Pain self‐management intervention supports successful attainment of self‐selected rehabilitation goals—secondary analysis of a randomized controlled trial

Abstract Objectives (i) Describe patients' self‐selected activity‐related rehabilitation goals, and (ii) compare attainment of these rehabilitation goals among people with persistent tension‐type neck pain receiving a group‐based pain and stress self‐management intervention (PASS) or individual physiotherapy (IPT). Methods Before intervention and random allocation to PASS or IPT, 156 people (PASS n = 77, IPT n = 79), listed three self‐selected activity‐related rehabilitation goals by use of the Patient Goal Priority Questionnaire (PGPQ). For each activity goal, participants rated limitations in activity performance, self‐efficacy and fear of activity performance, readiness to change to improve performance, and expectations of future activity performance. At follow‐ups (10 weeks, 20 weeks, 1 year and 2 years after inclusion), participants also responded to a question on changes made to improve activity performance. Mann–Whitney U test was used to evaluate between‐group differences. Results There were between‐group differences in favour of PASS in the attainment of self‐selected rehabilitation goals with regard to activity limitations and satisfaction with activity performance at all follow‐ups. Conclusions PASS was more successful than IPT for the attainment of self‐selected rehabilitation goals, improvements in activity limitations and satisfaction with activity performance as measured by PGPQ. The PASS programme emphasized the importance of applying active pain‐ and stress‐coping techniques in personal ‘risk situations’ for pain flare‐ups, which appear to support people with persistent tension‐type neck pain to make changes in their lives to improve activity performance. Patient or Public Contribution Patient engagement in rehabilitation by self‐selected goals was investigated, but patients were not involved in the design or conduct of the study.


| INTRODUCTION
The biopsychosocial perspective on pain highlights the significance of applying cognitive-behavioural principles in pain rehabilitation. [1][2][3] Rehabilitation strategies that address cognitive and behavioural factors involved in the maintenance of musculoskeletal pain by promoting active pain-coping skills and self-management have shown beneficial effects on pain-related disability. [4][5][6] The rationale for selfmanagement interventions is to induce health behaviour assumed to yield sustainable long-term favourable effects on disability. 7,8 Person-centred care and the patient's active involvement in rehabilitation have gained increased attention. [9][10][11][12] An important part of a person-centred approach is the identification and evaluation of patient-specific rehabilitation goals. 13,14 Clinical tools can support the goal-setting process: assist in drawing out and setting priority to the patient's valued treatment goals, and monitor the treatment process and goal achievements. 15 Patient-specific instruments are useful for bringing forward and acknowledging the individual's own valued activities, such as activities that are perceived important for the person's participation in everyday life and for actively involving the patient in the rehabilitation process. [16][17][18] The patient's active involvement in rehabilitation goal-setting has been suggested to increase motivation, active engagement and satisfaction with rehabilitation interventions. 19 Studies have found a positive association between patient involvement in goal-setting and rehabilitation outcome. 20,21 Despite support for the use of patient-specific instruments for goal-setting, they are used to a very small extent in clinical rehabilitation. 18,22,23 The Patient Goal Priority Questionnaire (PGPQ) has been described as a clinical tool to assist the patient in identifying and prioritizing activity-related goals in pain rehabilitation, likewise to assist shared decision-making and patient engagement in the evaluation of rehabilitation progress and outcome. 24 The activity goals rated by PGPQ are self-selected by the patient and thus person-specific as opposed to generic. Each prioritized activity goal is followed by a rating of generic aspects relating to activity performance; that is, limitations in activity performance, as well as satisfaction, self-efficacy and fear of activity performance, readiness to change to improve performance and expectations of future activity performance. One study has examined test-retest reliability of the self-report scales in PGPQ in a sample of people having persistent pain by calculating intra-class correlation coefficients (ICCs) and the reported ICCs ranged from .35 to .81. 25 The PGPQ has been used in pain rehabilitation research for the identification of patients' treatment goals 26 and for the assessment of clinically important changes related to activity performance and goal achievements. 27,28 One study examined the concurrent validity of the PGPQ to a generic measure of disability, the Pain Disability Index (PDI), in a sample of people having persistent pain. The PGPQ was negatively and moderately correlated with the PDI, indicating that the patients' perceptions of behavioural performance in prioritized activity goals as measured by PGPQ were moderately correlated to a generic measure of disability. 24 In a randomized controlled trial (RCT) of a multicomponent pain and stress self-management group intervention (PASS), the PGPQ was included in a comprehensive self-assessment questionnaire used for evaluation of intervention effects. Previous publications have reported results on posttreatment effects in the primary outcome: pain control, self-efficacy for activities of daily living (ADL) and disability 29 and on long-term effects at 2 30,31 and 9 years. 32 The results on primary outcomes have shown that the PASS had a better effect on pain control, pain-related self-efficacy for ADL, disability and pain catastrophizing than a control treatment: individual physiotherapy (IPT), for patients with persistent tension-type neck pain, in both short-term and long-term. 29,30 Also, treatment gains in self-efficacy for activity performance was an important predictor for favourable long-term outcomes on pain-related disability. 31 However, the outcome on the participants' attainment of self-selected rehabilitation goals collected by PGPQ before the intervention and at follow-ups (at 10 weeks, 20 weeks, 1 year and 2 years after inclusion) has not yet been reported.
Hence, the objective of this study was to (i) describe patients' self-selected activity-related rehabilitation goals, and (ii) compare attainment of these rehabilitation goals as measured by PGPQ among people with persistent tension-type neck pain participating in a twoarmed RCT receiving either PASS intervention or IPT.

| Study design
This study is a secondary analysis of data from a two-armed pragmatic 33 RCT 34 evaluating between-group differences and within-group changes over time with five time-points of data collection, on activity performance in self-selected rehabilitation goals.

| Participants and procedures
Detailed descriptions of participants and procedures have been reported in previous publications. [29][30][31] In brief, people with neck pain seeking physiotherapy treatment at nine primary healthcare (PHC) centres in eight towns in Sweden were consecutively recruited from September 2004 to April 2006. They were considered eligible if they were aged 18-65 years and had tension-type neck pain of persistent duration; that is, more than 3 months. Reasons for exclusion were: insufficient fluency in Swedish, medical history of psychotic disorder, pregnancy, ongoing treatment for neck pain or possible depression indicated by a score of ≥11 points on the depression subscale of the Hospital Anxiety and Depression Scale (HADS-D). 35 An a-priori power calculation based on data from a pilot study, 36 estimated that a sample size of 55 per group would be sufficient to detect a 10% between-group difference on the primary outcomes variables: 'pain control' and 'self-efficacy for performing activities in spite of pain'. No power calculation was undertaken regarding outcome on the PGPQ. The number of participants available for analyses of PGPQ was considered acceptable for analyses but the capacity to ensure power to detect important between-group differences was not calculated a priori. Before enrolment to the study, the participants provided informed consent to participate.
The study was approved by the Ethical Review Board of Uppsala University (Ups02-088). After completing the baseline selfassessment questionnaire, the participants were randomly allocated to receive either the intervention PASS or the control condition IPT. Allocation was stratified by the PHC centre. PASS and IPT were delivered at the PHC centres by experienced physiotherapists. The physiotherapists delivering PASS and IPT did not have access to the content of the questionnaire and were unaware of participants' responses to the PGPQ. Follow-ups were conducted 10 weeks, 20 weeks, 1 year and 2 years after inclusion by use of self-assessment questionnaires, which were mailed to the participants. In the present study, responses to the PGPQ at baseline and follow-ups were analysed.

| PASS and IPT interventions
PASS was a multicomponent PASS consisting of seven weekly group sessions of 1.5 h each, and an additional booster session at 20 weeks after the initial session. Each session consisted of applied relaxation, 37 body-awareness exercises 38 and interactive lectures, emphasizing two-way communication and group discussions concerning issues related to pain self-management. 29 The rationale was to teach the patient active pain-and stress-coping skills by identifying personal 'risk situations' in everyday life (i.e., activities, movements or thoughts believed to cause pain or stress) and applying techniques in these situations to manage physical arousal, and thereby prevent the pain from starting or increasing. The PASS participants attended an average of seven (range: 4-8) group sessions. 29 IPT was individually administered physiotherapy sessions in accordance with current practice at the PHC centres and was not standardized; that is, type of treatment, frequency of visits and duration of contact were left to the judgement and agreement between the physiotherapist and the patient. The sessions involved several treatment modalities: spinal mobilization techniques and massage, acupuncture, transcutaneous electric nerve stimulation and individually tailored exercise programmes. The IPT participants received an average of eleven (range: 1-52) individual sessions. 29

| Data collection
The PGPQ was used to collect data concerning patients' priorities of rehabilitation goals. 24 The participants selected three activity-related rehabilitation goals; that is, activities they were unable or had difficulties performing due to pain, and they wanted to improve through rehabilitation. They ranked the relative importance of the activities from 1 to 3, with 1 representing the most important activity (referred to as Activity 1, 2 and 3 in this paper). For each activity, the participants rated the following aspects: 8. At follow-ups, aspects 6 and 7 were omitted, and an additional rating was included concerning the amount of changes made to improve activity performance, scored on a 4-point ordinal scale (0 = none, 1 = a few changes, 2 = some changes, 3 = many changes).
At the follow-ups, each individual's original self-selected rehabilitation goals as stated in the baseline questionnaire were filled in before the questionnaire was sent to the participant.

| Data management and analyses
The PDI 39,40 was used to categorize the patients' self-selected rehabilitation goals listed in the PGPQ according to activity domains. The PDI has been used in other studies to categorize the patient-specific activities in the PGPQ into generic activity domains. 24,25 The PDI is a generic instrument designed to assess (5) sexual behaviour; (6) self-care and (7) life-support activities.
The two authors performed the categorization of activities to PDI domains using qualitative analysis. Both authors independently examined the data set and categorized each activity goal into one of the seven PDI domains. Then, the authors compared the categorizations. There was almost a complete inter-rater agreement. The few disagreements were discussed until consensus was reached and the categorization was revised. Activity goals that were interpreted as covering more than one PDI domain were allotted to a separate category. An additional category was included and labelled 'general functional ability', comprising general activities not tied to any specific task or situation, for example, sitting, standing, walking and handling objects. In addition, priorities that were not considered activity-related rehabilitation goals, such as 'get well', 'get less pain' and 'feel less tired', were categorized as 'Not activity-related goals'.
Mann-Whitney U test 41 was used to evaluate differences between groups at baseline and follow-ups (10 weeks, 20 weeks, 1 year and 2 years), in measures of aspects relating to activity performance as rated by PGPQ. Data were analysed per protocol; that is, based on participants with available data at each follow-up and according to group allocation at baseline. 42 To adjust for multiple testing; that is, five time points of follow-ups and for all three goal activities, a p-value less than .01 was accepted as statistically significant. 41 Analyses were conducted using IBM SPSS Statistics 21 for Windows. 43

| RESULTS
Baseline characteristics of the 156 participants who were originally included in the RCT, PASS n = 77; IPT n = 79, are displayed in Table 1.
At follow-ups, 20%-37% of participants failed to respond to the selfassessment questionnaires. Figure 1 provides a flowchart illustrating participation in the study over the follow-ups. Activities related to 'Recreation and Hobbies' were the most frequent rehabilitation goals in Activity 2 (46%) and Activity 3 (41.5%). Among the activity-related goals deemed covering more than one activity domain, the most frequent combination was 'Occupation and Education' and 'Recreation and Hobbies'. An example of such a combined goal is 'Sitting upright in front of a computer at work or a TV screen to watch a TV programme for more than 15 min'. Very few the participants' rehabilitation goals were categorized as not activity-related (4.4%).    In previous publications from this RCT, we have reported favourable effects of PASS on pain control, pain-related self-efficacy for ADL and disability as measured by generic instruments. [29][30][31] In this study, we aimed to explore effects on the self-selected person-specific activity-related outcomes. The importance of the self-selected personspecific goal-setting within physiotherapy has been raised as a means to identify goals that are perceived as meaningful and valued by the participant and worth investing time and effort to achieve. 19 Indeed, it is plausible that the self-selected activity goals in PGPQ enhanced T A B L E 3 Responses to PGPQ items for PASS and IPT at baseline and follow-ups and comparison of differences between groups based on participants with available data at follow-ups Item 8 follow-up: Amount of changes made by the patient to improve activity performance (0-3, higher score more changes made) .023 .002 .003 .007

PASS
.014 .303 .003 .006 .416 .003 .013 .034 PASS-group 7 (4-9)  The present study is a secondary analysis of an RCT conducted and completed more than 10 years ago, which may be considered a limitation. The main publication, 29 outlining results on posttreatment effects in primary outcome variables, was published in 2010. The IPT provided in the study was equivalent to a regular practice of physiotherapy in PHC at the time of data collection. Since contemporary IPT for people with persistent tension-type neck pain is delivered in a similar way today as when the RCT was undertaken, the results are still valid for current clinical practice. Moreover, as the aim of the present study was to compare goal attainment in goals set by the patients themselves, we believe that the research question addressed is of great interest and importance for the ongoing knowledge development of person-centred rehabilitation. Selfselected rehabilitation goals were to a larger extent attained and maintained in the group receiving the self-management programme (PASS) than those receiving IPT. IPT may have included instructions for self-management yet it did not lead to similar goal attainment of the patients' self-selected goals.
It is a limitation that 20%-37% of participants failed to respond to follow-up questionnaires despite reminders. In addition, the power calculation was based on the primary outcome variables in the RCT, not on the outcome of PGPQ. Still, the number of participants available for analyses of PGPQ was considered acceptable for analyses but limited the capacity to ensure power to detect important between-group differences. Hence it is possible that we failed to detect differences in intervention effects that were present. The vast majority of participants were women, thus the results should be generalized to men with caution.

| CONCLUSION
This study showed that the PASS intervention was more successful than IPT for the attainment of self-selected activity-related rehabilitation goals with regard to improvements in activity limitations and satisfaction with activity performance as measured by PGPQ. The differences were sustained over time at follow-ups up to 2 years postintervention. Immediately after completing the intervention, the PASS group reported having done more changes in their life to improve activity performance than the IPT group. The PASS programme emphasized the importance of applying active pain-and stress-coping techniques in personal 'risk situations' to manage pain flare-ups, which appear to support people with persistent tension-type neck pain to make changes that improve activity performance.