Implementing a behavioral medicine approach in physiotherapy for patients with musculoskeletal pain: a scoping review

Supplemental Digital Content is Available in the Text. The interventions in the scoping review were in agreement with the definition of behavioral medicine in physiotherapy, but reported behavior change techniques were few.


Introduction
Patients seeking for musculoskeletal pain care are common in physiotherapists' clinical practice. Modern pain management should be guided by a biopsychosocial theoretical approach, 14 which is complex. In intervention research on musculoskeletal pain, physiotherapists often study variations of what are called "behavioral and cognitive components," and positive sustainable evidence of applying these components has greatly increased during the past decade. 7,26,32,39,[59][60][61] However, how to effectively integrate the "behavioral and cognitive components" in the management of musculoskeletal pain from a biopsychosocial theoretical approach is challenging.
Behavioral medicine in physiotherapy, informed by the International Society for Behavioral Medicine's definition of behavioral medicine, 30 ie, the "integration of psychosocial, behavioral and biomedical knowledge in analyses of patients'/clients' behaviors in activities of importance for participation and in choosing and applying treatment and behavior change methods and evaluating outcomes," 47 can guide us to a better integration of important behavioral, psychosocial, and physical components in all phases of patient encounters. According to the behavioral medicine approach, patients with musculoskeletal pain should be coached to self-manage, to change behavior when needed, and to reduce dependency on health care. 15,16,46,50 A uniform description of how to integrate behavioral, psychosocial, and physical/biomedical knowledge into interventions in musculoskeletal research could better guide practitioners and researchers.
Moving from randomized controlled trials to implementing behavioral medicine in physiotherapy practice is a challenging task. A systematic review of physiotherapists' usage of behavior change techniques in promoting physical activity showed that only a small number of techniques were identified as being used in clinical practice. 32 Furthermore, Fritz et al. 24 studied the effects of multifaceted implementation methods on changing physiotherapists' clinical behavior when treating patients with musculoskeletal pain. They concluded that the methods could support the change in the short term, but that sustaining of the change needed different strategies and/or doses than those used in the study. Thus, although knowledge of the evidence-based behavior change techniques exists, the techniques are not well implemented, and psychosocial, behavioral, and biomedical knowledge is not optimally integrated in physiotherapy 12,13 or specifically in management of musculoskeletal pain. 2,18 A global common description for these interventions in musculoskeletal research could help practitioners and researchers to better communicate the treatments to patients, health care, and policymakers and thus set the stage for more effective implementation.
There are few studies that explicitly describe the integrative concept of behavior medicine in physiotherapy, 25,29,48 and even fewer that describe its implementation. 18,24 A scoping review would highlight the key components, identify the limitations of the current interventions using a behavioral medicine approach in physiotherapy, and provide direction for further research in the field.
The aim of the present scoping review was to study the intervention components and patient outcomes of studies integrating "behavioral and cognitive components" in physiotherapy, to match the interventions with a definition of behavioral medicine in physiotherapy, and to categorize the behavior change techniques targeted at patients with musculoskeletal pain in (1) randomized controlled effect trials or (2) implementation in clinical practice trials.

Methods
The systematic recommendations for a scoping review were used to conduct this study. 11,33 In addition, the PRISMA-ScR checklist was used when reporting the results of this scoping review. 54,55 This scoping review could not be registered in the PROSPERO registry due to current regulations of the registry. No formal protocol was written before conducting this review. However, an outline of the study was written and discussed between the authors before starting the study.
The methods are presented separately, when relevant, for the 2-fold aim of this study, ie, for the (1) randomized controlled effect trials and (2) implementation in clinical practice trials.

Eligibility criteria
To be included in this scoping review, the articles needed to meet the following inclusion criteria, separately for the (1) randomized controlled effect trials and (2) implementation in clinical practice trials.
Randomized controlled effect trials: peer-reviewed journal articles; patients with musculoskeletal pain; studies integrating physical, behavioral, and cognitive components in physiotherapy; and English language. Implementation in clinical practice trials: quasiexperimental or experimental studies; patients with musculoskeletal pain; implementation of physical, behavioral, and cognitive components in physiotherapy in clinical practice; description of the implementation intervention on physiotherapists was provided (ie, what was the implementation intervention supporting the uptake of physiotherapists' new working approach); patient outcomes reported; English language, and peer-reviewed journal articles or unpublished manuscripts (through contact with identified authors).

Information sources and search strategy
To identify relevant studies, the combination of PubMed, MED-LINE, PsycINFO, CINAHL Plus, and Web of Science Core databases was searched on several occasions with the final search being performed on the 19th of October 2019 separately for (1) randomized controlled effect trials and (2) implementation in clinical practice trials. The searches were conducted with topicrelevant MeSH search terms. The full electronic search strategy, which was the same for all databases, is shown in Tables 1 and 2.

Selection of sources of evidence
The first author conducted the database searches. The other authors contributed by identifying studies through other sources, ie, through personal contacts. All eligible studies' titles and abstracts were screened by the first author, who also decided which studies were included in the next step. If the decision for inclusion was perceived as uncertain, the study was included in the next step. In the next step, the full-text articles were downloaded to be assessed by all authors, and the decision for final inclusion was made in agreement.

Data charting process and parameters
Data were tabulated after jointly developing headings for the tables according to the aim of this scoping review. The data charting was divided between the authors, and finally, the correctness of the data in the finished tables was checked by all authors. Table 3 for the randomized controlled effect trials includes the following: reference, country, aim, sample, experimental intervention, control intervention, and patient outcomes. Table 4 for the implementation in clinical practice trials includes the following: reference, country, aim, target group for the implementation and context, patient sample for the intervention, intervention implemented on the patients, control intervention implemented on the patients, and the patient outcomes. Table 5 shows that (1) randomized controlled effect trials and (2) implementation in clinical practice trials both contributed to the content and included the following variables: integrated psychosocial aspects, integrated behavioral aspects, integrated biomedical/physical aspects, behavior change techniques (explicitly reported in the study or interpreted by the authors of the present scoping review), and cluster categorizing of behavior change techniques.

Synthesis of results
A critical quality appraisal was not conducted. The studies were grouped according to the 2-fold aim of this study: (1) randomized controlled effect trials and (2) implementation in clinical practice trials. All included study characteristics, listed in detail above in data charting process and parameters section, are summarized in Tables 3 and 4. The matching of the study interventions with the definition of behavior medicine in physiotherapy 47 was performed from the integration of psychosocial, behavioral, and biomedical knowledge perspective for (1) randomized controlled effect trials and (2) implementation in clinical practice trials separately and is presented in Table 5. This synthesis required an analytical interpretation of the intervention contents to determine whether the content fit with the psychosocial, behavioral, and biomedical knowledge integration. In the studies clearly reported or by the current scoping review authors interpreted behavior change techniques were categorized according to the taxonomy by Michie et al. 38 and are presented in Table 5.

Results
The results are presented both separately and together for the (1) randomized controlled effect trials and (2) implementation in clinical practice trials.

Selection of sources of evidence
Five databases were searched; see Table 1 (randomized controlled effect trials) and Table 2 (implementation in clinical practice trials) for the search strategy, the terms, and number of eligible studies from each database. Figure 1 presents the PRISMA chart for the study selection, number of studies in each phase of the selection process, and major reasons for study exclusion; see also appendix for excluded full-text articles (available online as supplemental digital content at http://links. lww.com/PR9/A75). In total, 18 studies, 15 randomized controlled effect trials and 3 trials of implementation in clinical practice, were included in this scoping review.

Characteristics and results of individual sources of evidence
Studies that contributed to the results were from Australia, 51 Ireland, 40 Norway, 9,10,56,57 Sweden, 4,6,8,19,23,28,34,35,42,43,45,49 the United Kingdom, 27,58 and the United States. 3,44 Target groups for the patient interventions varied. One study was targeted at adolescents, 28 3 were targeted at older persons, [8][9][10]44 and the rest of the studies were targeted at people of working age. 3,4,6,19,23,27,34,35,40,42,43,45,49,51,[56][57][58] The patient outcomes showed mostly significant differences in favor of the experimental intervention in the randomized controlled effect trials. The results for the short-term effects (3month follow-up) were mixed, but the long-term effects were better than those of the control condition. Six studies had 12 months of follow-up, 6,27,35,40,51,56 one had 24 months, 43 one had 36 months, 57 and one had 10 years. 19 The study characteristics, experimental and control interventions, and patient outcomes of the included randomized controlled effect trials regarding investigations of a behavioral medicine approach in physiotherapy for patients with musculoskeletal pain are presented in Table 3.
The 3 implementation in clinical practice trials showed no significant differences in short-term follow-up. One of these studies 23 reported a 12-month follow-up showing a difference in the percentage of patients on sick leave in favor of the experimental group. The study characteristics, implemented interventions, and patient outcomes of the included studies that implemented a behavioral medicine approach in a physiotherapists' clinical practice among patients with musculoskeletal pain are presented in Table 4.

Synthesis of the results for the matching of the interventions with the definition of behavioral medicine in physiotherapy
Both the randomized controlled effect trials and the implementation in clinical practice trials mostly integrated psychosocial, behavioral, and biomedical/physical aspects in the experimental patients' intervention condition, except in one study. The study by Ludvigsson et al. 35 did not clearly report the integration of behavioral aspects in the experimental intervention except that pacing as a behavior change technique was included, implying that patients were learning to alternate between activities and rest. The majority of behavior change techniques reported in randomized controlled effect trials were categorized 38 as "information of natural consequences," "feedback and monitoring," "goals and planning," and "shaping knowledge," and the majority of techniques reported in the implementation in clinical practice trials were categorized 38 as "information of natural consequences," "feedback and monitoring," and "goals and planning." All included randomized controlled effect trials and the implementation in clinical practice trials matched the definition of behavioral medicine in physiotherapy.
The results for matching the experimental interventions of the studies with the definition of behavior medicine in physiotherapy, ie, the "integration of psychosocial, behavioral and biomedical knowledge in analyses of patients'/clients' behaviors in activities of importance for participation and in choosing and applying treatment and behavior change methods and evaluating outcomes" 47 are presented in Table 5. Table 5 also includes studies reporting behavior change techniques and the categorization of the techniques according to Michie et al. 38 4. Discussion

Summary of evidence
Synthesis of the results for the matching of the patient interventions with an existing definition of behavioral medicine in physiotherapy for the randomized controlled effect and the implementation in clinical practice trials showed that the Table 1 Search strategy for behavioral medicine approach in physiotherapy regarding its effects on patient outcomes studied in randomized controlled trials for patients with musculoskeletal pain.
Databases (search date October 19, 2019) PubMed, MEDLINE, PsycINFO, CINAHL Plus, Web of Science Core Search terms ("Behavioural medicine"[All fields] OR "behavioral medicine"[MeSH terms] OR ("behavioral"[All fields] AND "medicine"[All fields]) OR "behavioral medicine"[All fields]) AND ("physical therapy modalities"[MeSH terms] OR ("physical"[All fields] AND "therapy"[All fields] AND "modalities"[All fields]) OR "physical therapy modalities"[All fields] OR "physiotherapy"[All fields]) AND ("pain"[MeSH terms] OR "pain"[All fields]) AND ("random allocation"[MeSH terms] OR ("random"[All fields] AND "allocation"[All fields]) OR "random allocation"[All fields] OR "randomized" interventions mostly integrated psychosocial, behavioral, and biomedical/physical aspects, and thus, showed conformity with the existing definition of behavioral medicine in physiotherapy. 47 The reported behavior change techniques in all trials were few and commonly occurred in the categories 38 ""information of natural consequences," "feedback and monitoring," and "goals and planning." The patient outcomes in the randomized controlled effect trials for the long-term follow-ups showed mostly positive effects in comparison to the control condition. The implementation in clinical practice trials reported no differences in the short term.
The matching of intervention components with the definition of behavioral medicine in physiotherapy was somewhat difficult due to overlap between the psychosocial and behavioral knowledge Table 2 Search strategy for the impact on outcomes for patients with musculoskeletal pain of an implementation of behavioral medicine approach in physiotherapists' clinical practice.  Table 3 Characteristics and patient outcomes of the included randomized controlled effect trials regarding investigations of a behavioral medicine approach in physiotherapy for patients with musculoskeletal pain.

Reference, Country
Aim Sample Experimental intervention

Control intervention Results of patient outcomes
Archer et al., 3 USA Study the effect of a cognitive-behavioral-based physical therapy program (CBPT) compared to an education program (EP).
Patients 6 wk after lumbar laminectomy, .21 y, n 5 86 Standard care including advice about lifting and driving restrictions. CBPT program (in-person session and over the telephone) aimed to decrease fear of movement and increase self-efficacy, including behavioral selfmanagement, problem solving, cognitive restructuring, and relaxation. Treatment manual was given.
Standard care including advice about lifting and driving restrictions. EP included sessions of benefits of physiotherapy, biomechanics after surgery, daily exercise, promoting healing, stress reduction, sleep, energy, communication with health care, and preventing injury.
For patients who had CBPT, disability and pain intensity decreased and physical function and general health increased significantly more compared to those with EP at 3 mo follow-up.
Bring et al., 6 Sweden Study the effect of an individually tailored behavioral medicine approach in physiotherapy delivered through internet compared to same intervention delivered faceto-face or a control group having self-care instructions.
Patients with acute whiplash associated disorders, aged 18-65 y, n 5 55 Individually tailored behavioral medicine intervention, based on functional behavioral analysis and everyday activity goals specifying physical, cognitive, and behavioral skills relevant for goal attainment. Enhancement of selfmanagement skills and level of functioning, strategies for maintenance, and relapse prevention. Seven treatment modules were included.
Written self-care instructions about physical symptoms, relaxation, neck and shoulder range of motion exercises and daily walks.
Significant differences (favoring the individually tailored behavioral medicine intervention groups) between the groups over time (up to 12-mo followup) in disability, selfefficacy in activities, catastrophizing, and fear of movement, but not in pain intensity.

Cederbom et al., 8 Sweden
Study the effect of an individually tailored behavioral medicine approach in physiotherapy compared to one-time delivered advice on physical activity.
Older women with chronic musculoskeletal pain, aged . 65 y, n 5 23 Behavioral medicine intervention integrated with physiotherapy. Individual functional behavior analysis of physical, psychological, social, and physical environmental factors affecting ability in specific everyday activities. Advice on physical activity and its benefits, goal setting, selfmonitoring, feedback, problem-solving strategies, strategies for maintenance, and relapse prevention.
Standard care including one-time only advice on physical activity and its benefits.
No significant differences between groups in pain intensity, disability, or morale were found at any of the follow-ups (postintervention, 3 mo after intervention).
Cederbom et al., 9,10 Norway Study the effects of an individually tailored behavioral medicine approach in physiotherapy compared to standard care.
Older persons, .75 y, with musculoskeletal pain, n 5 105 Functional behavioral analyses of the physical, psychological, social, and environmental factors related to the goal behaviors and treatment goals. Improve physical, behavioral, cognitive, or social skills, improve selfefficacy, decrease fear of falling and fear of movement in the goal behavior, generalize the skills to other behaviors, strategies to maintain new behavior, and problemsolving strategies Advice on and increase of physical activity and its benefits, functional exercises, and selfmonitoring of physical activity.
Standard care including one-time only advice on physical activity and its benefits.
There were differences in pain-related disability, pain severity, health-related quality of life, management of everyday activities, and self-efficacy in goal behaviors favoring the individually tailored behavioral medicine approach in physiotherapy intervention group. The effect on pain severity was maintained at 3-mo follow-up.
(continued on next page) 5 (2020) e844 www.painreportsonline.com No significant between group over time (6-mo follow-up) difference was shown in disability. A significant interaction effect was shown for the EQ-5D index in favor of the personcentered physiotherapy rehabilitation program based on a cognitive-behavioral approach.
(continued on next page) Educational selfmanagement guide, exercises according to guidelines for acute whiplash-associated disorders, return to normal activities, manual therapy was allowed.
Stress inoculation training with exercise decreased disability significantly more than exercise only at 6-wk, and 6-and 12-mo followups.
(continued on next page) 5 (2020) e844 www.painreportsonline.com or components/aspects to be integrated in the assessment, treatment, and evaluation of patients with musculoskeletal pain. Because a behavior can be defined as movement or activities or as the cognitive, emotional, or physiological response of an individual, 52 it is easy to see how behavioral aspects can spill over to the psychosocial and perhaps physical aspects of the definition of behavioral medicine and can thus be difficult to categorize. The definition of behavioral medicine in physiotherapy actually demands the integration of the behavioral, psychosocial, and physical aspects during both analysis and treatment, which was difficult to identify in the included studies. Frequently, the integration was described thoroughly for the analysis but not clearly for the treatment, ie, how the physiotherapists took into account the results of the analysis in the treatment. For example, the identification of fear of movement was mentioned but how to manage this fear in the treatment was not. 9,10,34 Similar results were shown in a recent systematic review in which it was not possible to identify how the cognitive-behavioral components used in physiotherapy were actually operationalized. 26 The reported intervention components in this study varied quite a lot. Frequently reported components ( Table 5) were improve selfefficacy and reduce fear and catastrophizing, generally discuss pain beliefs, increase activities and pain self-management strategies, improve stress management, rehearsal of behaviors No between group over time differences were found in any of the outcomes at 2-mo follow-up.  The studies' experimental interventions' matching with the definition of behavior medicine in physiotherapy: "Integration of psychosocial, behavioral, and biomedical knowledge in analyses of patients'/clients' behaviors in activities of importance for participation and in choosing and applying treatment and behavior change methods and evaluating outcomes" 47  Information of awareness of thoughts and beliefs in behavior.
Alternate between activities and rest (pacing).
Neck-specific exercises, information of neck functioning, postural control, isometric and other progressive neck-specific exercises, home exercise, instructions to continue exercises, and breathing exercises.
Awareness of thoughts and beliefs in behavior, activitybased goal setting, pacing, reinforcement of pain management education, and strategies for relapse prevention.  Functional behavior analysis to identify cognitive skills necessary for goal achievement, and cognitive basic skill acquisition.
Behavior goal identification, self-monitoring of behavior in activities, functional behavior analysis to identify behavioral skills necessary for goal achievement, apply the skills in complex behaviors, ie, cognitive and motor behaviors, and problem-solving strategies, and skill generalization to daily activities Functional behavior analysis to identify the physical skills necessary for goal achievement, and physical and basic skill acquisition. in daily activities, facilitate exercise, and improve function, but several other components had also been used. Interventions and their components in the included studies have certainly been thoughtfully developed. However, it might be possible to further refine the interventions by, eg, using the behavior change wheel system and its underpinning framework for systematically designing interventions. 37 The behavior change wheel relies on 3 main functions: capability (optimize psychological and physical abilities and thus enabling the behavior), opportunity (optimize social and physical environment to enable the behavior), and motivation (balancing activating and inhibiting mechanisms of the behavior). 5,37 The systematic use of, eg, the behavior change wheel in future studies could also lead us to more uniform definition of behavioral medicine interventions in physiotherapy framework. Thus, questions for research remain regarding for which patients, which components, in which order, and how many components need to be integrated to maximize patient outcomes. In addition, adherence to the intervention is important, ie, facilitating treatment integrity must be considered when designing interventions and their components for implementing in physiotherapy with a behavioral medicine approach. Being able to integrate the behavioral, psychosocial and physical aspects in assessment, analysis, and treatment and to adhere to a complex treatment approach, such as that in behavioral medicine, in both research context and clinical practice demands versatile and often new practical skills and knowledge 22,24,41 as well as new competencies in clinical reasoning. 15,17 In addition, barriers, eg, time, staff, and monetary resources, 21 as well as support from managers 1,53 in carrying out implementation in clinical practice trials and following patients in the long term must be considered in intervention designs. The most common behavior change techniques in the included trials were in the categories "information of natural consequences," "feedback and monitoring," and "goals and planning." 38 "Information of natural consequences" included information on health consequences; "feedback and monitoring" included giving feedback on behavior, self-monitoring of behavior, and the outcome of behavior; and "goals and planning" included action planning, problem solving/coping planning, goal setting for outcome and behavior, and reviewing goals. The major reported categories include many more behavior change techniques 38 that were not used in the studies. This suggests that the patient outcomes in these studies might have been positively different if more techniques were used, as previously concluded. 20,32 The low number of behavior change techniques used in the included studies might be due to limited knowledge about the techniques that could have been chosen. 31 Beyond the major categories, 7 others were identified; however, they were presented only in a few studies. 3,4,40,46,49 Why the behavior change techniques were chosen for the respective studies was not clearly stated in any of the studies. This can be associated with the fact that the results of the integration of the behavioral, psychosocial, and physical aspects in the analysis of behavior during activities could not clearly be identified in the treatment, ie, the translation of the results from the behavior analyses to the treatments was not obvious. Reasons for the purposeful selection of behavior change strategies for the treatment should be focused on future studies.
Nine randomized controlled effect trials 6,19,27,35,40,43,51,56,57 had follow-ups between 1 and 10 years and were able to show more stable positive experimental intervention effects across the studies. Interventions aiming to change the behavior of the patient most likely need to have long-term follow-ups to show the actual effects of the techniques used to change the pain behaviors.
Regarding the 3 implementation in clinical practice studies, only one study 23 reported 1-year follow-up with the difference shown in the decreased percentage of patients requiring sick leave in the behavior medicine intervention. However, this study has not yet been published and thus demands caution when drawing conclusions.

Limitations and strengths
A scoping review has both limitations and strengths. A scoping review can be seen as a preliminary assessment of available studies on the topic in question. One of the strengths is that ongoing research can be included in the review. However, this kind of research should be at least submitted for peer-review to ensure that the data collection and results have been finalized. Including research in progress can also be a limitation because conclusions from that kind of articles are not final as they are for articles that have been accepted for publication and completed a peer-review process. A limitation can also be that we did not register this scoping review, which implies that eg, the transparency of our research was not optimal. Thus, other researchers' input regarding, eg, search strategy was not possible and could have influenced the results. Further limitations of the current scoping review are the nonexistent quality assessment 11,33 and conclusions that are only descriptively summarized, which can have negative impact on the reliability of the results. However, the main aim of this review was not to investigate the effects of any intervention, where the quality assessment is very important. The main aim was to describe how the interventions matched with a definition of behavioral medicine in physiotherapy and to categorize the reported behavior change techniques for patients with musculoskeletal pain with the goal of informing future research, policies, and practice.
The search terms related to the behavioral medicine approach in physiotherapy may have biased the results because there was a large variation in the terms that could be chosen. The screening of the search results for inclusion was done by one person, which may have biased the results. However, in case of uncertainties, the study was included to the next step where full-text articles were assessed by all authors leading to agreement of final inclusion. Another limitation is that only search terms in English were used, therefore potentially relevant studies in other languages may have been excluded. In addition, the analysis of the matching of the interventions with the behavioral medicine definition may not be complete because there might exist several intervention components and behavior change techniques that were not explicitly reported by the authors of the included studies. However, the current article is a scoping review and thus does not try to give an appearance of being a comprehensive systematic review with associated meta-analysis or meta-synthesis. One more strength of this study was the use of a previously published definition of the behavioral medicine in physiotherapy, with which the included studies were matched, thus giving us quite a clear platform for comparing the interventions and definition.

Conclusion
The synthesis of the results for the matching of the patient interventions with an existing definition of behavioral medicine in physiotherapy for the randomized controlled effect and the implementation in clinical practice trials showed that the interventions mostly integrated psychosocial, behavioral, and biomedical/physical aspects, and thus complied with the existing definition of behavioral medicine in physiotherapy. The reported behavior change techniques were few and were commonly in the categories "information of natural consequences," "feedback and monitoring," and "goals and planning." The short-term patient outcomes were mixed, but the long-term follow-ups showed mostly positive effects.

Future challenges
The goal of this scoping review was to possibly inform future research, policies, and practice and not to explicitly, critically review effects for any evidence. The future challenges are, eg, that we need to keep refining interventions integrating psychosocial, behavioral, and biomedical/physical aspects, carefully choose the components and techniques that should be included and further develop these for physiotherapeutic purposes, but also, how to implement the components and techniques efficiently when integrating behavioral medicine in physiotherapy. Furthermore, beyond challenges for integrating psychosocial, behavioral, and biomedical/physical aspects, support on organizational and leadership levels in the implementation of behavioral medicine in physiotherapy should also be investigated to find the optimal contextual aspects for effective implementation.

Disclosures
The authors have no conflicts of interest to declare.