Level 2 predictor, and the linear time × treatment cross-level interaction. We then performed a second model for each outcome variable that retained the same predictors as the linear model but added quadratic time (i.e., time squared) as a Level 1 predictor and the cross-level interaction between quadratic time and treatment. All models utilized restricted maximum likelihood with a random effect for participant ID. Finally, we compared the linear and quadratic models on fit using an ANOVA with maximum likelihood to evaluate the difference between each model’s Akaike Information Criterion (AIC).

To present results as concisely as possible, we only report findings from multilevel regression models with better model fit (i.e., significantly lower AIC). If models were not significantly different, then we report results from the linear model to facilitate interpretation. We graphed significant time × treatment interactions for interpretation purposes. Further, we plotted the highest order interaction term (i.e., if both the linear time × treatment and quadratic time × treatment interactions were significant, then the quadratic time × treatment interaction was graphed) with model-implied trajectories and 95% confidence intervals. We used the alpha level of 0.05 for determining statistical significance across analyses. We calculated Cohen’s d effect sizes for significant effects within multilevel models. To determine the strength of an effect, we utilized Cohen’s⁵⁴ cutoff guidelines (0.20 = weak, 0.50 = moderate, 0.80 = strong).

Results

Table 1 presents the raw means and standard deviations of the outcomes across timepoint by treatment and Table 2 presents the medians and interquartile ranges of outcomes across timepoint by treatment. There were no significant differences between Tai Chi and Wellness on the outcomes at baseline (all p-values ≥ .12). Table 3 presents results from multilevel regression models for each outcome. There were no significant differences between linear and quadratic models in fit for the PAR-Strength (p = .59), PAR-Exercise (p = .44), CST (p = .34), and Walk 50FT (p = .61). Therefore, linear models were used for these four outcomes. However, no significant effects emerged in the linear models predicting these outcomes (all p-values ≥ .07).

Table 1.

Raw Means and Standard Deviations of Outcomes Across Timepoints by Treatment

	PAR-Strengtha		PAR-Flexibilitya		PAR-Exercisea		CSTb		Walk 4Mb		Walk 50FT
	Wellness	Tai Chi	Wellness	Tai Chi	Wellness	Tai Chi	Wellness	Tai Chi	Wellness	Tai Chi	Wellness	Tai Chi
Baseline	1.55 (5.24)	.38 (.76)	.45 (.82)	.38 (.62)	9.71 (9.07)	12.63 (10.28)	12.29 (3.49)	13.00 (3.14)	5.54 (.89)	5.50 (1.61)	10.47 (3.68)	10.56 (2.34)
Post-Intervention	.19 (.49)	.65 (1.38)	.26 (.58)	2.17 (1.85)	11.71 (10.09)	11.35 (8.15)	13.23 (7.94)	13.92 (5.03)	4.70 (1.02)	5.23 (1.19)	9.64 (2.74)	11.27 (4.03)
3-Month Follow-Up	.30 (.61)	.39 (.79)	.80 (1.28)	1.28 (1.24)	13.07 (15.65)	7.28 (4.89)	12.03 (3.95)	13.05 (5.66)	4.56 (.41)	5.52 (1.56)	9.49 (1.30)	10.91 (3.27)
9-Month Follow-Up	.22 (.53)	.35 (.68)	.36 (.69)	.88 (1.26)	11.31 (9.28)	12.39 (11.43)	10.96 (2.83)	14.61 (3.75)	4.55 (.36)	5.49 (1.53)	8.69 (.76)	11.01 (4.28)

Note. Standard deviations are presented in parentheses. Wellness: n = 26. Tai Chi: n = 27.

^aComputed using 7-Day Physical Activity Recall.

^bAdministered as part of the Short Physical Performance Battery.

Abbreviations: CST = Chair Stand Test; PAR-Exercise = Physical Activity Recall-Exercise; PAR-Flexibility = Physical Activity Recall-Flexibility; PAR-Strength = Physical Activity Recall-Strength; Walk 4M = 4-Meter Walk; Walk 50FT = 50-Foot Walk

Table 2.

Raw medians and interquartile ranges of outcomes across timepoints by treatment.

	PAR-Strength1 (hours)		PAR-Flexibility1 (hours)		PAR-Exercise1 (hours)		CST2 (seconds)		Walk 4M2 (seconds)		Walk 50FT (seconds)
	Wellness	Tai Chi	Wellness	Tai Chi	Wellness	Tai Chi	Wellness	Tai Chi	Wellness	Tai Chi	Wellness	Tai Chi
BL	0.00 (0.00–0.69)	0.00 (0.00–0.25)	0.00 (0.00–0.50)	0.00 (0.00–0.75)	6.88 (3.25–14.00)	9.00 (4.75–17.12)	12.00 (10.44–13.00)	12.33 (10.96–15.78)	5.55 (4.99–6.02)	5.06 (4.57–5.97)	9.52 (9.00–10.56)	9.75 (9.12–11.43)
PT	0.00 (0.00–0.00)	0.00 (0.00–0.81)	0.00 (0.00–0.00)	2.00 (0.94–2.81)	10.50 (4.25–13.25)	8.88 (6.69–13.38)	11.00 (9.57–13.85)	12.88 (10.14–17.79)	4.50 (4.00–5.06)	5.00 (4.50–5.52)	9.33 (8.50–9.73)	10.00 (9.00–12.52)
3M	0.00 (0.00–0.38)	0.00 (0.00–0.25)	0.00 (0.00–1.50)	1.25 (0.38–1.75)	8.75 (3.50–16.75)	6.50 (4.12–9.00)	11.28 (9.25–14.00)	10.12 (9.62–15.00)	4.50 (4.25–5.00)	5.00 (4.65–6.00)	9.25 (9.00–10.16)	10.00 (8.62–11.50)
9M	0.00 (0.00–0.00)	0.00 (0.00–0.50)	0.00 (0.00–0.12)	0.50 (0.00–1.25)	8.00 (3.69–18.25)	9.50 (6.75–16.00)	10.36 (9.44–12.00)	15.46 (11.37–16.46)	4.60 (4.23–4.86)	4.80 (4.57–5.82)	9.00 (8.64–9.10)	9.75 (8.23–11.40)

Interquartile ranges are presented in parentheses. Wellness: n = 26. Tai Chi: n = 27.

3 M: 3-Month Follow-up; 9 M: 9-Month Follow-Up; BL: Baseline; CST: Chair Stand Test; PAR-Exercise: Physical Activity Recall-Exercise; PAR-Flexibility: Physical Activity Recall-Flexibility; PAR-Strength: Physical Activity Recall-Strength; PT: Posttreatment; Walk 4M: 4-Meter Walk Test; Walk 50FT: 50-Foot Walk Test.

¹Computed using 7-Day Physical Activity Recall. Variable represents a sum of the total time spent engaging in the specific activity (strength training, flexibility exercise, or at least moderate exercise) per 7-day period.

²Administered as part of the Short Physical Performance Battery.

Table 3.

Multilevel regression models for outcome variables.

	Outcome Variable					Walk 50 FT2
	PAR-Strength1	PAR-Flexibility2	PAR-Exercise2	CST2	Walk 4M2
Effect: b (SE)
Intercept	0.45(0.13)***	0.06(0.10)	0.83(0.09)***	1.11(0.03)***	0.90(0.03)***	1.05(0.02)***
Linear Time	−0.07(0.04)	0.05(0.10)	0.03(0.03)	−0.001(0.01)	−0.11(0.03)***	−0.01(0.01)
Quadratic Time	–	−0.01(0.02)	–	–	0.02(0.01)***	–
Treatment (Tai Chi)	−0.09(0.18)	−0.35(0.15)*	0.21(0.13)	0.01(0.04)	−0.09(0.04)*	0.01(0.03)
Linear Time × Treatment (Tai Chi)	0.05(0.06)	0.47(0.13)***	−0.06(0.04)	0.01(0.01)	0.09(0.04)*	0.01(0.01)
Quadratic Time × Treatment (Tai Chi)	–	−0.09(0.03)**	–	–	−0.02(0.01)*	–

We only report estimates for Quadratic Time and Quadratic Time x Treatment for outcome variables that were evaluated using a quadratic model.

CST: Chair Stand Test; PAR-Exercise: Physical Activity Recall-Exercise; PAR-Flexibility: Physical Activity Recall-Flexibility; PAR-Strength: Physical Activity Recall-Strength; Walk 4M: 4-Meter Walk Test; Walk 50FT: 50-Foot Walk Test.

¹Cube root transformation.

²Log10 transformation.

^*p < .05;

^**p < .01;

^***p < .001.

The quadratic model for PAR-Flexibility demonstrated a better fit compared to the linear model (p < .001). In this model, there was a main effect of treatment, b = −0.35, SE = 0.15, p = .02, d = −0.34. Also, the linear time × treatment interaction, b = .47, SE = 0.13, p < .001, d = 0.31, and quadratic time × treatment interaction, b = −0.09, SE = 0.03, p = .001, d = −0.30, were significant. The graph of the quadratic time × treatment interaction for PAR-Flexibility revealed that participants in Tai Chi greatly increased their time engaged in flexibility exercises from baseline to 3-month follow-up, but then greatly declined from 3-month follow-up to 9-month follow-up (see Figure 1). Participants in Wellness slightly increased in time engaged in flexibility exercises from baseline to posttreatment but gradually decreased from posttreatment to 9-month follow-up. Participants in Tai Chi demonstrated significantly greater time engaged in flexibility exercises at posttreatment and 3-month follow-up compared to participants in Wellness as evidenced by non-overlapping 95% confidence intervals at these timepoints. Participants in both treatments reported similar levels of time engaged in flexibility exercises at the 9-month follow-up.

Figure 1. Quadratic Time × Treatment Interaction for PAR-Flexibility.

Note. PAR-Flexibility was transformed using a log10 transformation. This graph revealed that participants in Tai Chi greatly improved from baseline to 4-month follow-up, but then rapidly declined from 3-month follow-up to 9-month follow-up. In contrast, PAR-Flexibility ratings in participants who received Wellness slightly improved from baseline to posttreatment but gradually decreased from posttreatment to 9-month follow-up. Participants in both treatments reported similar levels of PAR-Flexibility ratings at the 9-month follow-up.

Abbreviations: PAR-Flexibility = Physical Activity Recall-Flexibility; 1 = Baseline; 2 = Posttreatment; 3 = 3-Month Follow-Up; 4 = 9-Month Follow-Up

The quadratic model for Walk 4M exhibited a better fit relative to the linear model (p = .003). This model indicated significant main effects for linear time, b = −0.11, SE = 0.03, p < .001, d = −0.42, quadratic time, b = 0.02, SE = 0.01, p < .001, d = 0.35, and treatment, b = −0.09, SE = 0.04, p = .04, d = −0.30. The linear time × treatment interaction, b = 0.09, SE = 0.04, p = .01, d = 0.26, and quadratic time × treatment interaction, b = −0.02, SE = 0.01, p = .04, d = −0.21, were also significant in this model. The graph of the quadratic time × treatment interaction for the Walk 4M displayed that times to complete the test in Wellness greatly decreased (i.e., gait speed increased) from baseline to 3-month follow-up, but then increased (i.e., gait speed decreased) from 3-month follow-up to 9-month follow-up (see Figure 2). For Tai Chi, times to complete the Walk 4M slightly decreased from baseline to posttreatment (i.e., gait speed increased), were maintained from posttreatment to 3-month follow-up, and then slightly increased (i.e., gait speed decreased) from 3-month follow-up to 9-month follow-up to the extent that completion times were similar at baseline and 9-month follow-up. Changes in times to complete the Walk 4M were similar across both treatments as indicated by overlapping 95% confidence intervals.

Figure 2. Quadratic Time × Treatment Interaction for Walk 4M.

Note. This graph displayed that Walk-4M times for participants in Wellness greatly decreased from baseline to 4-month follow-up, but then increased from 4-month follow-up to 9-month follow-up. For participants in Tai Chi, Walk-4M times slightly reduced from baseline to posttreatment, were maintained from posttreatment to 4-month follow-up, and then slightly increased from 4-month follow-up to 9-month follow-up to the extent that times were similar at baseline and 9-month follow-up. Changes in Walk-4M times were similar in both treatments as indicated by overlapping 95% confidence intervals (shaded region) across timepoints, thereby implying that while these differences were statistically significant, they may not have been clinically significant.

Abbreviations: Walk 4M = 4-Meter Walk; 1 = Baseline; 2 = Posttreatment; 3 = 3-Month Follow-Up; 4 = 9-Month Follow-Up

Discussion

The present study evaluated Tai Chi relative to a Wellness intervention comparison condition on physical functioning outcomes in veterans with GWI. While we hypothesized that Tai Chi would be related to greater improvements than Wellness, results provide modest support for both group treatments in enhancing physical functioning outcomes. Yet, each intervention evidenced a unique effect on specific outcomes. Tai Chi was related to greater increases in time spent engaging in flexibility exercises whereas Wellness was associated with improvements in gait speed, though between-treatment effect sizes were small and confidence intervals between interventions were overlapping in the gait speed model. Furthermore, the effects of Tai Chi on engagement in flexibility exercises and Wellness on gait speed diminished at the 9-month follow-up. These results align with prior research demonstrating weak long-term efficacy for exercise and educational interventions.⁵⁵ The diminishing effects of the interventions at 9-month follow-up may be explained by participants not regularly practicing Tai Chi in the long term or shifting their focus onto other health goals over time. Booster sessions or motivational telephone check-ins may have prolonged intervention effects or encouraged continual practice of skills learned in treatment.⁵⁵

Regarding unique effects, participants who received Tai Chi reported greater time spent engaging in flexibility exercises at posttreatment and 3-month follow-up compared to those who received Wellness with a small effect size. This finding is consistent with previous research supporting Tai Chi’s beneficial effect on flexibility and related exercise.⁵⁶ The gentle stretching and bodily movement of Tai Chi may build self-efficacy in exercising and motivate individuals to engage in more flexibility exercise to improve their health.^57,58 Further, spending time on flexibility exercise is critical because continued practice on flexibility may facilitate downstream improvements in GWI symptoms, including cognitive functioning, depressed mood, and pain.^59–61 Evaluating the impact of greater flexibility exercise on other health domains, such as GWI symptom clusters, following Tai Chi may be an important avenue for future study.

While gait speed improved in both treatment conditions, greater increases in gait speed were associated with Wellness relative to Tai Chi. This result is unsurprising considering that many of the participants in Wellness set goals for increasing their physical activity.⁶² The psychoeducation and goal-attainment strategies utilized in Wellness may have motivated individuals to walk more in their daily lives, which led to improvements in mobility, specifically gait speed.⁴⁹ Moreover, other self-care goals that were not directly related to physical activity but fueled general improvements in health (e.g., diet, sleep) may have contributed to better gait speed in participants who received Wellness. In contrast, Tai Chi emphasizes slow, mindful movement, and it may be that those who received Tai Chi walked differently (i.e., more upright and elongated spine, more deliberate movement with a focus on balance) following the intervention, which resulted in a slower gait speed. Nevertheless, the between-treatment effect size was small and the 95% confidence intervals of changes in gait speed over time for both interventions were overlapping. Taken together, these results imply that both treatments likely have similar real-world benefits for this outcome. Future research is needed comparing Tai Chi and Wellness in terms of their effect on gait speed to replicate our findings.

From a broader perspective, our results are promising because they support two treatments for improving specific physical functioning outcomes in veterans with GWI. GWI is considered a progressive illness as research documents that symptoms may worsen over time with few treatments exhibiting a strong salutary effect on the deteriorating physical health characteristic of this condition.^5,21 Thus, any intervention that can slow declines in aspects of physical functioning may be worthwhile for providers to consider in the treatment of GWI. Further evaluation of Tai Chi and Wellness to address GWI is warranted.

The present study possesses some limitations. Although we examined the effect of Tai Chi and Wellness on a range of physical functioning outcomes, we found support for their efficacy only on time spent engaging in flexibility exercises and gait speed, respectively. Therefore, we tempered our interpretations for their impact on physical health broadly and emphasize the need for future research to replicate and extend our results. Differences between intervention conditions were minimal as evidenced by small between-treatment effect sizes and overlapping confidence intervals between interventions for gait speed. There are several potential explanations for the minimal differences between interventions. First, the relatively small sample size may have hindered our ability to detect large differences between interventions. Furthermore, the interventions under study may have exhibited comparable effects on physical functioning outcomes because they were both active treatments, designed comprehensively and rigorously with strong theoretical- and research-driven frameworks.

Although objective measures were used for most of the outcomes, physical activity was indexed via self-report during the PAR, which is susceptible to biases in memory, context effects, and the demand characteristics of an RCT design. Participants may have misclassified their time spent in Tai Chi on the PAR or reduced their participation in other forms of physical activity while participating in Tai Chi, which could explain why some outcomes (e.g., PAR-Exercise) changed minimally from baseline to posttreatment. Depending on when the PAR was administered relative to the end of the intervention, it may have gauged different levels of physical activity, potentially adding variability to the measure. Additionally, it is important to note that time spent engaged in strength or flexibility activities as measured by the PAR does not necessarily equate to gains in strength or flexibility. Thus, use of objective measures of physical functioning, such as actigraphy and range of motion measures (to assess flexibility and strength), are recommended for future studies.

In addition, future RCTs would benefit from larger, more diverse samples given that the current study’s sample was small and consisted predominantly of white males. Due to the small sample size, we did not adjust multilevel regression models for covariates to maximize degrees of freedom and statistical power. As well, although attendance rates are consistent with prior research on intensive interventions for veterans, just over half of our participants did not receive a full “dose” (75% or more) of the study interventions.⁶³ Further research investigating these interventions in larger samples with greater variability in gender and ethnic/race characteristics will allow for consideration of relevant covariates (e.g., demographics, session attendance) in adjusted models to better test the robustness of treatment effects.

Conclusions

Despite study limitations, both Tai Chi and Wellness may be beneficial for improving physical functioning outcomes in veterans with GWI and pain. Specifically, Tai Chi may enhance engagement in flexibility exercise and Wellness may improve gait speed. The unique effects of these treatments may point to different mechanisms of action, which are worth examining in future studies. Yet, given the minimal differences between the interventions observed in the present study, future research elucidating the moderators of CIH interventions in veterans is warranted. In the end, our results are promising because we provide preliminary evidence for the utility of two treatments that may improve specific physical functioning outcomes in veterans with GWI, especially in accommodating physical difficulties associated with the condition that few treatments have addressed.²¹

Data availability statement

The data that support the findings of this study are available from the corresponding author, CP, upon reasonable request.