Journal Logo

Research Paper

Exercise combined with Acceptance and Commitment Therapy compared with a standalone supervised exercise programme for adults with chronic pain: a randomised controlled trial

Casey, Máire-Brída,*; Smart, Keith M.a; Segurado, Ricardoa; Hearty, Conorb; Gopal, Harib; Lowry, Damienc; Flanagan, Dearbhaild; McCracken, Lancee; Doody, Catherinea

Author Information
doi: 10.1097/j.pain.0000000000002487

Abstract

1. Introduction

Chronic pain is a highly prevalent and costly health problem3,33,34 that has a profound impact on an individual's physical and emotional well-being and quality of life.33 Although there are many treatment options available for people with chronic pain, the effects of these interventions are modest and more research examining long-term benefits is required.46,49 Exercise and psychological therapies such as cognitive behavioural therapy (CBT) are examples of conservative treatments that are known to be effective in reducing pain and disability.11,39,49,50 These interventions can be provided individually or they can be combined and delivered in the form of a multidisciplinary pain management programme (PMP).4 Although PMPs are recommended for chronic pain,39 systematic reviews have reported only small benefits of these programmes compared with usual care or physical interventions for chronic low back17 or nonspecific spinal pain.30 A recent systematic review and meta-analysis conducted by our research group found low-quality evidence for superior effects in favour of multidisciplinary-based rehabilitation (MBR) compared with active physical interventions for pain and disability, predominantly in patients with chronic low back pain.6 There is a need for more high-quality studies examining the effectiveness of multidisciplinary PMPs for all types of chronic pain and particularly further research to help determine which elements of these programmes are effective.28

Acceptance and Commitment Therapy (ACT) is a third wave psychological therapy that encourages individuals to change their relationship with feared or avoided thoughts and physical sensations using mechanisms such as acceptance, mindfulness, and value-based action.14 Three systematic reviews of ACT for chronic pain have been published in the last 5 years, which show small to moderate effects across a range of outcomes in favour of ACT vs inactive controls.13,15,47 However, a recently updated Cochrane review of psychological therapies for chronic pain urges cautious interpretation of the results of published studies related to ACT50 and a need for more adequately powered randomised controlled trials (RCTs) comparing ACT with active control interventions has been highlighted.15

To the best of our knowledge, this is the first RCT to compare the effectiveness of a combined group-based exercise and ACT treatment programme (ExACT) with an active control intervention for chronic pain. The primary aim of the study was to determine whether ExACT was superior for reducing pain interference at 12-week follow-up, compared with a supervised exercise programme alone. Secondary objectives were to investigate whether ExACT had a significant effect for study participants compared with a supervised exercise programme regarding pain severity, self-perception of change, patient satisfaction, depression, anxiety and healthcare costs, pain self-efficacy, pain catastrophizing, fear avoidance, pain acceptance, and committed action following treatment and at 12-week follow-up. We also sought to examine whether ExACT would have a significant effect on objective physical activity measures (step count, distance travelled, and active minutes) posttreatment, compared with the supervised exercise programme.

2. Methods

2.1. Study design and setting

The ExACT trial was a single-centre, 2-armed, parallel group, randomised controlled superiority trial. Participants were recruited from a consultant-led pain clinic and musculoskeletal outpatient clinics of a large academic teaching hospital (Mater Misericordiae University Hospital) in Dublin, Ireland. Treatments were provided in the pain clinic and in the physiotherapy department of the hospital. Ethical approval was granted by the Mater Misericordiae University Hospital Institutional Review Board (Reference 1/378/1864). The trial was prospectively registered on ClinicalTrials.gov (NCT03050528), and a trial protocol was published.5 The results are reported in accordance with the CONSORT guidelines.37

2.2. Participants

Adults (aged ≥ 18 years) with any type of chronic pain (other than cancer pain), diagnosed by a medical doctor, and who reported a score of ≥2 on the Brief Pain Inventory Interference Scale were eligible for inclusion in the study.

The exclusion criteria were as follows: need for further diagnostic evaluation (determined by a medical doctor), presence of major medical or psychiatric disorder that would impede ability to participate with treatment, presence of active cancer or cancer-related pain, unstable inflammatory condition (eg, rheumatoid arthritis or gout), presence of substance misuse, surgical or interventional procedure (eg, spinal cord stimulator, rhizotomy, epidural, or intra-articular injection) within the preceding 3 months, concurrent participation (or participation within the preceding 3 months) in a supervised exercise programme or psychological or physiotherapy treatment, previous participation in any multidisciplinary PMP, or presence of any contraindication to participation in a gymnasium or pool-based exercise programme (eg, shortness of breath at rest, unstable diabetes or epilepsy, recent myocardial infarction, stroke, pulmonary embolism, asthma attack, weight > 125 kg (19.5 stone), or waist circumference > 50 inches [restriction due to weight limit on hydrotherapy evacuation equipment]).

Participants were screened for study eligibility by a medical doctor. Those who met the eligibility criteria were informed about the study, and written information in the form of a patient information leaflet was provided. Individuals who expressed an interest in participating in the study were contacted by telephone by the lead researcher (M-B.C.). Any questions were clarified on the phone, and those who remained interested were invited to attend an individual face-to-face appointment with the lead researcher in the hospital pain clinic. Baseline outcome measures were sent in advance by post, and the participants were asked to bring the completed questionnaires with them on the day of their appointment. Informed consent was obtained in writing by the lead researcher before participation in the study.

2.3. Randomisation

Randomisation took place after completion of baseline measures as recommended by the CONSORT guidelines.37 Randomisation was coordinated by the senior author (C.D.), who had no involvement in eligibility screening, enrolment, or treatment processes. On receipt of signed consent forms and baseline measures, participants were assigned a unique code and randomised using an online randomisation database.40 The computer-generated randomisation schedule applied a permuted block design (block sizes 4, 6, and 8) to ensure that the groups were balanced periodically. The block size was concealed, and the randomisation list remained with C.D. until the study was complete. Allocation of participants was communicated to administrative staff in the pain clinic by C.D. through email. Participants were informed of their group allocation in writing by the administrative staff, who sent notification in sealed, opaque envelopes.

2.4. Interventions

The trial interventions have been described in detail in the published protocol,5 and the main components are outlined in this article.

  • (1) Combined Exercise and Acceptance and Commitment Therapy (ExACT)

Combined Exercise and ACT was an 8-week multidisciplinary pain programme combining exercise and psychological therapy. It was a face-to-face, group-based treatment, with up to 11 individuals per group (average 9.6 participants per group). Participants attended once per week, for a duration of three and a half hours. Each day began with a 2-hour session with a psychologist, based on the psychological approach ACT. Acceptance and Commitment Therapy is designed to promote psychological flexibility through methods that encourage openness, awareness, and engagement. The overall aim is to reduce pain-avoidant behaviours, in the service of living a rich and meaningful life.14 Group discussions, experiential exercises, and mindfulness practice (Supplemental data, Table 1, available at https://links.lww.com/PAIN/B497) were led by a psychologist who had been trained in ACT and had more than 6 years of experience treating chronic pain. The content of the sessions was adapted from an ACT treatment manual used in a recently published study2 and available from the web site www.contextualscience.org/better_living_with_illness.

After a break for lunch, participants attended a 1.5-hour supervised exercise class in the physiotherapy department of the hospital. The classes were delivered by a senior physiotherapist who had over 8 years of experience treating patients with musculoskeletal pain. They featured 2 components (1) education or advice and (2) exercise (Supplemental Data, Table 2, available at https://links.lww.com/PAIN/B497). The education or advice sessions covered topics including goal setting, understanding pain, pacing, and managing setbacks. The exercise classes took place in a gymnasium setting (4 sessions) or hydrotherapy pool (4 sessions).

  • (2) Supervised Exercise

The supervised exercise intervention also consisted of 8 face-to-face sessions of 1.5-hour duration, delivered on a weekly basis to groups of up to 11 participants (average 9.6 participants per group). The intervention was delivered by the same senior physiotherapist and mirrored the supervised exercise component of the combined treatment programme as described above.

2.5. Outcome measures

The outcome measures included in this trial have been described in detail in the published trial protocol.5 A brief summary is provided below.

2.5.1. Demographic data

Demographic data collected included age, sex, education level, relationship status, and employment status. Data related to pain history included duration of pain symptoms and healthcare professional diagnosis.

2.5.2. Primary outcome

The primary outcome for the study was pain interference, measured at 12-week follow-up. Pain interference was assessed with the Brief Pain Inventory Interference Scale.7 This is a 7-item self-report questionnaire that measures the extent to which pain interferes with functions such as general activity, walking ability, normal work, mood, relations with people, enjoyment of life, and sleep. A reduction of 1 point on the interference scale has been suggested to represent a clinically meaningful change.10 Details of the psychometric properties of the study outcome measures can be found in the published trial protocol.5

2.5.3. Secondary outcomes

Secondary outcomes were assessed postintervention and again at 12-week follow-up. Pain intensity was assessed using the Brief Pain Inventory (BPI)-pain severity subscale.7 Depression and anxiety were measured with the Patient Health Questionnaire19 and the General Anxiety Disorder.41 Patient satisfaction with treatment was measured using a single question (question 7) from the Client Satisfaction Questionnaire-8.20 The question asked “In an overall, general sense, how satisfied are you with the service you have received?” and 4 potential responses were provided (very satisfied, mostly satisfied, indifferent or mildly dissatisfied, and quite dissatisfied). Participants' perceived level of improvement was assessed with the Patient Global Impression of Change Scale.16 Healthcare utilisation (general practitioner visits, other healthcare professional visits, emergency department visits, and number of days hospital inpatient stay in the preceding 12 weeks) was recorded at baseline and 12-week follow-up. The occurrence of any adverse events was monitored by the treating clinicians throughout the 8-week intervention period.

2.5.4. Treatment process measures

Treatment process measures were also assessed postintervention and again at 12-week follow-up. Pain self-efficacy was assessed with the Pain Self-Efficacy Questionnaire,29 pain catastrophising with the Pain Catastrophising Scale,42 and fear of movement and reinjury with the Tampa Scale of Kinesiophobia.35 Pain acceptance was measured using the Chronic Pain Acceptance Questionnaire-836 and committed action with the Committed Action Questionnaire-8.22

2.5.5. Physical activity outcomes

Physical activity was measured using a Fitbit Zip wearable activity tracker. Data collected included average weekly step count, distance travelled, and active minutes. Participants were provided with the activity tracker at the baseline appointment and were instructed in how to use it. The activity trackers were worn by participants during waking hours for 1 week before starting the interventions and for the full duration of the interventions (9 weeks in total). The main time points for analysis were the baseline week and the last week of the intervention.

2.6. Statistical analyses

2.6.1. Sample size

Sample size was estimated using a target power of 80%, at a type I error rate of 0.05, and was calculated relative to the primary outcome measure: pain interference subscale of the BPI. The statistical test assumed was an independent samples t test for group differences in the change from baseline to subsequent assessment, assuming the randomisation ensures no systematic baseline or other covariate group differences. The minimal clinically significant difference for the interference scale of the BPI is 1 unit (SD of improvement of 2 units).10

Calculation produced a suggested sample size of 64 per group. Allowing for a potential attrition rate of 20%, our final sample size is 80 participants per group.

2.6.2. Outcome analyses

Outcome analyses were conducted by 2 of the authors (M-B.C. and R.S.) who were blinded to treatment group allocation. SPSS version 24 was used for all analyses, and no interim analyses were conducted. Descriptive statistics were calculated for all outcome measures at each time point. The primary analysis compared the effect of the interventions on the primary outcome: pain interference at 12-week postcompletion of the intervention. All outcome analyses were conducted according to an intention-to-treat principle, ie, all randomised participants were included in the main analysis and were analysed as randomised, regardless of protocol adherence. Secondary analysis included the analysis of the primary outcome postintervention and analysis of the secondary and treatment process outcomes detailed previously postintervention and at 12-week follow-up. Linear mixed models on the outcome measures over time were fitted to evaluate the effectiveness of both interventions, which intrinsically adjusted for pretreatment scores. Statistical significance for the primary outcome was assessed from a P value < 0.05 from the group by time interaction term. The Bonferroni method was used to appropriately adjust the overall level of significance for multiple secondary outcomes, resulting in a corrected P value significance level of P < 0.002. As a key component of the reporting of the analyses of outcomes, the mean changes (irrespective of statistical significance), confidence intervals, and correlations of the measures between the assessment periods are reported.

2.6.2.1. Missing data

Careful attention was paid to ensure that all participants were fully assessed at all time points. Baseline data were checked for missing data by the lead researcher at the baseline assessment, and participants were encouraged to complete any missing answers. The trial supervisor followed up on any missing data by telephone or post at the subsequent time points. Multiple imputation was not conducted because studies have demonstrated that linear mixed modelling is sufficient to control for missing data.32,45

2.6.3. Sensitivity analysis

The following sensitivity analysis was undertaken and is reported in accordance with the study protocol.

  • (1) A per-protocol analysis: The per-protocol analysis excluded participants found to be ineligible after randomization and those who attended less than 50% of the intervention. Both intention-to-treat and per-protocol analysis sets are reported.

2.7. Deviations from the registered trial protocol

A total of 175 participants were recruited, rather than 160 as was originally planned to ensure that the study remained adequately powered. The eligibility criterion related to the BPI interference subscale score was reduced from ≥ 3 to ≥ 2 to be more inclusive. We also redefined some of the outcomes (Pain Self-Efficacy Questionnaire and Pain Catastrophising Scale) from secondary outcomes to treatment process measures to reflect how these measures represent potential therapeutic mechanisms. The intervention fidelity assessment described in the protocol was modified. The fidelity assessment of the ACT intervention was conducted as planned, but we did not test fidelity of the exercise intervention because of limited time and resources. A cost-consequence analysis had also been planned, but the data collected in relation to cost was insufficient to allow an informative analysis.

3. Results

3.1. Participant flow

The flow of participants through the trial is illustrated in Figure 1. Between February 2017 and December 2018, 378 individuals attending the Mater Misericordiae University Hospital for treatment of a chronic pain condition were assessed for eligibility to take part in the study. Most of these patients attended the Department of Pain Medicine (n = 331), and the remainder were recruited from the rheumatology department (n = 47). Sixty-one individuals (16.14%) were found to be ineligible, and 142 (37.6%) declined to participate (Fig. 1). In total, 175 participants agreed to take part and provided written consent before enrolment. Participants were recruited in groups of approximately 20 and then individually randomised to receive either the ExACT or supervised exercise alone. This process was repeated 9 consecutive times over the course of the study, with 87 individuals randomised to ExACT and 88 individuals to the supervised exercise intervention.

F1
Figure 1.:
Flow chart. CONSORT, Consolidated Standards of Reporting Trials; ExACT, Exercise combined with Acceptance and Commitment Therapy; ITT, intention to treat.

Twenty-five trial participants (14.3%) failed to attend the intervention to which they were assigned. The rate of nonattendance was higher in the supervised exercise group, with n = 15 participants (17%) failing to attend any of the treatment sessions, compared with n = 10 participants (11.5%) from the ExACT group. A further 27 trial participants (15.4%) withdrew prematurely from the treatment programme having attended at least 1 session. Fifteen participants (17%) withdrew prematurely from the supervised exercise programme, compared with 12 individuals from the ExACT intervention (13.8%). Reasons for failing to commence the interventions and reasons for premature withdrawal from the treatment programmes are presented in Supplemental Data (supplemental tables 3 and 4, available at https://links.lww.com/PAIN/B497). The most common reason reported for both nonparticipation and withdrawal was illness or injury. None of these injuries occurred during the treatment programmes or were attributed by participants to the intervention. One participant was withdrawn from the ExACT group because they were found to be ineligible to participate in the exercise component of the intervention. Apart from this particular individual, attempts were made to follow-up all participants, regardless of their attendance, postintervention and at 12-week follow-up. All participants who provided data were included in the intention-to-treat analyses.

Overall, the trial participants attended a mean of 4.42 of the 8 sessions of the intervention to which they had been allocated (SD 2.71). Attendance was higher in the ExACT group (mean 4.94 sessions, SD 2.73), compared with the supervised exercise group (mean 3.9 sessions, SD 2.6) (Supplemental Data, Table 5, available at https://links.lww.com/PAIN/B497).

A total of 135 participants (77.1%) completed the self-report outcome measures postintervention, and 139 participants (79.4%) returned the questionnaires at 12-week follow-up. The response rate postintervention was slightly higher from participants in the ExACT group (80.5%), compared with a response rate of 73.9% in the supervised exercise group. Equal numbers of participants returned the questionnaires at 12-week follow-up (79% response rate in both groups).

3.2. Participant characteristics

The sociodemographic characteristics of the trial participants are presented in Table 1. Baseline characteristics were well balanced between the treatment groups. Participants of the supervised exercise group reported a slightly longer duration of chronic pain symptoms at baseline (10.07 years) compared with 8.83 years in the ExACT group. The mean number of pain sites at baseline was reported to be 4.29 (SD 2.37). The body chart featured in the BPI questionnaire was used to record participants' areas of symptoms. There are no specific guidelines for interpreting this information in the BPI user guide.7 The method of differentiation of pain sites that we used is illustrated in Supplemental Data Figure 1 (available at https://links.lww.com/PAIN/B497). High levels of education were observed in both groups, with 86.3% of the overall number of trial participants completing secondary school education or higher. Despite these high levels of education, a substantial number of participants were unemployed (73.7%) (Table 1).

Table 1 - Participant sociodemographic characteristics.
Participant characteristic Supervised exercise (n = 88), n (%) or mean ± SD ExACT (n = 87), n (%) or mean ± SD
Age (y) 47.69 ± 11.67 48.41 ± 10.39
Gender (female) 64 (72.7%) 60 (69%)
Duration of chronic pain symptoms (y) 10.07 ± 8.29 8.83 ± 7.52
Number of pain sites 4.44 ± 2.34 4.14 ± 2.39
Relationship status
 Married/cohabiting 46 (52.3%) 51 (58.6%)
 In a relationship, not cohabiting 9 (10.2%) 9 (10.3%)
 Single 33 (37.5%) 27 (31%)
Highest level of education achieved
 Primary school 11 (12.8%) 10 (11.6%)
 Secondary school 51 (59.3%) 46 (53.5%)
 Third level (undergraduate) 18 (20.9%) 23 (26.7%)
 Higher level (postgraduate degree) 5 (5.8%) 6 (7%)
 Other 1 (1.2%) 1 (1.2%)
Work status
 Not working 61 (70.1%) 68 (78.2)
 Working full time 15 (17.2%) 7 (8%)
 Working part-time 9 (10.3%) 9 (10.3%)
 Student 2 (2.3%) 2 (2.3%)
 Maternity leave 0 1 (1.1%)
ExACT, Exercise combined with Acceptance and Commitment Therapy.

A total of 51 different diagnoses were provided for the study participants by a variety of healthcare professionals (Supplemental Data, Table 6, available at https://links.lww.com/PAIN/B497). Diagnoses were based on many different factors including pain location (eg, low back pain), mechanism of injury (eg, knee pain after road traffic accident), and pathology (eg, trochanteric bursitis). Other diagnoses were nonspecific such as myofascial-mediated body pain and chronic widespread pain.

The participant characteristics for baseline measures are presented in Table 2. The scores were well balanced between the treatment groups across most of the health outcomes measured.

Table 2 - Baseline clinical characteristics.
Outcome measure (scoring range) Supervised exercise ExACT Total
n Mean ± SD n Mean ± SD n Mean ± SD
BPI interference score (0-10)* 87 6.92 ± 1.57 87 6.86 ± 1.75 174 6.89 ± 1.65
BPI composite severity score (0-10)* 88 6.53 ± 1.60 87 6.17 ± 1.62 175 6.35 ± 1.61
PHQ-9 (0-27)* 88 14.19 ± 6.32 87 14.43 ± 6.06 175 14.31 ± 6.18
GAD-7 (0-21*) 88 11.45 ± 5.84 87 11.01 ± 5.6 175 11.23 ± 5.71
PSEQ (0-60) 88 25.31 ± 12.27 87 23.86 ± 11.49 175 24.59 ± 11.88
PCS total (0-52)* 87 28.77 ± 12.17 86 32.36 ± 11.44 173 30.55 ± 11.92
PCS rumination (0-16)* 87 9.62 ± 4.21 87 11.07 ± 4.08 174 10.34 ± 4.20
PCS magnification (0-12)* 88 5.53 ± 3.34 87 6.22 ± 3.37 175 5.87 ± 3.36
PCS helplessness (0-24)* 88 13.53 ± 5.77 86 15.06 ± 5.62 174 14.29 ± 5.73
TSK (17-68)* 84 40.36 ± 7.19 86 41.88 ± 7.77 170 41.13 ± 7.51
CPAQ-8 total (0-48) 88 20.49 ± 7.87 87 19.07 ± 7.92 175 19.78 ± 7.90
CPAQ-8 AE (0-24) 88 12.19 ± 6.0 87 11.53 ± 5.50 175 11.86 ± 5.75
CPAQ-8 PW (0-24) 88 8.3 ± 4.61 87 7.54 ± 4.98 175 7.92 ± 4.80
CAQ-8 (0-48) 88 28.4 ± 7.87 87 26.89 ± 8.28 175 27.65 ± 8.09
Average daily step count (Fitbit Zip) 64 7388 ± 3801 73 7481 ± 3420 137 7438 ± 3590
Distance travelled (Fitbit Zip) 64 5.14 ± 2.72 73 5.26 ± 2.51 137 5.21 ± 2.60
Active minutes (Fitbit Zip) 59 232 ± 112 71 238 ± 102 130 235 ± 106
No. of GP visits in the last 3 months 86 2.52 ± 2.5 84 2.49 ± 2.86 170 2.51 ± 2.68
No. of medical specialist visits in the last 3 months 87 0.98 ± 0.86 86 0.85 ± 0.88 173 0.91 ± 0.87
No. of visits to HCPs in the last 3 months 87 1.03 ± 2.22 86 1.38 ± 3.09 173 1.21 ± 2.68
No. of ED visits in the last 3 months 87 0.2 ± 0.53 87 0.17 ± 0.72 174 0.18 ± 0.63
No. of nights as hospital inpatient in the last 3 months 87 0.06 ± 0.32 87 0.09 ± 0.76 174 0.07 ± 0.58
*Higher scores indicate worse symptoms.
Higher scores indicate better function.
BPI, Brief Pain Inventory; CAQ, Committed Action Questionnaire; CPAQ, Chronic Pain Acceptance Questionnaire; CPAQ-8 AE, activity engagement subscale; CPAQ PW, pain willingness subscale; ExACT, Exercise combined with Acceptance and Commitment Therapy; GAD, General Anxiety Disorder; GAD-7, General Anxiety Disorder-7; GP, general practitioner; HCP, healthcare professional; PCS, Pain Catastrophising Scale; PHQ, Patient Health Questionnaire; PHQ-9, Patient Health Questionnaire-9; PSEQ, Pain Self-Efficacy Questionnaire; TSK, Tampa Scale of Kinesiophobia.

3.3. Outcome analyses

Unadjusted mean (SD) values and sample sizes for all continuous outcomes at baseline, postintervention, and 12-week follow-up are reported in Table 3. The results of the intention-to-treat analyses for the primary and key secondary outcomes are stated below and described in detail in Table 4. The intention-to-treat analyses for the treatment process measures are described in Table 5. Data related to satisfaction with treatment and Participant Global Impression of Change, healthcare utilisation, and physical activity are presented in Supplemental data (supplemental tables 7-12, available at https://links.lww.com/PAIN/B497). Effect sizes were interpreted as small 0.2, medium 0.5, and large 0.8, according to the guideline provided by Cohen.8 Results of the per-protocol analysis are presented in Supplemental Data, Tables 13-16 (available at https://links.lww.com/PAIN/B497).

Table 3 - Unadjusted mean (SD) values and sample sizes for all continuous outcome measures at baseline, postintervention, and 12-week follow-up.
Outcome measure (scoring range) Baseline Postintervention 12-week follow-up
Supervised exercise ExACT Supervised exercise ExACT Supervised exercise ExACT
n Mean ± SD n Mean ± SD n Mean ± SD n Mean ± SD n Mean ± SD n Mean ± SD
BPI interference score (0-10) 87 6.92 ± 1.57 87 6.86 ± 1.75 63 6.53 ± 2.04 69 6.36 ± 2.12 70 6.48 ± 2.08 69 6.30 ± 2.30
BPI composite severity score (0-10) 88 6.53 ± 1.60 87 6.17 ± 1.62 64 6.46 ± 1.71 68 5.82 ± 1.75 70 6.38 ± 1.73 69 5.92 ± 2.04
PHQ-9 (0-27) 88 14.19 ± 6.32 87 14.43 ± 6.06 62 13.61 ± 6.18 69 12.16 ± 6.13 66 11.95 ± 5.93 69 11.49 ± 6.48
GAD-7 (0-21) 88 11.45 ± 5.84 87 11.01 ± 5.6 61 10.28 ± 5.63 69 8.81 ± 5.37 68 8.71 ± 5.86 69 8.88 ± 5.36
PSEQ (0-60) 88 25.31 ± 12.27 87 23.86 ± 11.49 61 27.28 ± 11.33 69 30.68 ± 11.2 66 28.05 ± 12.9 66 29.62 ± 10.58
PCS total (0-52) 87 28.77 ± 12.17 86 32.36 ± 11.44 58 26.26 ± 13.03 64 26.5 ± 10.88 59 22.09 ± 12.90 57 24.07 ± 10.87
PCS rumination (0-16) 87 9.62 ± 4.21 87 11.07 ± 4.08 58 9.05 ± 4.77 65 8.89 ± 4.29 60 7.88 ± 4.87 58 8.62 ± 4.12
PCS magnification (0-12) 88 5.53 ± 3.34 87 6.22 ± 3.37 58 4.84 ± 3.54 64 5.44 ± 2.68 59 3.98 ± 3.17 57 4.32 ± 3.04
PCS helplessness (0-24) 88 13.53 ± 5.77 86 15.06 ± 5.62 58 12.36 ± 6.14 65 11.88 ± 5.30 59 11.05 ± 6.12 58 11.36 ± 5.07
TSK (17-68) 84 40.36 ± 7.19 86 41.88 ± 7.77 54 38.78 ± 7.52 60 38.67 ± 6.82 51 37.9 ± 8.35 59 37.92 ± 7.75
CPAQ-8 total (0-48) 88 20.49 ± 7.87 87 19.07 ± 7.92 61 21.61 ± 8.07 66 22.35 ± 6.90 65 22.42 ± 9.32 66 22.41 ± 7.55
CPAQ-8 AE (0-24) 88 12.19 ± 6.0 87 11.53 ± 5.50 62 13.31 ± 5.68 67 13.93 ± 4.88 65 12.86 ± 5.61 68 13.44 ± 4.56
CPAQ-8 PW (0-24) 88 8.3 ± 4.61 87 7.54 ± 4.98 61 8.31 ± 4.65 67 8.34 ± 4.18 67 9.61 ± 4.99 66 9.00 ± 4.68
CAQ-8 (0-48) 88 28.4 ± 7.87 87 26.89 ± 8.28 61 28.00 ± 7.59 66 29.76 ± 6.97 66 28.58 ± 8.12 67 28.07 ± 6.87
Average daily step count (Fitbit Zip) 64 7388 ± 3801 73 7481 ± 3420 46 7442 ± 3981 58 7576 ± 3787 Not measured at this time point
Distance travelled (Fitbit Zip) 64 5.14 ± 2.72 73 5.26 ± 2.51 46 5.11 ± 2.85 58 5.29 ± 2.75 Not measured at this time point
Active minutes (Fitbit Zip) 59 232 ± 112 71 238 ± 102 43 234.95 ± 117 56 231.8 ± 110.6 Not measured at this time point
No. of GP visits in the last 3 months 86 2.52 ± 2.5 84 2.49 ± 2.86 Not measured at this time point 67 2.15 ± 2.24 65 1.83 ± 1.76
No. of medical specialist visits in the last 3 months 87 0.98 ± 0.86 86 0.85 ± 0.88 Not measured at this time point 69 0.83 ± 1.54 68 0.59 ± 0.85
No. of visits to HCPs in the last 3 months 87 1.03 ± 2.22 86 1.38 ± 3.09 Not measured at this time point 67 0.97 ± 2.18 66 1.5 ± 2.69
No. of ED visits in the last 3 months 87 0.2 ± 0.53 87 0.17 ± 0.72 Not measured at this time point 68 0.21 ± 0.56 68 0.16 ± 0.54
No. of nights as hospital inpatient in the last 3 months 87 0.06 ± 0.32 87 0.09 ± 0.76 Not measured at this time point 68 0.21 ± 1.24 68 0.03 ± 0.17
BPI, Brief Pain Inventory; CAQ, Committed Action Questionnaire; CPAQ, Chronic Pain Acceptance Questionnaire; CPAQ-8 AE, activity engagement subscale; CPAQ PW; pain willingness subscale; ExACT, Exercise combined with Acceptance and Commitment Therapy; GAD, General Anxiety Disorder; GP, general practitioner; HCP, healthcare professional; PCS, Pain Catastrophising Scale; PHQ, Patient Health Questionnaire; PSEQ, Pain Self-Efficacy Questionnaire; TSK, Tampa Scale of Kinesiophobia.

Table 4 - Intention-to-treat analyses for primary outcome (BPI interference) and key secondary outcomes (BPI composite severity, PHQ-9, and GAD-7).
Outcome measure (scoring range) Baseline Postintervention P* Effect size 12-week follow-up P* Effect size
BPI interference (0-10) Supervised exercise Mean (95% CI) 6.92 (6.57 to 7.27) 6.62 (6.15 to 7.10) 6.59 (6.10 to 7.09)
∆ (95% CI) 0.29 (−0.14 to 0.73) 0.185 0.18 0.33 (−0.14 to 0.79) 0.17 0.21
ExACT Mean (95% CI) 6.86 (6.51 to 7.21) 6.4 (5.94 to 6.86) 6.36 (5.86 to 6.86)
∆ (95% CI) 0.46 (0.04 to 0.88) 0.03 0.26 0.5 (0.04 to 0.97) 0.03 0.29
Group difference −0.17 (−0.78 to 0.44) 0.58 0.10 −0.18 (−0.84 to 0.48) 0.59 0.11
BPI composite severity (0-10)§ Supervised exercise Mean (95% CI) 6.53 (6.20 to 6.87) 6.49 (6.10 to 6.87) 6.43 (6.01 to 6.85)
∆ (95% CI) 0.05 (−0.25 to 0.34) 0.75 0.03 0.10 (−0.25 to 0.45) 0.57 0.06
ExACT Mean (95% CI) 6.17 (5.83 to 6.51) 5.85 (5.46 to 6.23) 5.97 (5.55 to 6.39)
∆ (95% CI) 0.32 (0.03 to 0.61) 0.03 0.20 0.2 (−0.16 to 0.55) 0.28 0.12
Group difference (95% CI) −0.28 (−0.69 to 0.14) 0.19 0.17 −0.09 (−0.60 to 0.41) 0.71 0.06
PHQ-9 (0-27) Supervised exercise Mean (95% CI) 14.19 (12.89 to 15.50) 13.55 (12.13 to 14.96) 12.08 (10.65 to 13.50)
∆ (95% CI) 0.65 (−0.60 to 1.89) 0.31 0.10 2.12 (0.85 to 3.39) 0.001 0.34
ExACT Mean (95% CI) 14.43 (13.12 to 15.74) 12.43 (11.04 to 13.81) 11.69 (10.28 to 13.10)
∆ (95% CI) 2.00 (0.80 to 3.20) 0.001 0.33 2.74 (1.48 to 3.99) 0.000 0.45
Group difference −1.35 (−3.08 to 0.38) 0.13 0.22 −0.62 (−2.40 to 1.16) 0.49 0.10
GAD-7 (0-21) Supervised exercise Mean (95% CI) 11.46 (10.25 to 12.66) 10.56 (9.28 to 11.85) 9.47 (8.19 to 10.74)
∆ (95% CI) 0.90 (−0.24 to 2.04) 0.12 0.15 1.99 (0.99 to 2.99) 0.000 0.34
ExACT Mean (95% CI) 11.01 (9.80 to 12.22) 9.14 (7.90 to 10.38) 9.16 (7.88 to 10.43)
∆ (95% CI) 1.87 (0.78 to 2.96) 0.001 0.33 1.86 (0.86 to 2.85) 0.000 0.33
Group difference −0.97 (−2.55 to 0.60) 0.22 0.17 0.13 (−1.28 to 1.54) 0.86 0.02
*From the linear mixed model test for the time point × group interaction term. All models included an unstructured correlation between time points. A P value indicating a statistically significant difference for primary outcome P < 0.05. Bonferroni-corrected P value for secondary outcomes P < 0.002.
Cohen D computed as mean difference relative to pooled SDs (baseline SDs used in both within-group and between-group calculations).
Scores on the BPI interference subscale range from 0 to 10, with higher scores indicating worse pain interference (3).
§Scores on the BPI composite severity subscale range from 0 to 10, with higher scores indicating greater severity of pain (3).
Score of 10 or more is diagnostic of depression, and scores of 5, 10, 15, and 20 were suggested to represent mild, moderate, moderately severe, and severe depression, respectively (8).
Score of 10 or more is diagnostic of generalised anxiety, and scores of 5, 10, and 15 have been suggested to represent mild, moderate, and severe levels of anxiety, respectively (9).
∆, change from baseline. BPI, Brief Pain Inventory; CI, confidence interval; ExACT, Exercise combined with Acceptance and Commitment Therapy; GAD-7, General Anxiety Disorder-7; PHQ-9, Patient Health Questionnaire-9.

Table 5 - Intention to treat analyses for treatment process measures (PSEQ, PCS, TSK, CPAQ, CAQ).
Outcome measure (scoring range) Baseline Post intervention P* Effect size 12 week follow up P* Effect size
PSEQ (0-60)  Supervised exercise Mean (95% CI) 25.31 (22.81 to 27.81) 26.92 (24.31 to 29.54) 27.59 (24.90 to 30.27)
∆ (95% CI) −1.62 (−3.92 to 0.68) 0.17 0.13 −2.28 (−4.73 to 0.17) 0.07 0.19
ExACT Mean (95% CI) 23.86 (21.35 to 26.38) 30.06 (27.53 to 32.59) 28.95 (26.27 to 31.63)
∆ (95% CI) −6.20 (−8.41 to −3.99) 0.000 0.54 −5.09 (−7.54 to −2.64) 0.000 0.44
Group difference 4.58 (1.39−7.77) 0.005 0.39 2.81(−0.66 to 6.28) 0.112 0.24
PCS (total) (0-52)  Supervised exercise Mean (95% CI) 28.80 (26.31 to 31.29) 26.92 (24.07 to 29.78) 24.54 (21.62 to 27.46)
∆ (95% CI) 1.88 (−0.53 to 4.29) 0.13 0.15 4.26 (1.43 to 7.10) 0.003 0.35
ExACT Mean (95% CI) 32.38 (29.87 to 34.88) 26.31 (23.51 to 29.11) 23.99 (21.04 to 26.93)
∆ (95% CI) 6.07 (3.74 to 8.39) 0.000 0.53 8.39 (5.53 to 11.25) 0.000 0.73
Group difference −4.19 (−7.54 to −0.84) 0.02 0.35 −4.13 (−8.16 to −0.10) 0.045 0.35
PCS (rumination) (0-16) Supervised exercise Mean ± SD (95% CI) 9.62 (8.75 to 10.50) 9.25 (8.17 to 10.34) 8.42 (7.32 to 9.52)
∆ (95% CI) 0.37 (−0.53 to 1.27) 0.42 0.08 1.2 (0.17 to 2.24) 0.020 0.29
ExACT Mean (95% CI) 11.07 (10.19 to 11.95) 8.89 (7.83 to 9.94) 8.60 (7.49 to 9.71)
∆ (95% CI) 2.18 (1.32 to 3.05) 0.000 0.53 2.47 (1.42 to 3.52) 0.000 0.61
Group difference −1.82 (−3.06 to −0.57) 0.005 0.44 −1.27 (−2.74 to 0.21) 0.09 0.31
PCS (magnification) (0-12)  Supervised exercise Mean (95% CI) 5.53 (4.83 to 6.24) 5.01 (4.25 to 5.77) 4.39 (3.65 to 5.14)
∆ (95% CI) 0.52 (−0.19 to 1.23) 0.15 0.16 1.14 (0.43 to 1.86) 0.002 0.34
ExACT Mean (95% CI) 6.22 (5.51 to 6.93) 5.46 (4.72 to 6.21)) 4.28 (3.53 to 5.03)
∆ (95% CI) 0.76 (0.06 to 1.45) 0.03 0.26 1.94 (1.22 to 2.66) 0.000 0.58
Group difference  −0.24 (−1.23 to 0.76) 0.64 0.07 −0.8 (−1.81 to 0.21) 0.12 0.24
PCS (helplessness) (0-24) Supervised exercise Mean (95% CI) 13.53 (12.34 to 14.73) 12.65 (11.29 to 14.01) 11.62 (10.22 to 13.02)
∆ (95% CI) 0.88 (−0.38 to 2.14) 0.17 0.15 1.91 (0.41 to 3.41) 0.01 0.33
ExACT Mean (95% CI) 15.06 (13.85 to 16.27) 11.88 (10.56 to 13.20) 11.19 (9.79 to 12.59)
∆ (95% CI) 3.18 (1.96 to 4.40) 0.000 0.57 3.87 (2.36 to 5.38) 0.000 0.69
Group difference −2.29 (−4.05 to −0.54) 0.01 0.40 −1.96 (−4.09 to 0.17) 0.07 0.34
TSK (17-68) Supervised exercise Mean (95% CI) 40.28 (38.67 to 41.88) 39.39 (37.61 to 41.18) 38.61 (36.62 to 0.61)
∆ (95% CI) 0.88 (−0.62 to 2.39) 0.25 0.12 1.66 (−0.09 to 3.41) 0.06 0.23
ExACT Mean (95% CI) 41.88 (40.29 to 43.48) 39.36 (37.63 to 41.09) 38.82 (36.91 to 40.73)
∆ (95% CI) 2.52 (1.09 to 3.95) 0.001 0.32 3.06 (1.43 to 4.69) 0.000 0.39
Group difference −1.64 (−3.71 to 0.44) 0.12 0.22 −1.40 (−3.79 to 0.99) 0.25 0.19
CPAQ total (0-48) Supervised exercise Mean (95% CI) 20.49 (18.83 to 22.15) 21.61 (19.86 to 23.35) 22.66 (20.72 to 24.60)
∆ (95% CI) −1.12 (−2.74 to 0.50) 0.18 0.14 −2.17 (−3.91 to −0.43) 0.02 0.28
ExACT Mean (95% CI) 19.07 (17.40 to 20.74) 22.26 (20.55 to 23.98) 22.43 (20.50 to 24.40)
∆ (95% CI) −3.19 (−4.78 to −1.61) 0.000 0.40 −3.36 (−5.09 to −1.63) 0.000 0.42
Group difference 2.08 (−0.19 to 4.35) 0.07 0.26 1.19 (−1.27 to 3.64) 0.34 0.15
CPAQ activity engagement (0-24)  Supervised exercise Mean (95% CI) 12.19 (10.98 to 13.41) 13.27 (12.04 to 14.50) 13.03 (11.85 to 14.20)
∆ (95% CI) −1.07 (−2.24 to 0.09) 0.07 0.18 −0.83 (−1.99 to 0.32) 0.16 0.14
ExACT Mean (95% CI) 11.53 (10.31 to 12.75) 13.93 (12.73 to 15.14) 13.52 (12.36 to 14.69)
∆ (95% CI) −2.40 (−3.54 to −1.27) 0.000 0.44 −1.99 (−3.13 to −0.85) 0.001 0.36
Group difference 1.33 (−0.29 to 2.96) 0.11 0.23 1.16 (−0.46 to 2.78) 0.16 0.20
CPAQ pain willingness (0-24)  Supervised exercise Mean (95% CI) 8.30 (7.29 to 9.30) 8.31 (7.26 to 9.36) 9.63 (8.53 to 10.73)
∆ (95% CI) −0.01 (−1.04 to 1.01) 0.977 0.002 −1.34 (−2.40 to −0.27) 0.01 0.29
ExACT Mean (95% CI) 7.54 (6.53 to 8.56) 8.25 (7.26 to 9.27) 8.91 (7.81 to 10.02)
∆ (95% CI) −0.71 (−1.71 to 0.29) 0.16 0.14 −1.37 (−2.44 to −0.30) 0.01 0.28
Group difference 0.69 (−0.74 to 2.12) 0.34 0.28 0.04 (−1.47 to 1.54) 0.96 0.24
CAQ (0-48)  Supervised exercise Mean (95% CI) 28.40 (26.70 to 30.10) 27.80 (26.09 to 29.51) 28.40 (26.66 to 30.14)
∆ (95% CI) 0.60 (−1.03 to 2.23) 0.47 0.08 −0.001 (−1.75 to 1.75) 0.999 0.00
ExACT Mean (95% CI) 26.89 (25.18 to 28.59) 29.41 (27.73 to 31.09) 27.93 (26.20 to 29.67)
∆ (95% CI) −2.52 (−4.12 to −0.93) 0.002 0.3 −1.05 (−2.79 to 0.70) 0.24 0.13
Group difference 3.12 (0.84 to 5.41) 0.008 0.39 1.04 (−1.43 to 3.52) 0.41 0.13
*From the linear mixed model test for the time point × group interaction term. All models included an unstructured correlation between time points. Bonferroni-corrected P value indicating a statistically significant difference P < 0.002.
Cohen D computed as mean difference relative to pooled SDs (baseline SDs used in both within-group and between-group calculations).
∆, change from baseline; CAQ, Committed Action Questionnaire; CI, confidence interval; CPAQ, Chronic Pain Acceptance Questionnaire; ExACT, Exercise combined with Acceptance and Commitment Therapy; PCS, Pain Catastrophising Scale; PSEQ, Pain Self-Efficacy Questionnaire; TSK, Tampa Scale for Kinesiophobia.

3.3.1. Analyses for primary outcome: pain interference

No statistically significant differences in pain interference were found between the ExACT and supervised exercise groups postintervention or at 12-week follow-up (Table 4).

Statistically significant improvements in pain interference were observed within the ExACT group, both postintervention and at 12-week follow-up. There was no significant difference within the supervised exercise group postintervention or at 12-week follow-up. The mean improvements noted within both of the intervention groups did not reach the threshold for a minimal clinical important difference (1 point on the BPI interference scale).10 Small effect sizes were noted within the ExACT group postintervention and at 12-week follow-up, whereas the magnitude of the effects observed within the supervised exercise group were minimal (Fig. 2).

F2
Figure 2.:
Mean change in BPI interference score by study arm at baseline, postintervention, and 12-week follow-up based on the ITT analysis. Error bars represent 95% CIs of estimated marginal mean. BPI, Brief Pain Inventory; CI, confidence interval; ExACT, Exercise combined with Acceptance and Commitment Therapy; FU, follow-up; ITT, intention to treat.

3.3.2. Analyses for continuous secondary outcomes and treatment process measures

No statistically significant differences were found between the ExACT and supervised exercise groups for any of the secondary outcomes or treatment process measures at either time point, when assessed from the Bonferroni corrected P value < 0.002 (Table 4 and 5). These analyses are exploratory because the trial was not powered to detect significant differences in these outcomes. If the outcomes were assessed in relation to the conventional significance level of P < 0.05, the results indicate superior outcomes in favour of ExACT for pain self-efficacy and committed action postintervention and pain catastrophising at both time points. However, the magnitudes of the effects were small and do not seem to be clinically significant.10

3.3.3. Participant Satisfaction and Global Impression of Change

Overall, satisfaction levels were high postintervention, with 81.6% of all respondents reporting that they were either very satisfied or mostly satisfied with the treatment programmes. When the randomised groups were examined separately, satisfaction levels were higher in the ExACT group, with 92% of respondents reported to be very satisfied or mostly satisfied, compared with 70% in the supervised exercise group (Supplemental Data, Table 7, available at https://links.lww.com/PAIN/B497). A χ2 test showed a statistically significant association between satisfaction and randomised group, with higher levels of satisfaction reported in the ExACT group, χ2 (df = 3, n = 125)= 14.72, P =0.002. Satisfaction remained high at 12-week follow-up, and again when the groups were compared satisfaction levels remained higher in the ExACT group. However, there was no statistically significant association between satisfaction and randomised group at 12-week follow-up, χ2 (df = 3, n = 127)= 5.443, P = 0.136.

Participants' global impression of change was assessed postintervention and again at 12-week follow-up. The frequency (n) and percentage of responses are reported for each individual category, as recommended by Dworkin et al.10 A higher number of ExACT group participants reported feeling better with a noticeable change (categories 5-7) at both time points compared with participants from the supervised exercise group (Supplemental Data, Table 8-9, available at https://links.lww.com/PAIN/B497).

3.3.4. Healthcare utilisation

Healthcare utilisation was measured at baseline and at 12-week follow-up. No significant differences were observed between the ExACT and supervised exercise groups for any of the outcomes measured (Supplemental Data, Table 10, available at https://links.lww.com/PAIN/B497).

3.3.5. Physical activity measures—average daily step count, distance travelled, and active minutes

A total of 137 trial participants (78.28%) were compliant with wearing of the Fitbit activity trackers in the baseline week, before commencing the programme. Participants were considered compliant if they had worn the device for 4 of 7 days during the baseline and final week. Compliance levels were reasonably well balanced between the 2 groups, although a lower compliance rate was observed in the supervised exercise group (74.4%), compared with 83.9% in the ExACT group. The compliance rate reduced for both groups in the last week of the intervention. The rate of compliance was again lower in the supervised exercise group for the final week of the intervention (52.3%) vs 66.7% in the ExACT group (Supplemental Data, Table 12, available at https://links.lww.com/PAIN/B497).

No statistically significant differences were observed between the ExACT and supervised exercise groups in relation to any of the measures of physical activity postintervention (step count, distance travelled, or active minutes) (Supplemental data, Table 11 and Figure 2, available at https://links.lww.com/PAIN/B497). The mean difference between the ExACT and supervised exercise intervention in change from baseline to postintervention for average daily step count was 190.32 in favour of the supervised exercise intervention. Similar nonsignificant between-group differences in favour of supervised exercise were observed for distance travelled and active minutes from baseline to postintervention.

As specified in the trial protocol,5 we also sought to explore the proportion of participants who were reaching the global recommendations for physical activity for health. Although healthy adults are advised to achieve an average of 10,000 steps/day, a normative range for individuals with disability or chronic illness has been estimated at 6500 to 8500 steps/day.44 The data related to average daily step count were recategorised to quantify the number of participants reaching the lower and upper limits of the normative range of steps per day for adults with disability or chronic illness. At baseline, 57% of trial participants were observed to be reaching the lower limit of 6500 steps per day and 35% were achieving the upper limit of 8500 steps per day. Nearly 1 in 4 participants (24.82%) were achieving the global recommendation for healthy adults of 10,000 steps per day (Supplemental Data, Table 13, available at https://links.lww.com/PAIN/B497). A minimal change was observed in the proportions of participants for each of the physical activity level categories postintervention (Supplemental Data, Table 14, available at https://links.lww.com/PAIN/B497).

3.3.6. Adverse events

No adverse events attributable to participation in the trial were reported.

3.4. Treatment fidelity

Treatment fidelity of the ACT intervention was assessed by an independent psychologist who was trained and experienced in ACT for chronic pain. All 8 ACT sessions, from 1 treatment group were audio recorded, and the recordings were provided to the psychologist for review. In the absence of a gold standard ACT fidelity measurement (FM) tool and the absence of a FM tool specific to a chronic pain population, an ACT FM tool that has been modified for chronic pain was used to evaluate frequency of ACT consistent and inconsistent behaviours demonstrated by the psychologist delivering the intervention.31 The FM tool assesses therapists' behaviours in 4 areas: therapist stance, open response style, aware response style, and engaged response style, and the therapist's level of ACT consistency and inconsistency during a single treatment session is rated on a scale from 0 to 36. The assessing psychologist rated all 8 sessions of the programme, but it should be noted that sessions 1 and 2 contained a large amount of non–ACT-related content (eg, group introductions and outline of the treatment programme). The levels of ACT consistent behaviours estimated over the course of the 8-week intervention ranged from low (for session 1) to high (sessions 6-8) (mean 20.2, SD 8.13). Very low levels of ACT inconsistent behaviours were reported (mean 0.87, SD 0.83).

3.5. Sensitivity analysis: per-protocol analysis

The per-protocol analysis included only participants who attended ≥ 4 sessions of the intervention, as had been prespecified in the study protocol.5 We chose to include individuals who had attended at least 50% of the intervention in the per-protocol analysis because low attendance rates are not uncommon for these type of complex interventions for chronic pain and we anticipated that only a small proportion of participants would achieve full attendance. We considered that attendance of 4 sessions would be sufficient to have some impact on participants. One hundred fourteen participants (65.14% of total sample) were included in the per-protocol analysis and the results showed a very similar pattern to the primary intention-to-treat analysis. Per-protocol analyses for primary and secondary outcomes, treatment process measures, healthcare utilisation, and physical activity measures are reported in Supplemental Data, Table 15-18, available at https://links.lww.com/PAIN/B497.

4. Discussion

The primary result of this RCT indicated no significant difference between the ExACT and supervised exercise groups for the primary outcome: pain interference at 12-week follow-up. Few RCTs have compared ACT with an active control intervention for chronic pain, but our result is consistent with systematic reviews that have not found ACT to be superior to active controls.13,15 Our result is also in line with a systematic review investigating MBR for low back pain, which noted that the effects of MBR compared with usual care or physical interventions were modest,17 and a subsequent systematic review, which found no clinically significant differences in pain or disability outcomes in a comparison of physical, behavioural or psychologically informed, and combined interventions for nonspecific spinal pain.30 By contrast, our research group have published a systematic review comparing MBR with active physical interventions for adults with any type of chronic noncancer pain,6 which found medium to large effects in favour of MBR for disability in the short-term and long-term. However, it was recommended that the results be interpreted with caution because of low quality of evidence.

The ExACT trial result may be influenced by a number of factors that could be explored in future research. First, the participant characteristics may have had an impact on clinical outcomes. For example, a large proportion of the participants reported that they were not working. This was surprising considering the high levels of education reported. The reasons are unknown but may partly relate to the setting of the study in an area of Dublin with relatively high levels of deprivation.43 Unemployment has been identified as a predictor of worse outcomes for pain intensity, physical functioning, mental health, and social functioning after contextual CBT.12 Future research exploring factors associated with unemployment may help identify challenges that could be addressed with treatment. The baseline level of pain severity reported by participants could be another relevant factor because it was high compared with other studies that reported improvements in favour of MBR interventions.26,27 Further studies of participants with varying degrees of baseline pain severity are warranted.

The nature and mode of delivery of the trial interventions may also be relevant. Both the ExACT and supervised exercise interventions were group programmes. The studies included in our systematic review that showed medium to large effects for disability in favour of MBR6 featured interventions that included individual treatment components.9,26,27 Participants were excluded from the ExACT trial if they had attended physiotherapy or psychology in the 3 months preceding the trial, but it is possible that provision of individual sessions in addition to group programmes may result in outcomes superior to those observed in the current study. It may also be worth increasing the dosage of the intervention because it is possible that individuals who have lived with chronic pain for many years may require treatment over a longer period.

Although the RCT was powered based on the primary outcome alone, we will also discuss some of the secondary and treatment process outcomes. No significant differences were observed between the groups for any of these outcomes when the Bonferroni method was used to correct for multiple analyses. As these secondary and treatment process outcomes are exploratory, it is worth considering the effect sizes from these analyses, together with the P values without the Bonferroni correction, with statistical significance assessed from a P value < 0.05 so as not to excessively reduce the unknown and variable statistical power. Similarly, no significant differences between the ExACT and supervised exercise groups were detected for the pain severity, depression, or anxiety. Superior treatment effects of a small magnitude were observed for pain self-efficacy in favour of the ExACT intervention postintervention. It is possible that some of the ACT processes including value awareness, acceptance, and committed action may have had a positive impact on pain self-efficacy, although this association was not tested specifically. The ExACT intervention also had superior effects on pain catastrophising both postintervention and at 12-week follow-up. Multimodal treatment has previously been shown to have large effects on pain catastrophising,38 and a number of the ACT processes are very relevant for catastrophising, such as contact with the present moment, values clarification, and cognitive defusion.

The lack of significant difference observed between the ExACT and supervised exercise groups for pain acceptance was unexpected because the practice of acceptance received considerable attention throughout the ExACT intervention and was not a feature of the supervised exercise programme. The result is in contrast with a previous RCT that reported a superior effect on pain acceptance in favour of an interdisciplinary ACT intervention, compared with applied relaxation.18 The improvement in pain acceptance observed within the supervised exercise group at 12-week follow-up was not anticipated, but similar changes have been observed for control interventions in other studies.1,21,48 A between-group difference in committed action was found in favour of the ExACT group postintervention, but not at 12-week follow-up. Research related to committed action is only beginning to emerge, but the trial results seem to be in line with 2 non-RCTs that reported small improvements in committed action after interdisciplinary ACT treatment.23

Physical activity levels were unchanged in both the ExACT and supervised exercise groups. This result may relate to a lack of potential to improve because participants of both groups reported baseline levels of activity considered to be within a normative range for individuals with chronic illness.44 Another reason for the lack of change in physical activity may be the failure to use specific strategies to enhance physical activity, such as targeted use of the Fitbit trackers for goal setting, and individual feedback was not provided.

It was interesting to note that a higher proportion of ExACT group participants reported feeling better with a noticeable change compared with the supervised exercise participants. Combined Exercise and ACT group participants also reported higher levels of satisfaction with treatment. A similar finding has been reported previously in favour of ACT, compared with CBT.48

Strengths of this RCT include the successful recruitment and retention of participants and achievement of the target sample size. The trial was conducted to a high standard, with linear mixed model analyses used to minimise the impact of missing data. The similar results of the per-protocol analysis support the validity of the primary intention-to-treat analyses. The broad diagnostic profile of participants is representative of people attending hospital pain clinics for treatment of disabling chronic pain, unlike many other trials that limit inclusion to specific conditions such as low back or neck pain. The same trial physiotherapist and psychologist delivered the exercise and ACT interventions, minimising possible differences in clinician expertise and personal communication style that may confound the results.

There are a number of limitations to be acknowledged. As is generally the case with trials of complex interventions for chronic pain, it was not feasible to blind the clinicians or study participants to group allocation. A large proportion of participants did not start or complete treatment, and the withdrawal and drop-out rates were higher in the supervised exercise group, compared with the ExACT group. Attendance rates were also lower in the supervised exercise group. The reasons for this are unknown but may indicate a preference for ExACT. However, the questionnaire response rates were similar for both groups, and intention-to-treat analyses were applied to minimise risk of bias. The grouping of Patient Global Impression of Change responses had not been prespecified (categories 1-4 and 5-7) but was performed in an effort to quantify the proportion of participants who reported a noticeable improvement in their symptoms. Regarding follow-up, it would have been preferable to follow-up participants at 6 months, rather than the shorter time frame of 12 weeks, but this was not feasible within the allocated time frame for this study. Finally, it would have been preferable to perform a more comprehensive fidelity assessment of the ACT interventions. Our assessment was restricted to 1 treatment group because of limited resources. Furthermore, there are no established guidelines for the assessment of ACT fidelity. The tool we used requires further evaluation of it psychometric properties, and the utility and reliability of this tool has not been tested in a chronic pain population.31 Further research related to the assessment of ACT fidelity is required.

The ExACT trial is the first RCT to investigate the effectiveness of a combined Exercise and ACT intervention, compared with a supervised exercise programme for people with chronic pain. The overall RCT result indicated that there was no significant difference between the ExACT programme and the supervised exercise intervention for the primary outcome pain interference at 12-week follow-up. Further investigation may help to identify factors that predict a response to these types of interventions. In general, it seems important to continue to identify key processes of therapeutic change and to explore how the treatment effects for these processes can be enhanced to maximise the impact on clinical outcomes, including pain and functional outcomes for people with chronic pain.24,25

Conflict of interest statement

Lance McCracken has received payment for providing clinical training in Acceptance and Commitment Therapy, the psychological approach examined in this trial. The remaining authors have no conflict of interest to declare.

Appendix A. Supplemental digital content

Supplemental digital content associated with this article can be found online at https://links.lww.com/PAIN/B497.

Acknowledgments

The authors would like to acknowledge all of the staff in the MMUH Department of Pain Medicine and the MMUH Physiotherapy and Rheumatology Departments who supported this study. The authors thank Dr Odhran McCarthy, Psychologist, who performed the treatment fidelity assessment for the ACT intervention. The authors thank all of the trial participants and the Health Research Board who provided the funding to support the study through a Research Training Fellowship for Healthcare Professionals (Grant reference HPF-2016-1685).

References

[1]. Åkerblom S, Perrin S, Rivano Fischer M, McCracken LM. The mediating role of acceptance in multidisciplinary cognitive-behavioral therapy for chronic pain. J Pain 2015;16:606–15.
[2]. Brassington L, Ferreira NB, Yates S, Fearn J, Lanza P, Kemp K, Gillanders D. Better living with illness: a transdiagnostic acceptance and commitment therapy group intervention for chronic physical illness. J Contextual Behav Sci 2016;5:208–14.
[3]. Breivik H, Collett B, Ventafridda V, Cohen R, Gallacher D. Survey of chronic pain in Europe: prevalence, impact on daily life, and treatment. Eur J Pain 2006;10:287–333.
[4]. British Pain Society. Guidelines for Pain Management Programmes for adults. Available at: https://wwwbritishpainsocietyorg/static/uploads/resources/files/pmp2013_main_FINAL_v6pdf 2013. Last accessed October 20, 2021.
[5]. Casey MB, Smart KM, Segurado R, Hearty C, Gopal H, Lowry D, Flanagan D, McCracken L, Doody C. Exercise combined with Acceptance and Commitment Therapy (ExACT) compared to a supervised exercise programme for adults with chronic pain: study protocol for a randomised controlled trial. Trials 2018;19:1–14.
[6]. Casey MB, Smart KM, Segurado R, Doody C. Multidisciplinary-based rehabilitation (MBR) compared with active physical interventions for pain and disability in adults with chronic pain: a systematic review and meta-analysis. Clin J Pain 2020;36:874–86.
[7]. Cleeland CJ. The Brief Pain Inventory User guide. Available at: https://wwwmdandersonorg/documents/Departments-and-Divisions/Symptom-Research/BPI_UserGuidepdf 2009. Last accessed October 20, 2021.
[8]. Cohen J. Statistical Power Analysis for the Behavioral Sciences. New York: Lawrence Erlbaum Associates, 1988.
[9]. Den Hollander M, Goossens M, De Jong J, Ruijgrok J, Oosterhof J, Onghena P, Smeets R, Vlaeyen J. Expose or protect? A randomized controlled trial of exposure in vivo vs pain-contingent treatment as usual in patients with complex regional pain syndrome type 1. PAIN 2016;157:2318–29.
[10]. Dworkin RH, Turk DC, Wyrwich KW, Beaton D, Cleeland CS, Farrar JT, Haythornthwaite JA, Jensen MP, Kerns RD, Ader DN, Brandenburg N, Burke LB, Cella D, Chandler J, Cowan P, Dimitrova R, Dionne R, Hertz S, Jadad AR, Katz NP, Kehlet H, Kramer LD, Manning DC, McCormick C, McDermott MP, McQuay HJ, Patel S, Porter L, Quessy S, Rappaport BA, Rauschkolb C, Revicki DA, Rothman M, Schmader KE, Stacey BR, Stauffer JW, von Stein T, White RE, Witter J, Zavisic S. Interpreting the clinical importance of treatment outcomes in chronic pain clinical trials: IMMPACT recommendations. J Pain 2008;9:105–21.
[11]. Geneen LJ, Moore RA, Clarke C, Martin D, Colvin LA, Smith BH. Physical activity and exercise for chronic pain in adults: an overview of Cochrane Reviews. Cochrane Database Syst Rev 2017;1:CD011279.
[12]. Gilpin HR, Stahl DR, McCracken LM. A theoretically guided approach to identifying predictors of treatment outcome in Contextual Cognitive Behavioural Therapy for chronic pain. Eur J Pain 2019;23:354–66.
[13]. Hann KEJ, McCracken LM. A systematic review of randomized controlled trials of Acceptance and Commitment Therapy for adults with chronic pain: outcome domains, design quality, and efficacy. J Contextual Behav Sci 2014;3:217–27.
[14]. Hayes SC. Acceptance and commitment therapy, relational frame theory, and the third wave of behavioral and cognitive therapies. Behav Ther 2004;35:639–65.
[15]. Hughes LS, Clark J, Colclough JA, Dale E, McMillan D. Acceptance and commitment therapy (ACT) for chronic pain: a systematic review and meta-analyses. Clin J Pain 2017;33:552–68.
[16]. Kamper SJ, Maher CG, Mackay G. Global rating of change scales: a review of strengths and weaknesses and considerations for design. J Man Manip Ther 2009;17:163–70.
[17]. Kamper SJ, Apeldoorn AT, Chiarotto A, Smeets RJ, Ostelo RW, Guzman J, van Tulder MW. Multidisciplinary biopsychosocial rehabilitation for chronic low back pain: Cochrane systematic review and meta-analysis. BMJ: Br Med J 2015;350:h444.
[18]. Kemani M, Olsson G, Lekander M, Hesser H, Andersson E, Wicksell RK. Efficacy and cost-effectiveness of acceptance and commitment therapy and applied relaxation for longstanding pain: a randomized controlled trial. Clin J Pain 2015;31:1004–16.
[19]. Kroenke K, Spitzer RL, Williams JB. The PHQ-9: validity of a brief depression severity measure. J Gen Intern Med 2001;16:606–13.
[20]. Larsen DL, Attkisson CC, Hargreaves WA, Nguyen TD. Assessment of client/patient satisfaction: development of a general scale. Eval Program Plann 1979;2:197–207.
[21]. Luciano JV, Guallar JA, Aguado J, López-del-Hoyo Y, Olivan B, Magallón R, Alda M, Serrano-Blanco A, Gili M, Garcia-Campayo J. Effectiveness of group acceptance and commitment therapy for fibromyalgia: a 6-month randomized controlled trial (EFFIGACT study). PAIN 2014;155:693–702.
[22]. McCracken LM, Chilcot J, Norton S. Further development in the assessment of psychological flexibility: a shortened Committed Action Questionnaire (CAQ-8): committed action in chronic pain. Eur J Pain 2015;19:677–85.
[23]. McCracken LM, Davies M, Scott W, Paroli M, Harris S, Sanderson K. Can a psychologically based treatment help people to live with chronic pain when they are seeking a procedure to reduce it? Pain Med 2015;16:451–9.
[24]. McCracken LM. Necessary components of psychological treatment for chronic pain: more packages for groups or process-based therapy for individuals? Eur J Pain 2020;24:1001–2.
[25]. McCracken LM. Beyond therapy types: mindful self-compassion and the future of process-based therapy for chronic pain. Eur J Pain 2021;25:729–30.
[26]. Monticone M, Ferrante S, Rocca B, Baiardi P, dal Farra F, Foti C. Effect of a long-lasting multidisciplinary program on disability and fear-avoidance behaviors in patients with chronic low back pain: results of a randomized controlled trial. Clin J Pain 2013;29:929–38.
[27]. Monticone M, Ambrosini E, Rocca B, Magni S, Brivio F, Ferrante S. A multidisciplinary rehabilitation programme improves disability, kinesiophobia and walking ability in subjects with chronic low back pain: results of a randomised controlled pilot study. Eur Spine J 2014;23:2105–13.
[28]. National Institute for Health and Care Excellence (NICE). NICE Guideline: Chronic Pain. Available at: https://wwwniceorguk/guidance/ng193. Last accessed October 20, 2021.
[29]. Nicholas MK. The pain self-efficacy questionnaire: taking pain into account. Eur J Pain 2007;11:153–63.
[30]. O'Keeffe M, Purtill H, Kennedy N, Conneely M, Hurley J, O'Sullivan P, Dankaerts W, O'Sullivan K. Comparative effectiveness of conservative interventions for nonspecific chronic spinal pain: physical, behavioral/psychologically informed, or combined? A systematic review and meta-analysis. J Pain 2016;17:755–74.
[31]. O'Neill L, Latchford G, McCracken LM, Graham CD. The development of the Acceptance and Commitment Therapy Fidelity Measure (ACT-FM): a delphi study and field test. J Contextual Behav Sci 2019;14:111–18.
[32]. Peters SAE, Bots ML, den Ruijter HM, Palmer MK, Grobbee DE, Crouse JR, O'Leary DH, Evans GW, Raichlen JS, Moons KGM, Koffijberg H. , METEOR study group. Multiple imputation of missing repeated outcome measurements did not add to linear mixed-effects models. J Clin Epidemiol 2012;65:686–95.
[33]. Raftery MN, Sarma K, Murphy AW, De la Harpe D, Normand C, McGuire BE. Chronic pain in the republic of Ireland—community prevalence, psychosocial profile and predictors of pain-related disability: results from the prevalence, impact and cost of chronic pain (PRIME) study, Part 1. PAIN 2011;152:1096–103.
[34]. Raftery MN, Ryan P, Normand C, Murphy AW, de la Harpe D, McGuire BE. The economic cost of chronic noncancer pain in Ireland: results from the PRIME study, Part 2. J Pain 2012;13:139–45.
[35]. Roelofs J, Sluiter JK, Frings-Dresen MHW, Goossens M, Thibault P, Boersma K, Vlaeyen JWS. Fear of movement and (re)injury in chronic musculoskeletal pain: evidence for an invariant two-factor model of the Tampa Scale for Kinesiophobia across pain diagnoses and Dutch, Swedish, and Canadian samples. PAIN 2007;131:181–90.
[36]. Rovner GS, Arestedt K, Gerdle B, Borsbo B, McCracken LM. Psychometric properties of the 8-item chronic pain acceptance questionnaire (CPAQ-8) in a Swedish chronic pain cohort. J Rehabil Med 2014;46:73–80.
[37]. Schulz KF, Altman DG, Moher D; CONSORT Group. CONSORT 2010 Statement: updated guidelines for reporting parallel group randomised trials. BMC Med 2010;8:18.
[38]. Schütze R, Rees C, Smith A, Slater H, Campbell JM, O'Sullivan P. How can we best reduce pain catastrophizing in adults with chronic noncancer pain? A systematic review and meta-analysis. J Pain 2018;19:233–56.
[39]. Scottish Intercollegiate Guidelines Network (SIGN). Management of chronic pain. SIGN publication no. 136. Available at: http://www.sign.ac.uk.Last accessed October 20,2021./
[40]. Sealed Envelope. Randomisation and Online Databases for Clinical Trials. Available at: www.sealedenvelope.com. Last accessed October 20, 2021.
[41]. Spitzer RL, Kroenke K, Williams JB, Lowe B. A brief measure for assessing generalized anxiety disorder: the GAD-7. Arch Intern Med 2006;166:1092–7.
[42]. Sullivan M, Bishop S, Pivik J. The pain catastrophizing scale: development and validation. Psychol Assess 1995;7:524–32.
[43]. Teljeur C, Darker C, Barry J, O'Dowd T. The Trinity National Deprivation Index for Health & Health Services Research 2016. Dublin: Trinity College, 2019.
[44]. Tudor-Locke C, Craig C, Aoyagi Y, Bell R, Croteau K, De Bourdeaudhuij I, Ewald B, Gardner A, Hatano Y, Lutes L, Matsudo S, Ramirez-Marrero F, Rogers L, Rowe D, Schmidt M, Tully M, Blair S. How many steps/day are enough? For older adults and special populations. Int J Behav Nutr Phys Activity 2011;8:80.
[45]. Twisk J, de Boer M, de Vente W, Heymans M. Multiple imputation of missing values was not necessary before performing a longitudinal mixed-model analysis. J Clin Epidemiol 2013;66:1022–8.
[46]. van Middelkoop M, Rubinstein SM, Verhagen AP, Ostelo RW, Koes BW, van Tulder MW. Exercise therapy for chronic nonspecific low-back pain. Best Pract Res Clin Rheumatol 2010;24:193–204.
[47]. Veehof MM, Trompetter HR, Bohlmeijer ET, Schreurs KM. Acceptance- and mindfulness-based interventions for the treatment of chronic pain: a meta-analytic review. Cogn Behav Ther 2016;45:5–31.
[48]. Wetherell JL, Afari N, Rutledge T, Sorrell JT, Stoddard JA, Petkus AJ, Solomon BC, Lehman DH, Liu L, Lang AJ, Hampton Atkinson J. A randomized, controlled trial of acceptance and commitment therapy and cognitive-behavioral therapy for chronic pain. PAIN 2011;152:2098–107.
[49]. Williams AC, Eccleston C, Morley S. Psychological therapies for the management of chronic pain (excluding headache) in adults. Cochrane Database Syst Rev 2012;11:Cd007407.
[50]. Williams ACC, Fisher E, Hearn L, Eccleston C. Psychological therapies for the management of chronic pain (excluding headache) in adults. Cochrane Database Syst Rev 2020;8:CD007407.
Keywords:

Exercise; Acceptance and Commitment Therapy; Pain management programme; Chronic pain; Randomised controlled trial

Supplemental Digital Content

© 2021 International Association for the Study of Pain