The Journal of Pain, Vol 21, No 1–2 (January/February), 2020: pp 71−81

Available online at www.jpain.org and www.sciencedirect.com

Physical Therapy Informed by Acceptance and

Commitment Therapy (PACT) Versus Usual Care Physical
Therapy for Adults With Chronic Low Back Pain:
A Randomized Controlled Trial
Emma Godfrey,*,x Vari Wileman,* Melissa Galea Holmes,* Lance M. McCracken,*
Sam Norton,* Rona Moss-Morris,* Sandra Noonan,y Massimo Barcellona,z and
Duncan Critchleyx
*
Health Psychology Section, Department of Psychology, Institute of Psychiatry, Psychology and Neuroscience, King’s College
London, London, UK, yGuy’s and St Thomas’ Hospital NHS, St Thomas’ Hospital, London, UK, zKing’s College Hospital NHS
Foundation Trust, London, UK, xDepartment of Physiotherapy, School of Population Health and Environmental Sciences, Faculty
of Life Sciences and Medicine, King’s College London, London, UK

Abstract: Chronic low back pain (CLBP) is a major cause of global disability and improving manage-
ment is essential. Acceptance and commitment therapy (ACT) is a promising treatment for chronic
pain but has not been modified for physical therapy. This randomized controlled trial (RCT) compared
physical therapy informed by ACT (PACT) against standard care physical therapy for patients with
CLBP. Patients with CLBP (duration ≥12 weeks, mean 3 years) were recruited from physical therapy
clinics in 4 UK public hospitals. The Roland-Morris Disability Questionnaire (RMDQ) at 3 months’ post-
randomization was the primary outcome. Two hundred forty-eight participants (59% female, mean
age = 48) were recruited and 219 (88.3%) completed measures at 3 and/or 12 months’ follow-up. At
3 months, PACT participants reported better outcomes for disability (RMDQ mean difference = 1.07,
p = .037, 95% CI = 2.08 to .07, d = .2), Patient Specific Functioning (p = .008), SF12 physical health
(p = .032), and treatment credibility (p < .001). At 12 months’ follow-up, there were no significant dif-
ferences between groups. PACT was acceptable to patients and clinicians and feasible to deliver.
Physical therapists incorporated psychological principles successfully and treatment was delivered
with high (≥80%) fidelity. Our results may inform the management of CLBP, with potential benefits
for patients, health care providers, and society.
Perspective: Psychologically informed physical therapy has great potential but there are chal-
lenges in implementation. The training and support included in the PACT trial enabled the interven-
tion to be delivered as planned. This successfully reduced disability in the short but not long term.
Findings could inform physical therapists’ treatment of CLBP.
© 2020 THE AUTHORS. Published by Elsevier Inc. on behalf of the U.S Association for%S1526-5900(20)
X0009-5%202001/02%21%1%2%71%81%ELE\gdef\pts@issuepubyear{2020}the Study of Pain. This is
an open access article under the CC BY-NC-ND license. (http://creativecommons.org/licenses/by/4.0/)
Key words: Chronic low back pain, physical therapy, acceptance and commitment therapy, random-
ized controlled trial.

Received January 17, 2019; Revised April 17, 2019; Accepted May 24, Address reprint requests to Emma Godfrey, MA (Oxon), PhD, Health Psy-
2019. chology Section, Department of Psychology, Institute of Psychiatry, Psy-
This paper presents independent research funded by the National Insti- chology and Neuroscience, King’s College London, 5th Floor Bermondsey,
tute for Health Research (NIHR) under its Research for Patient Benefit Guy’s Campus, London SE1 9RT, UK. E-mail: [email protected]
(RfPB) program (grant reference number PB-PG-1112-29055). The views 1526-5900/$36.00
expressed are those of the authors and not necessarily those of the NIHR © 2020 THE AUTHORS. Published by Elsevier Inc. on behalf of the U.S
or the Department of Health and Social Care. Association forthe Study of Pain. This is an open access article under the
The researchers were completely independent from the funders and CC BY-NC-ND license. (http://creativecommons.org/licenses/by/4.0/)
there were no conflicts of interest, relationships, or activities that have https://doi.org/10.1016/j.jpain.2019.05.012
influenced the submitted work.
Trial registration: ISRCTN95392287.

71
72 The Journal of Pain RCT of Physical Therapy Informed by ACT for CLBP

L
ow back pain is the leading cause of global disabil- ACT for CLBP delivered by psychologists found patients
ity9,18,25 and urgently requires better manage- referred for physical therapy were somewhat resistant to
ment.9 Eighty percent of the adult population seeing a psychologist and consequently recommended
experiences a significant episode of disabling low back combining ACT with physical therapy.39 A mixed method
pain over their life. Most people recover, however, 10 to study reported challenges and opportunities to embed-
15% go on to develop chronic low back pain (CLBP), ding ACT within pain rehabilitations settings that include
defined as pain lasting over 12 weeks.11 Ninety percent physical therapists.3 We have developed a brief physical
of people with CLBP have a non-specific problem with no therapist-delivered intervention, theoretically under-
clearly identifiable cause for their pain.17 CLBP is increas- pinned by ACT, called PACT (physical therapy informed
ing in prevalence and is globally the second most fre- by ACT).19 The main objective of this trial was to evaluate
quent reason for time off work.18,25 Total costs the efficacy of PACT on functioning at the primary end
associated with back pain in the United States are esti- point of 3 months’ follow-up, compared with standard,
mated at between $100 and $200 billion per year, one- usual care (UC) physical therapy.
third due to healthcare costs, with the remainder due to
lost wages and lower productivity.29 It is a complex condi-
tion associated with psychological comorbidities, such as Methods
sleep disorders, anxiety, and depression.20 CLBP is often
ineffectively managed17 and thus debilitating for Study Design and Participants
patients, challenging for healthcare providers, and costly A phase II, assessor blind, 2-armed, parallel group,
for society.9,25 multicenter randomized controlled trial compared the
Many people with CLBP are referred for physical therapy efficacy of PACT with UC physical therapy treatment for
but within the range of treatments used by physical thera- patients with CLBP. Participants were recruited from sec-
pists, there is little consensus about which are the most ondary care physical therapy clinics in 3 UK NHS (Public)
effective and cost effective.13 Trials have shown only mod- Hospital trusts in London UK (Guy’s, St Thomas’, and
est improvements in pain and disability following usual Kings College Hospitals) and 1 in suburban/rural south
physical therapy treatment.2 Self-management programs east England (Ashford and St Peter’s Hospital). The trial
can be effective for people with CLBP14 and individualized received full Research Ethics Committee (REC) approval
treatment may facilitate better self-care.13 Recent guide- (National Research Ethics Committee South Central -
lines promote a combined psychological/physical approach Berkshire; 14/SC/0277) and conformed to current guide-
if previous treatments have proved ineffective or where lines for ethical research. The trial was registered pro-
there is a medium to high risk of chronicity.13,36 Psychologi- spectively: ISRCTN95392287.
cally informed practice is proposed as a middle way, inte-
grating traditional biomechanical and impairment-focused
practice with cognitive behavioral approaches.32 Interest in Participants
this approach is growing, however, many questions remain, Eligible patients were adults (aged ≥18 years), with
such as how much treatment is required and whether it can nonspecific CLBP with or without associated leg pain, of
be delivered with adequate fidelity.30 Physical therapists greater than 12 weeks’ duration, and reporting a score of
frequently report a lack of confidence in using psychologi- ≥3 points on the Roland Morris Disability Questionnaire
cal techniques successfully1 and may have difficulty identi- (RMDQ).41 Potential participants required a good under-
fying psychological factors associated with CLBP.8 Recent standing of spoken and written English to complete trial
reviews have concluded that this could be rectified with data collection and participate in the PACT program. Peo-
additional training and support.21 Other findings have sug- ple who had prior treatment from multidisciplinary CBT
gested patients with CLBP want to discuss personal issues pain management at any time and/or other physical ther-
with their physical therapist,28 although many physical apy treatment in the previous 6 months, or injection ther-
therapists perceive they lack the skills or confidence to apy within the last 3 months, were excluded. People with
address these concerns.1 This highlights the need to specific spinal pathology were excluded, as were people
develop psychologically informed interventions that are with severe psychiatric illness and/or current drug or alco-
suitable for physical therapists and the training to help hol misuse, as these issues require different treatment pri-
them provide it. orities. Potential participants referred to outpatient
Cognitive behavioral therapy (CBT) is an effective inter- physical therapy were identified over an 18-month period
vention for CLBP40 although it remains challenging to by physical therapists from each hospital center at their
implement approaches to physical therapy that are based triage sessions. They were provided with written and ver-
on CBT.7 Acceptance and commitment therapy (ACT), a bal information about the PACT trial and invited to par-
newer third wave CBT, has a good evidence base in the ticipate. Interested participants were then contacted by
treatment of chronic pain18,27 but its underlying princi- the research associate to screen for eligibility. All partici-
ples of psychological flexibility have not been applied to pants gave written informed consent before taking part.
physical therapy. ACT focuses on improving functioning,
rather than reducing pain, using acceptance, mindfulness
strategies, and values-based action.23,34 This approach is Randomization and Masking
particularly suitable for CLBP, as a focus on symptom Participants were randomized to receive either PACT or
reduction is frequently counter-productive.38 A trial of UC physical therapy. Random allocation to the 2 groups
Godfrey et al The Journal of Pain 73
employed random block sizes stratified by recruiting cen- clinicians. All treatment in the trial took place in the
ter (Guy’s & St Thomas’, King’s College, and Ashford & St physical therapy clinics based at the participating hospi-
Peter’s Hospitals); implemented via the King’s College tals. Training the physical therapists effectively in PACT
London Clinical Trials Unit online system, with emails treatment was an integral part of the study. To assess
generated automatically and sent to relevant physical fidelity, a randomly selected sample of 20% of the audio-
therapy staff at study sites. Face-to-face treatment meant recorded PACT sessions was rated by 2 trained, indepen-
it was not possible to blind participants or the physical dent assessors. The randomization was stratified by ses-
therapists delivering the interventions. However, the sion (initial face-to-face, 2-week face-to-face, 1-month
research associate conducting outcome assessment and telephone call) and physical therapist to ensure at least
the trial statistician analyzing the data were blinded to 1 session per physical therapist was assessed.
group allocation. No hypothesis was proposed to partici-
pants about the superiority of either treatment and sepa-
rate groups of clinicians delivered PACT and UC physical Outcomes
therapy to avoid contamination. Self-reported questionnaires were completed by
patients at baseline, 3 (primary end point) and
12 months, either online or via postal questionnaires, to
Procedures avoid any influence of the study team on the responses.
PACT was a brief physical therapy intervention, The primary outcome was patient-reported functioning
guided by principles of ACT, designed to promote self- at 3 months, assessed with the RMDQ.42 The RMDQ is
management. PACT consisted of 3 individual treatment a widely used well-validated measure with good reli-
sessions as follows: two 60-minute face-to-face sessions ability, where a 2- to 3-point change from baseline is
2 weeks apart conducted in a private room, plus one considered clinically meaningful.41 Demographic data
20-minute telephone call 1 month later. Treatment collected at baseline included: age, gender, ethnicity,
included an initial physical assessment with feedback, marital and work status, and educational attainment.
identification of value-based goals, individualized phys- Secondary outcomes included all core domains recom-
ical exercise prescription, addressing barriers and facili- mended in chronic pain research (IMMPACT recommen-
tators to self-management, and skills training to dations).15 Secondary outcome measures were: the
promote psychological flexibility. It excluded manual Patient Health Questionnaire-9 (PHQ-9),31 to assess
therapy. Total contact time was designed to be similar depression; the Generalized Anxiety Disorder-7 (GAD-
to the average amount of time patients with CLBP 7),44 to assess anxiety; the Patient-Specific Functional
receive as part of UC physical therapy treatment, as Scale (PSFS)45 and Work and Social Adjustment Scale
reported in UK RCTs for CLBP where UC physical therapy (WSAS)35 to assess functioning; a life satisfaction scale;
was used as the control arm.22 The aim was to maximize and a pain numeric analogue scale to assess pain sever-
the potential for a treatment effect within a timeframe ity. Global Improvement,4 Outcome Satisfaction,26 and
that was similar to the contact received, on average, in Treatment Credibility5 questionnaires were completed
standard physical therapy, as this was considered a feasi- at both follow-ups. Process measures were chosen with
ble way to ensure eventual implementation and cost the intention to stay theoretically clear without redun-
effectiveness. However, PACT altered the context, con- dancy, as well as the need to maintain reasonable par-
tent and duration of physical therapy treatment, so that ticipant burden from the assessment. Process variables
it was delivered in fewer but longer sessions compared included the Chronic Pain Acceptance Questionnaire-8
to usual care in the UK. Further details of PACT treat- (CPAQ-8)16; and Committed Action Questionnaire-8
ment are reported in the protocol paper.19 (CAQ-8)33; as well as the Pain self-efficacy Questionnaire
Eight experienced (Band 6 and 7) physical therapists (PSEQ).37 Nested qualitative studies were completed
received a bespoke training package, including a manual with 20 PACT patients and all PACT physical therapists
and 2-day face-to-face training program, followed by on- (reported elsewhere). In addition, proposed therapeutic
going monthly group supervision from a clinical/health mechanisms of action (process variables)16,33 were
psychologist and a physical therapist. Differences in assessed. Three bespoke treatment fidelity measures,
boundaries between psychologists and physical thera- 1 for each PACT session, were developed to appraise
pists were carefully communicated during training, as physical therapists’ adherence to the PACT intervention.
this was not designed to alter these boundaries. A A cost-consequences estimation of the economic impact
patient manual individualized to patient needs was pro- of the interventions on CLBP was completed using
vided during the first session. UC physical therapy was patient data from 2 health-related quality of life
provided by physical therapists (Bands 5−8) employed in measures, the EQ-5D-5L,24 and MOS Short Form-12v2
the Public Hospitals and comprised any treatment consid- (SF-12)47 Questionnaires. Serious adverse events were
ered suitable by the treating physical therapist, including reviewed by the chief investigator and reported to an
individual physical therapy and/or back rehabilitation independent trial steering committee for consideration.
classes, dynamic control classes, manual therapy, and
hydrotherapy. All PACT sessions were audio recorded to
check treatment fidelity. Attendance at UC sessions was Patient Involvement
documented to record volume (duration and frequency) The PACT study was developed with contributions
and components (1:1, class) of UC physical therapy by from 4 dedicated patient representatives, who were
74 The Journal of Pain RCT of Physical Therapy Informed by ACT for CLBP
recruited from local physical therapy services and Results
included 2 participants from the proof of concept feasibil-
ity study. Their contributions included the development of Participant Characteristics
PACT treatment components and materials to determine Between November 2014 and March 2016, 660
appropriateness of content, language and format; feed- patients were informed about the study by their physi-
back on key documents, such as patient information cal therapist and 478 (72%) agreed to be screened for
sheets and consent forms; piloting of assessments, includ- eligibility. Two hundred forty-eight (518%) met eligi-
ing questionnaire content and delivery via an online ques- bility criteria and consented to participate and so were
tionnaire database, to determine acceptability, length of randomized to receive either PACT (n = 124) or UC physi-
surveys and estimated completion time, and resonance of cal therapy (n = 124). Of 124 people randomized to
items within surveys to check relevance and acceptability. PACT, 17 (14%) received no treatment and 4 (3%) had
One patient was a coapplicant on the grant funding the UC physical therapy. Of 124 randomized to UC physical
trial and 2 patients were patient and public involvement therapy, 30 (24%) received no treatment and 2 (16%)
representatives on the Trial Steering Committee to ensure patients had PACT. Administrative delay and some con-
it addressed issues relevant to service users. fusion about attending appointments that had already
been allocated led to a few people inadvertently receiv-
ing the wrong treatment. Overall, 204 patients (83%)
Sample Size completed follow-up assessments at 3 months and 181
The trial was designed to detect a standardized mean (73%) at 12 months. In total, 219 patients (88.3%) pro-
difference of .4 in the primary outcome (RMDQ; 5% sig- vided data on at least 1 follow-up occasion and were
nificance, 80% power) assuming attrition of 20%. This retained for the intention-to-treat analysis, irrespective
difference equates to a 3-point difference between of whether they received treatment. Of those receiving
groups (assuming the standard deviation of the RMDQ PACT, 23 had 1 face-to-face session, 14 had 2 face-to-
is 7.4, as suggested by our small feasibility study and face sessions, and 66 received both face-to-face sessions
previous research in a similar population12 where a 2- to plus the telephone session. Overall, 103 patients (83%)
3-point difference in the RMDQ score is considered clini- completed at least 1 session of PACT. Of those allocated
cally important.41 We calculated that in total 240 partici- to UC physical therapy, 92 (74%) received UC treatment
pants needed to be randomized. and they all had at least 1 face-to-face session with a
physical therapist, with the majority referred to some
form of group-based intervention after this (eg, back
Statistical Analysis rehabilitation or hydrotherapy classes). On average, par-
Data were analyzed using Stata version 14.1 statistical ticipants in UC attended 3 hours of physical therapy (eg,
software. Estimates of treatment effect at the 3-month three 30 minutes 1:1 sessions and one and a half 60-min-
and 12 months’ follow-up followed the intention-to- ute classes), compared to 2 hours treatment in PACT.
treat principle. Between-group differences (treatment There were no patients who were withdrawn or opted
effect) were estimated for the primary outcome (RMDQ) to withdraw from the trial. Patient flow through the
at the postintervention assessments. Estimates of treat- study is presented in Fig. 1.
ment effect at the 3- and 12 months’ postrandomization Baseline demographic and clinical characteristics by
follow-up assessments were based on adjusted mean dif- treatment group are presented in Tables 1 and 2. Fifty-
ferences using linear-mixed models following the inten- nine percent of participants were female, 59% described
tion-to-treat principle.48 A 3-level model was estimated their ethnicity as white, and participants’ average age
including random effects for the patient to account for was 48 years. Our participants had RMDQ mean scores
repeated assessment over time and a random effect for that are typical of people with CLBP seeking physical
physical therapist to account for partial-clustering of therapy.12 Patients in our sample were in the mid-range
patients by physical therapist in the PACT arm. Covariates of pain intensity scores and many were experiencing
in the model included an indicator variable for group mild depression and anxiety symptoms, but these were
assignment, an indicator for follow-up time, an interac- below the level where treatment should be considered.
tion term for group by time, the baseline level of the out- There was good variability across the range of scores at
come variable and indicator variables for center, as this baseline, with sound balance across the groups achieved
was a stratification factor in the randomization. Residual by randomization. At the baseline assessment, 8 individu-
diagnostics indicated heteroscedasticity for RMDQ. Stan- als in the PACT arm and 6 patients in the UC arm had a
dard errors that are robust to violations of the normality RMDQ score <3 (although they had scored ≥3 during
and homoscedasticity assumption were estimated by screening and so were eligible). Of these, 6 patients in
bootstrapping with 1,000 replications. Treatment effects the PACT arm and 3 patients in the UC arm were retained
were converted into standardized mean differences as in the intention-to-treat analysis as they provided data.
Cohen’s d to allow comparison of effect sizes across out-
comes. Costs associated with delivering PACT and UC
were estimated using a combination of actual resource Primary Outcome
used, derived from logs kept by the physical therapists Twenty-nine patients provided no postbaseline data
and NHS Executive reference costs for 2015/16. and so were not included in the analysis sample. We
Godfrey et al The Journal of Pain 75

Figure 1. Consort Flow diagram.

included careful sensitivity analysis to determine RMDQ is considered clinically meaningful.41 The inten-
whether excluding these patients impacted on the tion-to-treat adjusted mean difference between groups
results. There was a trend for those who were not indicated people who received PACT reported signifi-
retained for the intention-to-treat analysis or not com- cantly better functioning (RMDQ) at the primary end
pleting 12 months’ follow-up to be younger, male, point of 3 months than those receiving UC (mean differ-
unmarried, have less education, and to report worse ence = 1.07, P = .037, 95% CI = 2.08 to 007, d = .2).
health at baseline. However, only the difference in age Clinically important reductions in RMDQ levels were
for those completing the 12 months’ follow-up was sta- maintained at 12 months in the PACT group, although
tistically significant, with younger patients more likely the intention-to-treat adjusted mean difference com-
to be lost to follow-up. pared to UC was reduced and nonsignificant (mean dif-
Fig. 2 shows the treatment effects based on the inten- ference .38, P = .52, 95% CI = 1.54 to .78, d = .1).
tion to treat sample. The mean reduction in RMDQ score Sensitivity analysis with the per-protocol sample (only
from baseline to 3 months in the PACT group was 3.4, included those meeting the inclusion criteria at base-
compared to 2.1 in the UC group (Table 3, Supplemen- line, ie, RMDQ ≥3 and receiving at least 1 session of
tary Materials), where a change of 2 to 3 units in the PACT or UC) and using baseline observation carried
76 The Journal of Pain RCT of Physical Therapy Informed by ACT for CLBP
Baseline Continuous Demographic and Clinical Characteristics by Treatment Group, Ran-
Table 1.
domized Sample
USUAL CARE PACT TOTAL
(N = 124) (N = 124) (N = 248)

N MEAN SD N MEAN SD N MEAN SD

Age 124 47.5 14.0 124 48.4 14.6 248 47.9 14.3
BMI 112 29.1 5.2 113 28.6 5.9 225 28.9 5.6
RMDQ*,41 124 10.8 5.8 124 10.7 5.7 248 10.7 5.7
Pain 123 6.1 1.9 123 6.1 2.1 246 6.1 2.0
PSFS45 122 4.7 2.3 120 4.6 2.3 242 4.7 2.3
WSAS35 124 16.7 9.3 124 17.2 9.5 248 16.9 9.4
Life satisfaction 122 5.9 2.6 120 5.8 2.6 242 5.9 2.6
PHQ931 124 7.4 5.7 124 7.6 6.2 248 7.5 5.9
GAD744 124 6.6 5.6 124 6.3 5.5 248 6.4 5.5
CPAQ16 123 24.3 8.6 124 25.3 8.3 247 24.8 8.5
CAQ35 123 32.3 8.8 124 30.6 8.7 247 31.5 8.8
PSEQ37 123 36.5 13.5 124 37.6 14.7 247 37.1 14.1
SF12 physical47 123 37.3 7.9 122 38.3 8.7 245 37.8 8.3
SF12 mental47 123 46.8 10.6 122 46.1 10.5 245 46.5 10.6

BMI, body mass index.

*Measures reference numbers in superscript.

Baseline Categorical Demographic and Clinical Characteristics by Treatment Group, Ran-

Table 2.
domized Sample
USUAL CARE PACT TOTAL
(N = 124) (N = 124) N = 248

N % N % N %

Gender Male 53 42.7 48 38.7 101 40.7

Female 71 57.3 76 61.3 147 59.3
Total 124 100.0 124 100.0 248 100.0
Ethnicity White 74 59.7 72 58.1 146 58.9
Mixed 7 5.6 7 5.6 14 5.6
Asian 9 7.3 10 8.1 19 7.7
Black 32 25.8 31 25.0 63 25.4
Other/unknown 2 1.6 4 3.2 6 2.4
Total 124 100.0 124 100.0 248 100.0
Marital status Unmarried 35 28.5 32 26.0 67 27.2
Married/partner 64 52.0 63 51.2 127 51.6
Separated/divorced/widowed 24 19.5 28 22.8 52 21.1
Total 123 100.0 123 100.0 246 100.0
Education No qualifications 17 13.8 24 19.7 41 16.7
High school diploma equivalent 34 27.6 36 29.5 70 28.6
AP equivalent 27 22.0 24 19.7 51 20.8
Degree/equivalent 45 36.6 38 31.1 83 33.9
Total 123 100.0 122 100.0 245 100.0
Work status Employed F/T 59 48.4 50 41.3 109 44.9
Employed P/T 28 23.0 31 25.6 59 24.3
Unemployed 17 13.9 17 14.0 34 14.0
Retired 18 14.8 23 19.0 41 16.9
Total 122 100.0 121 100.0 243 100.0

forward to impute missing data were carried out. These Secondary Outcomes and Process
confirmed the robustness of the intention-to-treat estimate Variables
with estimates of the treatment effect of 1.43 (P = .008) Participants who received PACT rated their treatment as
and .85 (P = .041), respectively (Table 4, Supplementary more credible compared then those receiving UC physical
Materials).
Godfrey et al The Journal of Pain 77

Figure 2. Standardized treatment effect sizes with 95% confidence intervals.

therapy at 3 months. PACT patients also reported better 40. This was as expected after only 2 days’ training in
outcome on the Patient-Specific Functional Scale, SF-12 this ACT informed physical therapy treatment.
physical health scale and Work and Social Adjustment
Scale at 3, but not 12 months (supplementary Materials).
No group differences were observed for measures of pain, Cost Consequences
mood, self-efficacy, or the ACT process variables (accep- The cost-consequences analysis revealed the total
tance [CPAQ-8] and committed action [CAQ-8]) at 3 or 12 CLBP-associated costs in the PACT arm were £19,776, or
months. Twenty-one trial participants reported adverse £193.88 per patient, compared to £20,286 or £220.50
events, 9 from PACT and 12 from UC. The Trial Steering per patient, in the UC physical therapy arm. However,
Committee concluded that no adverse events reported by PACT had additional one-off training costs of £11,958.
patients were related to treatment. UK NHS (public health service) resource use across the
3-time points was very comparable between the 2
groups. Over the 12-month follow-up period, resource
Fidelity utilization was similar between groups, with direct NHS
Eight physical therapists delivered PACT (mean age healthcare costs accounting for 25% (£151,345.77/
33, range = 24−44 years; 5 female). Bespoke fidelity £595,821.07) of total costs, while 13% (£77,673.00) was
measures were developed for this trial, including all the attributable to private costs paid for by the patient and
elements of treatment that were expected to be deliv- 62% (£366,802.30) to societal costs, such as time off
ered in each session. Seventy-two (20%) audio tapes work. Direct NHS healthcare costs reduced in both
were rated by independent assessors. Prior to the cali- groups over time and were £98.79 per person cheaper
bration of scores, overall agreement between raters was at 12 months’ follow-up in UC and £104.63 cheaper per
85% (474/560 decisions; 95% CI = 81−88%), when rating person in PACT. Costs associated with absenteeism fell
whether treatment elements were fully completed, par- from £597.25 to £223.63 per person in UC and from
tially completed, or not completed. Treatment fidelity £447.72 to £244.43 per person in PACT.
was calculated according to whether a minimum of
80% of treatment elements were rated as being com-
pleted/partially completed by the physical therapist, Discussion
from calibrated total scores for each individual session. This is the first trial to test the efficacy of an ACT-
The results confirmed physical therapist adherence to informed physical therapist-delivered intervention for
the PACT intervention was high,6 with overall 88% people with CLBP. PACT significantly improved partic-
(95% CI = 78−94%) treatment fidelity achieved across ipants’ back pain disability at the primary end point of
sessions (session 1: 97%, 95% CI = 84−100%; session 2: 3 months’ follow-up compared to UC physical therapy,
81%, 95% CI = 62−94%; session 3: 77%, 95% CI = 46 although effect sizes were small and not sustained at
−95%). Fidelity assessment also revealed that only a few 12 months. PACT participants achieved a clinically mean-
core ACT methods were delivered overall, with average ingful reduction of over 3 points on the RMDQ at both 3
ACT fidelity across all 72 sessions scored as 16.4 out of and 12 months compared to baseline levels; UC also
78 The Journal of Pain RCT of Physical Therapy Informed by ACT for CLBP
achieved a clinically meaningful reduction of 2 points at related skills could be successfully integrated into usual
3 months and 12 months, where 2 to 3 points are gener- physical therapy with additional staff training and sup-
ally judged to be clinically important.41 PACT participants port. Our 2-day training and on-going monthly supervi-
showed significantly greater improvements in secondary sion seem to have provided suitable support, enabling
measures of pain-related interference at 3 but not physical therapists to deliver PACT with high fidelity. Feed-
12 months. People who received PACT rated their treat- back within training sessions indicated that none of the
ment as having greater credibility. The cost-consequences methods trained deviated substantially from what the
analysis revealed no major cost differences between PACT physical therapists would consider within their scope of
and UC (with PACT training costs excluded, as this type of practice. This is imperative because higher levels of treat-
brief training could be incorporated into continuing pro- ment fidelity are associated with better retention rates
fessional development) and that resource use reduced and treatment outcomes.6 Our nested qualitative study of
over time in both groups, which could reflect improve- PACT physical therapist found treatment was acceptable
ments due to treatment or simply the natural history of and feasible. These findings suggest PACT could success-
the condition. PACT had lower attrition and total treat- fully broaden physical therapists’ scope of practice when
ment was completed within a mean of 2 hours, in contrast treating patients with CLBP. However, our results suggest
to 3 hours for UC. No between-group differences were structural barriers, like the availability of private rooms
observed on the remaining secondary outcomes including and supervision/support, need to be addressed for physical
pain, mood, self-efficacy, or the ACT process variables therapists to incorporate psychological techniques success-
(acceptance and committed action). PACT was designed fully.
to shift the focus from pain to daily functioning and as Both groups in the trial were comparable at baseline
reducing pain was not the primary aim of treatment, it demonstrating randomization worked well and both
was not surprising that there was no difference between received physical therapy, with separate groups of clini-
groups. Depression and anxiety scores in the PACT trial cians delivering the treatments, which limits bias and
were below clinical cut-offs at baseline and were mini- strengthens the validity of results. However, there were
mally targeted as part of treatment, which could explain some sources of bias that need to be considered. Some
the lack of change in these measures. aspects of the trial were more pragmatic than explanatory
Previous research established ACT was effective for (see PRECIS figure in Supplementary Materials) and as a
treating chronic pain,27,46 with small to medium effects result, a higher percentage of UC participants received no
on functioning and disability and suggested combining it treatment. However, the rate of non-attendance was simi-
with physical therapy might make it more acceptable to lar to previous reports from UK physical therapy trials in
patients with CLBP referred to that service.39 Our results this population.12 A weakness of the study was our inabil-
support these findings and are in line with a recent meta- ity to collect accurate information on the number and
analysis of CBT for non-specific low back pain, which type of care providers in the UC arm or comparable data
found the effect of CBT versus other recommended active on the UC sessions patients attended. In addition, as most
treatments ranged from small to moderate; and that participants who were lost to follow-up did not return
most studies maintained a clinically meaningful 30% questionnaires, we were not able to collect the reasons
decrease in the RMDQ over the long term.40 why they were lost to follow-up. The PACT trial restricted
Recruitment and retention data suggest the trial was eligibility to participants speaking good English, referred
well designed and implemented and that PACT was an to physical therapy in public hospitals in London and
acceptable treatment approach for patients with CLBP South East England and excluded those with severe psychi-
referred to physical therapy. Of the 660 patients atric comorbidities common in chronic pain, which some-
informed about the study by their physical therapist, what limits generalizability.
72% (478) were screened and 248 (51.8%) met selection PACT was not designed to turn physical therapists into
criteria and agreed to participate. Retention in the trial ACT practitioners and, as expected, the fidelity assess-
was excellent, with 219 (88.3%) providing follow-up ment showed that few ACT-consistent methods were
data. These findings are supported by our nested quali- delivered. Moreover, the theory-derived measures, used
tative study of 20 PACT participants, which also indi- to assess acceptance and committed action, indicated
cated treatment was acceptable to participants. This is no differences between groups. This means the mecha-
essential as treatment expectancy and credibility have nisms of action behind the improvement in disability
been shown to be associated with better outcome in with PACT are unclear and suggests the treatment needs
physical and CBT treatments of patients with CLBP.43 some redesign and/or the training could be enhanced,
The PACT treatment approach also has potential to to ensure PACT is delivered with higher competency
address physical therapists’ barriers to using psychological and with greater impact on these key processes.10 More
techniques effectively.3 Recent research has highlighted research is required to refine the PACT intervention,
some of the complexity in psychologically informed treat- enhance efficacy, and maintain effects over the longer
ment,30 such as treatment specification, cost, and inter- term. It would be valuable to explore how to optimize
vention fidelity, which we have explored in this trial. In impact on the specific intended processes of change,
addition, systematic reviews have concluded that physical namely acceptance and committed action, as an impor-
therapists’ lack of confidence in successfully delivering psy- tant question persist about how to facilitate this with
chological interventions might be rectified with additional clear implications for wider delivery and training. Fur-
training and support.21,7 We have demonstrated that ACT- thermore, in terms of mechanisms of action, it is
Godfrey et al The Journal of Pain 79
impossible to disentangle the benefits of employing a Ethics Committee Approval
more psychological model of treatment delivery (longer The trial was registered on 22/10/2014 prospectively
sessions, private rooms, and on-going therapist supervi- before enrolling the first patient; registration number:
sion) from the specific content of sessions. In future ISRCTN95392287. This study was approved by the
research, we intended to further investigate training, Research Ethics Committee (REC) approval (National
fidelity, and competency questions. There may also Research Ethics Committee South Central - Berkshire;
have been some contamination in UC, as CBT-based 14/SC/0277) and informed written consent was given by
methods are sometimes employed within routine physi- all patients.
cal therapy.7,8 It is possible that PACT treatment was too
brief, as it involved 1 hour less treatment than UC, and
increasing the dose with additional sessions might help Acknowledgments
maintain benefits. Additional training and support
materials might address these issues and could increase We would like to thank the patient advisors involved
access to PACT. In addition, it might be fruitful to inves- in this trial, especially John Pallet, who was a coappli-
tigate if the advantages of PACT over UC are clearer in cant on the grant, for their dedication, support, and
patients with greater disability or more psychosocial risk input throughout the trial; the physiotherapy, medical
factors at baseline. and health psychology students at the King’s College
The findings from the PACT trial are encouraging and London, who helped collect data, particularly Daniel
have the potential to improve the management of peo- Bourke for help with the cost-consequences analysis
ple with CLBP. The PACT treatment approach can and Shaira Hassan for help with the fidelity study; all
address physical therapists’ barriers to using psychologi- physical therapists at the participating sites for identify-
cal techniques,3 as well as patients’ concerns about find- ing potential participants and delivering the usual phys-
ing a credible treatment offering aspects of care they ical therapy treatment over the 18-month trial period.
value.28 As physical therapy is a common treatment for We would especially like to thank the PACT physical
CLBP and millions of patients are referred to physical therapists: Rosie Cruickshank, Francesca Howes, Chris-
therapists every year, even small additional benefits tina Sothinathan, Laura Oliver, James Womersley, Nicola
could have a considerable impact for patients, health- Kelly, Leon Palmer-Wilson, and Mike Carpenter for their
care providers, and society. Next steps might include enthusiasm for and dedication to the trial and thought-
refining the training and support for physical therapists, ful and consistent delivery of PACT treatment; Dr. Julie
as well as investigating whether to select patients and/ Denning for her help with the fidelity study; and finally
or provide additional treatment sessions. Further all the trial participants for their contribution to improv-
booster sessions should be considered and could be ing knowledge in this area.
delivered remotely, as in other musculoskeletal condi-
tions, to help maintain the effects of treatment over
time. More research is warranted to develop successful Supplementary data
care in the long term and to determine whether PACT is Supplementary data related to this article can be
effective and cost effective in a larger trial. found at https://doi.org/10.1016/j.jpain.2019.05.012.

References 5. Borokovec TD, Nau SD: Credibility of analogue ther-

apy rationales. J Behav Ther Exp Psychiatry 3:257-260,
1972
1. Alexanders J, Anderson A, Henderson S: Musculoskeletal
physiotherapists’ use of psychological interventions: A sys- 6. Borrelli B: The assessment, monitoring, and enhance-
tematic review of therapists’ perceptions and practice. ment of treatment fidelity in public health clinical trials. J
Physiotherapy 101:95-102, 2015. https://doi.org/10.1016/j. Public Health Dent 71(Suppl 1):S52-S63, 2011. https://doi.
physio.2014.03.008 org/10.1111/j.1752-7325.2011.00233.x
2. Artus M, van der Windt DA, Jordan KP, Hay EM: Low 7. Brunner E, Dankaerts W, Meichtry A, O’Sullivan K, Probst
back pain symptoms show a similar pattern of improvement M: Physical therapists’ ability to identify psychological fac-
following a wide range of primary care treatments: A sys- tors and their self-reported competence to manage chronic
tematic review of randomized clinical trials. Rheumatology low back pain. Phys Ther 98:471-479, 2018. https://doi.org/
49:2346-2356, 2010. https://doi.org/10.1093/rheumatology/ 10.1093/ptj/pzy01
keq245
8. Brunner E, De Herdt A, Minguet P, Baldew SS, Probst M:
3. Barker E, McCracken LM: From traditional cognitive- Can cognitive behavioural therapy based strategies be inte-
behavioural therapy to acceptance and commitment ther- grated into physiotherapy for the prevention of chronic
apy for chronic pain: A mixed-methods study of staff expe- low back pain? A systematic review. Disabil Rehabil 35:1-10,
riences of change. Br J Pain 8:98-106, 2014. https://doi.org/ 2013. https://doi.org/10.3109/09638288.2012.683848
10.1177/2049463713498865
9. Buchbinder R, van Tulder M, Oberg B, Costa LM, Woolf
4. Bombardier C: Outcome assessments in the evaluation of A, Schoene M, Croft P: Low back pain: A call for action. Lan-
treatment of spinal disorders: Summary and general recom- cet 391:2384-2388, 2018. https://doi.org/10.1016/S0140-
mendations. Spine 25:3100-3103, 2000 6736(18)30488-4
80 The Journal of Pain RCT of Physical Therapy Informed by ACT for CLBP
10. Carvalho SA, Gillanders D, Palmeira L, Pinto-Gouveia J, 21. Hall A, Richmond H, Copsey B, Hansen Z, Williamson E,
Castilho P: Mindfulness, selfcompassion, and depressive Jones G, Fordham B, Cooper Z, Lamb S: Physiotherapist-
symptoms in chronic pain: The role of pain acceptance. J Clin delivered cognitive-behavioural interventions are effective
Psychol 74:2094-2106, 2018. https://doi.org/10.1002/jclp.22689 for low back pain, but can they be replicated in clinical
practice? A systematic review. Disabil Rehabil 40:1-9, 2018.
11. Chen Y, Campbell P, Strauss VY, Foster NE, Jordan KP, https://doi.org/10.1080/09638288.2016.1236155
Dunn KM: Trajectories and predictors of the long-term
course of low back pain: Cohort study with 5-year follow- 22. Hay EM, Mullis R, Lewis M, Vohora K, Main CJ, Watson
up. Pain 159:252-260, 2018. https://doi.org/10.1097/j. P, Dziedzic KS, Sim J, Minns Lowe C, Croft PR: Comparison
pain.0000000000001097 of physical treatments versus a brief pain-management pro-
gramme for back pain in primary care: A randomised clini-
12. Critchley DJ, Ratcliffe J, Noonan S, Jones RH, Hurley cal trial in physiotherapy practice. Lancet 365:2024-2030,
MV: Effectiveness and cost-effectiveness of three types 2005. https://doi.org/10.1016/S0140-6736(05)66696-2
of physiotherapy used to reduce chronic low back pain dis-
ability: A pragmatic randomized trial with economic evalu- 23. Hayes SC, Strosahl KD, Wilson KG: Acceptance and
ation. Spine 32:1474-1481, 2007. https://doi.org/10.1097/ Commitment Therapy: An Experiential Approach to Behav-
BRS.0b013e318067dc26 ior Change. New York, NY, US, Guilford Press, 1999

13. de Campos TF: Low back pain and sciatica in over 16s: 24. Herdman M, Gudex C, Lloyd A, Janssen M, Kind P, Par-
Assessment and management NICE Guideline [NG59]. J kin D, Bonsel G, Badia X: Development and preliminary
Physiother 63:120, 2017. https://doi.org/10.1016/j.jphys. testing of the new five-level version of EQ-5D (EQ-5D-5L).
2017.02.012de Qual Life Res 20:1727-1736, 2011. https://doi.org/10.1007/
s11136-011-9903-x
14. Du S, Hu L, Dong J, Xu G, Chen X, Jin S, Zhang H, Yin H:
Self-management program for chronic low back pain: A sys- 25. Hoy D, Bain C, Williams G, March L, Brooks P, Blyth F,
tematic review and meta-analysis. Patient Educ Couns Woolf A, Vos T, Buchbinder R: A systematic review of the
100:37-49, 2017. https://doi.org/10.1016/j.pec.2016.07.029 global prevalence of low back pain. Arthritis Rheum
64:2028-2037, 2012. https://doi.org/10.1002/art.34347
15. Dworkin RH, Turk DC, Farrar JT, Haythornthwaite JA,
Jensen MP, Katz NP, Kerns RD, Stucki G, Allen RR, Bellamy 26. Hudak PL, Wright JG: The characteristics of patient sat-
N, Carr DB, Chandler J, Cowan P, Dionne R, Galer BS, Hertz isfaction measures. Spine 25:3167-3177, 2000. https://doi.
S, Jadad AR, Kramer LD, Manning DC, Martin S, McCormick org/10.1097/00007632-200012150-00012
CG, McDermott MP, McGrath P, Quessy S, Rappaport BA,
Robbins W, Robinson JP, Rothman M, Royal MA, Simon L, 27. Hughes LS, Clark J, Colclough JA, Dale E, McMillan
Stauffer JW, Stein W, Tollett J, Wernicke J, Witter J: Core D: Acceptance and commitment therapy (ACT) for
outcome measures for chronic pain clinical trials: IMMPACT chronic pain: A systematic review and meta-analyses.
recommendations. Pain 113:9-19, 2005. https://doi.org/ Clin J Pain 33:552-568, 2017. https://doi.org/10.1097/AJP.
10.1016/j.pain.2004.09.012 0000000000000425

16. Fish RA, McGuire B, Hogan M, Morrison TG, Stewart I: 28. Kamper SJ, Haanstra TM, Simmons K, Kay M, Ingram
Validation of the chronic pain acceptance questionnaire TGJ, Byrne J, Roddick JM, Setliff A, Hall AM: What do
(CPAQ) in an internet sample and development and prelim- patients with chronic spinal pain expect from their physio-
inary validation of the CPAQ-8. Pain 149:435-443, 2010. therapist? Physiother Can 70:36-41, 2018. https://doi.org/
https://doi.org/10.1016/j.pain.2009.12.016 10.3138/ptc.2016-58

17. Foster NE, Anema JR, Cherkin D, Chou R, Cohen SP, Gross 29. Katz JN: Lumbar disc disorders and low-back pain:
DP, Ferreira PH, Fritz JM, Koes BW, Peul W, Turner JA, Maher Socioeconomic factors and consequences. J Bone Joint
CG: Prevention and treatment of low back pain: evidence, Surg 88(Suppl 2):21-24, 2006. https://doi.org/10.2106/JBJS.
challenges, and promising directions. Lancet 391:2368-2383, E.01273
2018. https://doi.org/10.1016/S0140-6736(18)30489-6
30. Keefe FJ, Main CJ, George SZ: Advancing psychologi-
18. GBD 2016 Disease and Injury Incidence and Prevalence cally informed practice for patients with persistent muscu-
Collaborators: Global, regional, and national incidence, loskeletal pain: Promise, pitfalls, and solutions. Phys Ther
prevalence, and years lived with disability for 328 diseases 98(5):398-407, 2018 https://doi.org/10.1093/ptj/pzy024
and injuries for 195 countries, 1990-2016: A systematic anal-
ysis for the Global Burden of Disease Study 2016. Lancet 31. Kroenke K, Spitzer RL, Williams JB: The PHQ-9: validity
390:1211-1259, 2017. https://doi.org/10.1016/S0140-6736 of a brief depression severity measure. J Gen Intern Med
(17)32154-2 16:606-613, 2001. https://doi.org/10.1046/j.1525-1497.2001.
016009606.x
19. Godfrey E, Galea Holmes M, Wileman V, McCracken L,
Norton S, Moss-Morris R, Pallet J, Sanders D, Barcellona M, 32. Main CJ, George SZ: Psychologically informed practice
Critchley D: Physiotherapy informed by acceptance and for management of low back pain: Future directions in
commitment therapy (PACT): Protocol for a randomised practice and research. Phys Ther 91:820-824, 2011. https://
controlled trial of PACT versus usual physiotherapy care for doi.org/10.2522/ptj.20110060
adults with chronic low back pain. BMJ Open 6:e011548,
2016. https://doi.org/10.1136/bmjopen-2016-011548 33. McCracken LM, Chilcot J, Norton S: Further develop-
ment in the assessment of psychological flexibility: A
20. Gore M, Sadosky A, Stacey BR, Tai KS, Leslie D: The shortened committed action questionnaire (CAQ-8). Eur
burden of chronic low back pain: clinical comorbidities, J Pain 19:677-685, 2015. https://doi.org/10.1002/ejp.589
treatment patterns, and health care costs in usual care
settings. Spine 37:E668-E677, 2012. https://doi.org/10.1097/ 34. McCracken LM, Morley S: The psychological flexibility
BRS.0b013e318241e5de model: A basis for integration and progress in psychological
Godfrey et al The Journal of Pain 81
approaches to chronic pain management. J Pain 15:221- 42. Roland M, Morris R: A study of the natural history of
234, 2014. https://doi.org/10.1016/j.jpain.2013.10.014 back pain. Part I: Development of a reliable and sensitive
measure of disability in low-back pain. Spine 8:141-144,
35. Mundt JC, Marks IM, Shear MK, Greist JH: The work and 1983
social adjustment scale: A simple measure of impairment in
functioning. Br J Psychiatry 180:461-464, 2002. https://doi. 43. Smeets RJ, Beelen S, Goossens ME, Schouten EG,
org/10.1176/appi.ajp.163.1.28 Knottnerus JA, Vlaeyen JW: Treatment expectancy and
credibility are associated with the outcome of both physical
36. National Low Back and Radicular Pain Pathway 3.0 3rd and cognitive-behavioral treatment in chronic low back
ed, 2017 pain. Clin J Pain 24:305-315, 2008. https://doi.org/10.
1097/AJP.0b013e318164aa75
37. Nicholas MK: The pain self-efficacy questionnaire: Tak-
ing pain into account. Eur J Pain 11:153-163, 2007. https:// 44. Spitzer RL, Kroenke K, Williams JB, Lowe B: A brief mea-
doi.org/10.1016/j.ejpain.2005.12.008 sure for assessing generalized anxiety disorder: The GAD-7.
Arch Intern Med 166:1092-1097, 2006. https://doi.org/
38. Nicholas MK, George SZ: Psychologically informed inter- 10.1001/archinte.166.10.1092
ventions for low back pain: An update for physical therapists.
Phys Ther 91:765-776, 2011. https://doi.org/10.2522/ptj.20100278 45. Stratford P, Gill C, Westaway M, Binkley J: Assessing dis-
ability and change on individual patients: A report of a
39. Pincus T, Anwar S, McCracken LM, McGregor A, Graham
patient specific measure. Physiother Can 47:258-263, 1995.
L, Collinson M, McBeth J, Watson P, Morley S, Henderson J,
https://doi.org/10.3138/ptc.47.4.258
Farrin AJ: Delivering an optimised behavioural intervention
(OBI) to people with low back pain with high psychological
46. Vowles KE, McCracken LM, O’Brien JZ: Acceptance and
risk; results and lessons learnt from a feasibility randomised
values-based action in chronic pain: A three-year follow-up
controlled trial of contextual cognitive behavioural therapy
analysis of treatment effectiveness and process. Behav Res
(CCBT) vs. physiotherapy. BMC Musculoskelet Disord
Ther 49:748-755, 2011. https://doi.org/10.1037/0022-006X.
16:147, 2015. https://doi.org/10.1186/s12891-015-0594-2
76.3.397
40. Richmond H, Hall AM, Copsey B, Hansen Z, Williamson E,
Hoxey-Thomas N, Cooper Z, Lamb SE: The effectiveness of 47. Ware J Jr., Kosinski M, Keller SD: A 12-item short-form
cognitive behavioural treatment for non-specific low back health survey: Construction of scales and preliminary tests
pain: A systematic review and meta-analysis. PLoS One 10: of reliability and validity. Med Care 34:220-233, 1996.
e0134192, 2015. https://doi.org/10.1371/journal.pone.0134192 https://doi.org/10.2307/3766749

41. Roland M, Fairbank J: The Roland−Morris disability 48. White IR, Horton NJ, Carpenter J, Pocock SJ: Strategy
questionnaire and the Oswestry disability questionnaire. for intention to treat analysis in randomised trials with
Spine 25:3115-3124, 2000. https://doi.org/10.1097/ missing outcome data. BMJ 342:d40, 2011. https://doi.org/
00007632-200012150-00006 10.1136/bmj.d40