The complexity of integrating a behavioral medicine approach into physiotherapy clinical practice

ABSTRACT

Introduction and Aim: The implementation of a behavioral medicine (BM) approach in physiotherapy is challenging, and studies regarding the determinants are sparse. Thus, the aim of this study was to explore determinants of applying a BM approach in physiotherapy for patients with persistent pain across the micro-, meso-, and macro-levels. Methods: A qualitative multiple-case study design was used. Data were collected from four cases through semi-structured interviews with physiotherapists (PTs), patients, and managers; observations of video-recorded treatment sessions; and reviews of local directives and regulations. Data were analyzed with inductive content analysis and cross-case analysis, followed by mapping to the domains of determinants at the micro-, meso-, and macro-levels within the Implementation of Change Model. Results: Similar determinants were found across the cases. At the micro-level, these determinants concerned the PTs’ ambivalence toward a BM approach, a biomedical focus, embarrassment asking about psychosocial factors, BM knowledge, skills for applying the approach, and self-awareness. Others concerned the patients’ role expectations of the PT, patients as active or passive agents in the treatment process, patients’ focus on biomedical aspects, and confidence in the PT. At the meso-level, support from managers and peers, allocation of time, and expectations from the organization were identified as determinants. No determinants were identified at the macro-level. Conclusion: The complexity of integrating a BM approach into physiotherapy clinical practice arises from multiple determinants functioning as both facilitators and barriers. By selecting strategies to address these determinants, the implementation of a BM approach could be supported.

KEYWORDS:

• Behavioral medicine
• physiotherapy
• clinical practice
• determinant
• facilitator
• barrier

Introduction

Disability related to persistent musculoskeletal pain is influenced by a complex behavioral learning process. Behaviors are affected by emotions, cognitions, and the environment when an individual is struggling to cope with pain. Some behaviors can be beneficial in the short term but detrimental in the long term (Gatchel et al, 2007; Linton and Shaw, 2011). Thus, pain-related behaviors should be a focus of rehabilitation. A behavioral medicine (BM) approach to physiotherapy is therefore useful in assessing and treating patients with persistent musculoskeletal pain. This approach is recommended in the updated report by the Swedish Agency for Health Technology Assessment and Assessment of Social Services (2010). The recommendation is based on evidence supporting the use of psychosocial and behavioral strategies in physiotherapy (Åsenlöf, Denison, and Lindberg, 2005; Friedrich, Gittler, Arendasy, and Friedrich, 2005; Gatchel et al, 2007; Åsenlöf, Denison, and Lindberg, 2009), and later studies have strengthened the evidence (Brunner et al, 2013; Linton and Shaw, 2011; Williams, Eccleston, and Morley, 2012). BM is based on a biopsychosocial model of health and is an interdisciplinary field that integrates health-related biomedical, psychosocial, and behavioral knowledge to facilitate health promotion, diagnosis, treatment, and rehabilitation (International Society of Behavioural Medicine, 2014). A systematic method is necessary for incorporating a BM approach into physiotherapy clinical practice. Åsenlöf, Denison, and Lindberg (2005) describe a process model for the systematic application of a BM approach to physiotherapy clinical practice that is based on cognitive behavioral principles (Turk and Okifuji, 1993). The process model focuses on changing the patient’s behaviors that are relevant for goal attainment related to disabilities in daily life. Physical, psychological, and contextual factors are analyzed, and treatment is tailored to the individual patient. Support and maintenance of behavioral change are prioritized rather than merely symptom reduction (Åsenlöf, Denison, and Lindberg, 2005).

A fundamental problem for evidence-based healthcare is professionals’ poor adoption of new methods and guidelines. Similarly, the implementation of a BM approach to clinically relevant behaviors in physiotherapy is challenging (Gray and Howe, 2013; Sanders, Foster, Bishop, and Ong, 2013; Synnott et al, 2015). Physiotherapists (PTs) seem not to change their traditional biomedical method of working after participating in e.g., academic or other courses in BM. Some changes in knowledge, beliefs, and attitudes occur, but changes in practice are less forthcoming (Overmeer, Boersma, Main, and Linton, 2009; Sandborgh, Åsenlöf, Lindberg, and Denison, 2010; Stevenson, Lewis, and Hay, 2006). Over the last 10 years, increasing attention has been focused on improving the integration of clinical guidelines into healthcare and physiotherapy (Grimshaw et al, 2012; Jones et al, 2015). Different frameworks for implementation in clinical practice highlight the importance of exploring determinants of professional change (i.e., factors that may act as facilitators or barriers to implementation) to address when planning an implementation intervention (Baker et al, 2015; Flottorp et al, 2013; Grol, Wensing, Eccles, and Davis, 2013).

In the Implementation of Change Model, a problem analysis of the target group and setting, including an inventory of determinants, is stated as important (Grol, Wensing, Eccles, and Davis, 2013). Pope et al, (2006) claim that implementation is a multilevel phenomenon. The determinants of change in professional practice are related to different domains. These domains can be traced to the micro-, meso-, and macro-levels (McLaren and Hawe, 2005) (i.e., the perspective of the individual PT and his/her interactions with the patient, the workplace, and healthcare policies). At the micro-level, the domains concern the specific innovation (e.g., underlying research evidence, the complexity of the innovation, and how it matches current practice); individual health professional (e.g., knowledge and skills, attitudes, motivation, and individual characteristics); and patient (e.g., beliefs, knowledge, and motivation). At the meso-level, the domains concern professional interactions (e.g., colleagues, leadership, and organizational structure); and incentives and resources (e.g., the availability of financial resources and incentives). At the macro-level, the domain concerns social, political, and legal factors (e.g., rules, regulations, and laws). They also state that some domains are generic and others are related to the specific innovation or context. Dannapfel, Peolsson, and Nilsen (2013) found determinants of using research in physiotherapy clinical practice in Sweden at the individual level (i.e., attitudes and motivation concerning research use, research-related knowledge and skills); workplace level (i.e., leadership support, organizational culture, research-related resources, and knowledge exchange); and extra-organizational level (i.e., evidence-based practice guidelines, external meetings, networks and conferences, academic research, and education). However, studies of determinants of applying a BM approach in physiotherapy clinical practice are sparse. The aim of this study was to explore determinants of applying a BM approach in physiotherapy for patients with persistent pain across the micro-, meso-, and macro-levels.

Methods

Design, sampling, and participants

The study used a qualitative multiple-case design to explore possible determinants from different perspectives within a natural context (i.e., in a PT setting) (Yin, 2014). The domains of determinants within the Implementation of Change Model (Grol, Wensing, Eccles, and Davis, 2013) influenced the composition of cases that covered the micro-, meso-, and macro-levels. The cases consisted of a PT with a BM education (the individual health professional domain), a patient with persistent pain (the patient domain), and a manager (the professional interaction domain), acting in a context guided by local directives and regulations (the domain of incentives and resources and the domain of social, political, and legal factors). Multiple cases were chosen to extend and replicate emergent findings (Yin, 2014). The cases were selected to represent diverse environmental factors in the context of physiotherapy for patients with persistent pain: differences in county council affiliation, primary healthcare and hospital care, private and county council practice, and BM education.

A purposive sampling of clinically working PTs with a BM education was used (Polit and Beck, 2012). These PTs had graduated from an undergraduate physiotherapy program or had completed advanced courses in physiotherapy at a university in Sweden. In both types of physiotherapy education, the process model for a BM approach (Åsenlöf, Denison, and Lindberg, 2005) is integrated. The participants were selected 1 to 2 years after their graduation/completion of courses. One of the PT’s patients of working age with persistent pain treated by the included PT was asked to participate in the study as well as the PT’s manager. Four cases were included and are presented in Table 1. Participation was voluntary, and all participants gave their written informed consent after receiving oral and written information. Furthermore, they were allowed to withdraw without any consequences. The study was approved by the Regional Ethical Review Board, Uppsala, Sweden, Dnr 2012/017.

Table 1. Characteristics of the four included cases.

Case 1	Case 2
County council A, primary healthcare	County council B, hospital clinic
PT 1: undergraduate physiotherapy program with a BM approach; had worked for 12 months	PT 2: undergraduate physiotherapy program with a BM approach and an advanced course in BM; had worked for 9 months
Patient 1: pain duration, 40 years	Patient 2: pain duration, 10 years
Manager 1: occupational therapist, manager of the healthcare center; had worked as a manager for less than 1 year	Manager 2: PT, manager of PTs and occupational therapists; had worked as a manager for 21 years
Case 3	Case 4
County council B, primary healthcare	County council B, primary healthcare, private practitioner
PT 3: undergraduate physiotherapy program with a BM approach; had worked for 14 months	PT 4: advanced courses in BM; had worked for 12 years
Patient 3: pain duration, 10 months	Patient 4: pain duration, 6–7 years
Manager 3: PT, manager of PTs, and occupational therapists; had worked as a manager for 6 years

PT = physiotherapist.

Data collection

Data were collected through semi-structured interviews with the PTs, the patients, and the managers; through observations of video-recorded treatment sessions; and through reviews of documents regarding local directives and regulations. The different sources of data were used to stimulate a comprehensive understanding (Polit and Beck, 2012; Yin, 2014). All data were collected by the first author.

The data collection was conducted for each case as follows: (1) face-to-face interview with the PT before the treatment session, (2) video recordings of two treatment sessions, (3) telephone interview with the patient the day after the video-recorded treatment session, (4) face-to-face interview with the PT based on joint observation of selected video sequences from the recorded treatment sessions 1 week after the last video recording, (5) formal documents regarding local directives and regulations, and (6) face-to-face interview with the manager.

Video sequences of the recorded treatment sessions were used to facilitate the researcher’s observation and to stimulate the PTs’ recall during the interviews (Paskins, McHugh, and Hassell, 2014). For each case, two treatment sessions in the middle of the patient’s treatment period were recorded. The sequences were selected from the recordings if they concerned the core BM components in the process model by Åsenlöf, Denison, and Lindberg (2005) (Table 2). Three test video recordings were made to ensure that the authors consistently identified core components. Twenty-eight sequences were identified for analysis from the video recordings of the four cases; for details, see Table 2. The sequences were selected by the first author and validated by the last author.

Table 2. The number of selected sequences per core BM component from the video-recorded treatment sessions in relation to the cases.

Core BM components	The number of selected sequences	Cases
Identification and assessment of behavioral goal	7	1, 2, 3
Self-monitoring of target behavior	8	1, 2, 3, 4
Individual functional behavioral analysis	2	3, 4
Basic skills acquisition	8	1, 2, 3, 4
Applied skills acquisition	3	1, 2, 4
Maintenance and relapse prevention	0

The interview questions concerned the participant’s experiences of applying a BM approach in the recorded treatment sessions and in general. Four test interviews were conducted—two with PTs and two with patients—to refine the interview guide. Before the treatment sessions, the PTs were asked about their intentions for the treatment session. In conjunction with viewing the recorded video sequences, they were also asked about their experiences of applying a BM approach in relation to the situations in the recorded sequences. The patients were asked about their expectations of the previous day’s treatment session and whether they were met. They were also asked what they believed affected their pain-related disorders. The managers were asked about organizational factors that guided the PT’s work and about experiences of applying a BM approach in physiotherapy at their clinic.

Demographic data (i.e., age, sex, education, pain duration, years as a PT/manager) were collected at the beginning of each interview. The interviews lasted between 15 and 60 min and were digitally audio-recorded. All face-to-face interviews were conducted in the clinic. The managers and PTs were asked which formal local directives and regulations guided their work. Business plans, clinical guidelines, and rules and mission statements for primary healthcare were included.

Data analyses

The first author transcribed the selected video sequences from the recorded treatment sessions as observation narratives, and all interviews were transcribed verbatim. Inductive content analysis (Elo and Kyngäs, 2008) and cross-case analysis (Yin, 2014) were used. The transcribed interviews, the observation narratives, and the documents regarding local directives and regulations were read several times for familiarization. Sensitive meaning units related to the aim were identified in the manifest content of the texts. These meaning units were identified by the first author and validated by the other authors. The meaning units were decontextualized and placed in a document including meaning units from all texts. In the analytical process, the meaning units were coded and grouped according to similarities into subcategories and categories, based on abstraction of the manifest content. The categories illustrate determinants of the application of the BM approach. The cases were analyzed one at a time and then merged together in the cross-case analysis. The identified determinants across the cases were finally mapped to the domains of determinants in the Implementation of Change Model (Grol, Wensing, Eccles, and Davis, 2013). Examples of the abstraction process from meaning units to subcategories and categories, and the mapping to the domains are shown in Table 3.

Table 3. Example of the abstraction process: meaning units, codes, subcategories, categories, and domains of determinants in the Implementation of Change Model (Grol, Wensing, Eccles, and Davis, 2013).

Meaning unit	Code	Subcategory	Category	Domain
PT: The advantage of this exercise is that when you find that you can do it well, then you will also be ready to climb the stairs. (The PT is stepping up and down on a step.) Because on principle, it is the same movement. You break it down and practice it because it is too tough for you to climb a regular staircase. Therefore, we break it down and take small parts of it and practice it. (Observation narrative, case 4)	This exercise is a step toward climbing the stairs	Relates treatment to goal	Pedagogical skills	Determinants related to the individual PT
PT: How about (scratches his nose, looking down at the papers), for example, you use it like a diary? (Looking up at the patient.) Patient: Mm. I think that’s good. (Nods) It’s really hard to remember. PT: Exactly. (Looking down at the paper, demonstrates with the pen on the paper.) If you are writing, like, “then I did this”. Patient: Yes. PT: And then you rate it on a scale of 0–10 or write down how it felt, “well, this was not very painful, so I rate it a 3.” (…) So, take a paper at home and fill it in a bit and take notes… (Observation narrative, case 1)	Introduce self-monitoring with a diary. Unclear instructions and body language	Vague communication
She is good at what she does. She manages her professional role so I don’t have to worry. I go there and get answers to my questions and get help with what I need help with. (Patient 4)	Manage the professional role	Feels confidence in the PT	Confidence	Determinants related to the patient
He (the PT) sees the person. He sees all who come there. When I get there, he is talking to me, asking me how I feel and what I think. (Patient 3)	Sees the person	Interaction between the patient and PT
Yes, I understand that it is in the spine, that my discs are bulging. It is a question of the discs; the nerves get pinched in the spine, so to speak. When you get a crick, there is a hernia and so on pressing on a nerve. (Patient 1)	The pain is coming from the spine or the muscles	Gives a physical explanation	Biomedical focus
But I have the pain and the only thing I know is that I should lose weight. It is common sense that if you are thinner and do not weigh too much, it affects the body. That should be better for the body. (Patient 4)	Being overweight affects the body

The analysis was performed one case at a time and thereafter synthesized across multiple cases to clarify the extension and replication of the findings. Triangulation in the analyses concerned the multiple perspectives (Polit and Beck, 2012; Yin, 2014) of the patient, the PT, and the management, which consisted of the manager and the documents regarding local directives and regulations. The analysis was regularly discussed and validated among the authors to achieve consensus.

Results

The results revealed potential determinants of integrating a BM approach into physiotherapy clinical practice referring to the following domains in the Implementation of Change Model (Grol, Wensing, Eccles, and Davis, 2013) at the micro- and meso-levels: the BM approach, the individual PT, the patient, the professional interactions, and the incentives and resources, as shown in Table 4. No potential determinants were identified at the macro-level, i.e., related to the domain of social, political, and legal factors. Fourteen potential determinants were identified, and thirteen of those were replicated in at least three of the four cases. Case 2 was distinguished by only replicating 11 of the identified determinants. No extension of categories was found in relation to case 1, only replications. The results are presented using the domains as headings. Both quotations and observation narratives are used as descriptions from the four cases.

Table 4. Domains and categories related to the determinants of applying a BM approach at the micro and meso levels based on the within-case and cross-case analyses. The sources and cases are shown in brackets: PT=physiotherapists, P=patients, M=managers, D=documents regarding local directives and regulations, O=observations, ¹ =case 1, ²=case 2, ³=case 3, ⁴=case 4.

MICRO
Factors related to the BM approach
Ambivalence towards a BM approach (PT^1,2,3,4, P^1,3,4, M^1,3)
PT factors
Biomedical focus (PT^1,3,4, O^1,3,4)
Embarrassment asking about psychosocial factors (PT^1,3)
BM knowledge (PT^1,2,3,4, M^1,3, O^1,2,3)
Skills for applying the BM approach (PT^1,2,3,4, M^1, O^1,3)
Pedagogical skills (PT^2,3,4, M^1,3, O^1,4)
Self-awareness (PT^1,2,3,4)
Patient factors
Role expectations of the PT (P^1,3,4)
Patients as active and passive agents (PT^1,4, P^1,4, M^1,3, O^2,4)
Biomedical focus (P^1,2,3,4)
Confidence in the PT (P^1,2,3,4, M^3)
MESO
Professional interactions
Support from manager and peers (PT^2, M^1,2,3)
Incentives and resources
Allocation of time (PT^1,2,3, M^1,2,3, D^1,4)
The organization’s expectations (M^1,2,3, D^1,2,3,4)

Determinants related to the BM approach

One identified determinant concerned ambivalence among the PTs toward the BM approach. This ambivalence was found in all cases.

Ambivalence toward the BM approach

The BM approach was used in daily work as a tool when needed rather than an approach that informed the entire patient encounter. According to the managers, the PTs were free to choose their practice approach based on their professional knowledge and personal interest. It was unclear whether the PTs were willing to change to a BM approach in their practice. Parts of the BM approach (e.g., goal setting, goal reviewing, and self-monitoring) were experienced as beneficial to clinical practice. Ensuring that goals were measurable, related to daily activities, and important to the patient were useful aspects of goal setting. The PTs described that self-monitoring in diaries helped to motivate the patients by highlighting when they had been physically active. Additionally, self-monitoring clarified for the patients what affected their pain:

I think it is very useful for patients like him who might find it a little difficult, as we saw in these earlier video sequences, to put his finger on what exactly he is doing when he gets more pain. So it may be important for him to pay attention to that himself. (PT 1)

However, the PTs described that they were not convinced that the focus on behaviors related to daily activities was relevant and sufficiently beneficial for the patients. Patients had difficulty understanding the relevance of focusing on behaviors instead of pain alleviation, as well as in expressing how their pain-related disorders affect their daily living. The expectation was that the patients wanted a simple explanation and solution:

That is also because they (patients) have difficulties understanding the relevance. Usually, they have pain, and they want less pain. Then it might be hard to understand why one should focus on daily activities. (PT 2)

The PTs were also concerned that the approach was time-consuming. Parts of the BM approach did not require a great deal of preparation and thus were easier to use. Other parts, particularly those that require writing things down, were time-consuming. Thus, the parts that could be performed orally were preferred. The PTs considered it difficult to motivate the patients when they were not motivated themselves to use the BM approach:

So you first need to truly get the patient on board. And if I’m not convinced myself that it is so very important that we select activities, then it will be very difficult to convince a patient of it. (PT 4)

Determinants related to the individual PT

Several identified determinants were related to the individual PT. Some concerned attitudes regarding a biomedical focus and embarrassment asking about psychosocial factors. The biomedical focus was found in all cases, while embarrassment was only found in cases 1 and 3. The results also revealed determinants concerning the PTs’ BM knowledge, pedagogical skills and skills for applying the BM approach, and self-awareness, which were found in all cases.

Biomedical focus

Both the observations and interviews indicated that physical explanations for the movement disorders were dominant among the PTs, and the treatment focused on physical functions:

(About the treatment) It’s a mix of motor control, balance, stability exercises and some light weight lifting, and also exercises to get the muscles to relax. Some self-administered trigger point therapy: like standing with a ball against a wall, some stretching and so on. The strength training is about pelvic lifts, to start using and strengthen these muscles. (PT 1)

Embarrassment asking about psychosocial factors

The PTs found it embarrassing to ask the patient about home situations, thoughts, and emotions. Before asking, they felt they needed to build a relationship with the patient:

More about the home situation maybe, emotionally and lot of those things. This is maybe not something that you talk about in the beginning, but you build up a relationship with the patient first, to avoid badgering about this sensitive stuff too early, which might be needed in the end to get the puzzle together. (PT 1)

BM knowledge

The PTs’ knowledge about the BM approach was evident. They demonstrated intentions to identify important biopsychosocial aspects for the patients’ behavior change by asking about environmental, social, and psychological factors in terms of expectations of the treatment and avoidance of daily activities. They talked about how the patients’ pain-related disorders affected their behaviors in daily life:

PT: I was thinking, you paint sometimes? How much is it in the springtime? You finish your paintings for the exhibition, but… Patient: I have not had the time now. PT: I was thinking, when you paint, are you standing up or sitting down? Patient: Both. It depends because I have difficulties bending forward in a certain position because of my lower back problems. PT: Yes. Patient: However, I am standing too. Bending. However, mostly I sit. (Observation narrative, case 3)

Skills for applying the BM approach

The PTs had the knowledge and skills to tailor treatment to the individual patient and support behavior change. This was done by making individual adjustments to the treatments, problem-solving, and using prompts and reinforcement:

PT: If you consider how much you want to exercise, what is your plan? Patient: Well, in the beginning, there will be one hydrotherapy session a week and then maybe one session here. Then we’ll see about the future, if there will be maybe gym exercises twice a week and hydrotherapy once a week. PT: Yes, exactly, as a progression. Patient: To begin with, anyway. Then, we will see how it develops. PT: That’s a great plan. Patient: Yes, I think it seems like a good plan, really. PT: Does it feel like you would be able to manage this? Patient: Yes. PT: Good. And if you think about the time of day? For now, you have an appointed time for the hydrotherapy. However, this gym training, will it be a particular day? Patient: It will be sometime before the hydrotherapy, anyway. PT: Are you thinking of the same day? Patient: No, a day or so before. If I do the gym training on Mondays or something like that, maybe. PT: And what about your kids and other stuff, is there any time that works better considering them? Patient: No, my wife is at home. PT: Yes, exactly. Patient: She’s on maternity leave, so she’s at home with the kids. So regarding the kids, it doesn’t matter. I have to get going so that I will recover some day. PT: Does your wife support you in this? Patient: Yes, I think she does. (Observation narrative, case 2)

However, the PTs’ skills in applying other parts of the BM approach were incomplete. The PTs experienced difficulty applying the BM skills in practice and integrating the BM approach into their daily work. They talked about how the patients’ pain-related disorders affected behaviors in daily life, but the video sequences showed that they lost their focus on this during the encounters and instead mainly focused on how the symptoms could be reduced. It was deemed important by the PTs to offer the patient understandable explanations about causes and correlations. However, the PTs found this to be complex from a biopsychosocial perspective, and they indicated that it was easier to take a physical perspective.

The challenge is to not ignore the patient’s environment. It is so easy to just focus on one thing, whatever it may be - knee or something. (PT 3)

Pedagogical skills

The pedagogical skills mainly related to the ability to create patient-centered communication and motivate the patients. One PT described how the treatment was related to the goal:

PT: The advantage of this exercise is that when you find that you can do it well, then you will be ready to climb the stairs, too. (The PT is stepping up and down on a step.) Because on principle, it is the same movement. You break it down and practice it because it is too tough for you to climb the regular staircase. So we break it down and take small parts of it and practice it. (Observation narrative, case 4)

Additionally, more vague communication skills were found. This was apparent, for example, in not following up on earlier conversations and providing vague instructions for exercises:

PT: How would it be to (scratches his nose, looking down at the papers), for example, a little bit like a diary? (Looking up at the patient.) Patient: Mm. I think that’s good. (Nods) It’s really hard to remember. PT: Exactly. (Looking down on the paper, demonstrates with the pen on the paper.) If you are writing, like, ‘this I did then’. Patient: Yes. PT: And then you rate it on this scale of 0-10 or how it felt: well, this was not so very painful, so I rate a 3. (…) So take a paper at home and fill in a bit and take notes… (Observation narrative, case 1)

Self-awareness

The PTs were unaware of their own actions. When they watched the video sequences, they became aware that they did not use the BM approach as much as they had thought. They indicated that seeing themselves on video was beneficial.

When I look at these video recordings, I do not think I do it. Not as much as I thought I did. I thought I had a greater focus on the patient and his/her experiences. However, it seems not to be so. (PT 1)

Determinants related to the patient

Several determinants were related to the patient. They concerned the patients’ role expectations of the PT, which was found in cases 1, 3, and 4, and patients as active or passive agents in the treatment process, which was also found in all cases. Additionally, the patients’ attitudes regarding a focus on biomedical aspects and the patients’ confidence in the PT were found in all cases.

Role expectations of the PT

The patients demonstrated ambivalent expectations of the PTs. On the one hand, they expected the PT to act as a coach; on the other hand, they expected him or her to act as a doer. The patients considered the PTs’ coaching important and expressed the need for motivation to exercise. They wanted the PTs to explain the causes of the pain-related disorder, how to treat it, and how to adjust the exercises; furthermore, the patients wanted reinforcement from the PTs:

He (the PT) sees things that I haven’t noticed. I mean, in my improvements. That is positive for me, how he acts in this, because I get some praise, so to say. (Patient 1)

Expectations of the PT as a doer meant that the patients felt it was the PTs’ responsibility to resolve the pain-related disorder:

It is he (the PT) who should push me and tell me… It is he who is the expert and then, of course, I will obey him. And do as he says, even if the exercises feel awkward. (Patient 1)

Patients as active or passive agents

The patients expressed a willingness to take responsibility for their pain-related disorders. They wanted to be active in the treatment process and do things by themselves to achieve treatment results. Contrary to the patients’ statements, the PTs and managers viewed patients as having a passive attitude to treatment. The PTs and managers stated that the patients did not want to do very much themselves and that the patients preferred hands-on treatment.

Of course, I must do something myself; otherwise, nothing will happen. I cannot just go there once a week/fortnight and think that it will get better. I have to do the homework myself. I believe that ninety percent of it is what you do by yourself, both physically and mentally. (Patient 1)

Biomedical focus

Both the observations and interviews indicated that physical explanations for the movement disorders were also dominant among the patients. The patients explained that the pain was caused by a disease or injury, heredity or from being overweight:

Yes, I understand that it is in the spine, that my discs are bulging. It is a question of the discs; the nerves get pinched in the spine, so to speak. When you get a crick, there is a hernia and so on pressing on a nerve. (Patient 1)

Confidence

Confidence concerned the patients’ attitudes toward the PTs and was based on trust. The patients understood that rehabilitation of their pain-related disorder would take time. They trusted the PT’s knowledge and the manner in which the PT interacted with the patient (e.g., showed interest, took time, and saw the person behind the pain disorder).

He (the PT) sees the person. He sees all who come here. When I get here, he is talking to me, asking me how I feel and what I think. (Patient 3)

Determinants related to professional interactions

The determinants related to professional interactions concerned support from managers and peers, which were found in cases 1, 2, and 3.

Support from managers and peers

The managers wanted to support a BM approach, but felt that they could not provide sufficient support for initiating new clinical guidelines and problem-solving during implementation. They wanted to provide better support but expressed that it should also come from someone outside the clinic. Additionally, the managers claimed that peer support at the clinic facilitated the PTs’ methods of practice. It was considered that there was openness and mutual support among the PTs for discussing problems in daily practice:

The PTs who work at the clinic and I, we have weekly meetings where we regularly work on the different problems at the clinic. A part of this is, of course, to support each other. And they do a lot. They have many discussions with each other about different patients and get help and support: “How would you address this,” and “What do you think?” (Manager 3)

Determinants related to incentives and resources

The determinants related to incentives and resources concerned allocation of time and expectations from the organizations, which was found in all cases.

Allocation of time

Flexibility in the use of time was evident in the ways the PTs used the time in the clinic. The time allotted for first visits was longer than for follow-up visits. Additionally, the number of treatment sessions was flexible on the PTs’ part and was adapted to the patient’s needs, which meant that the PTs had control over their work schedule.

I don’t interfere much. I think they (the PTs) manage; they control it pretty much by themselves. If they see that they need more time, they take more time. So the time schedule, you control by yourself as a physiotherapist. (Manager 1)

In clinical practice, workloads are high, and at times, there was a waiting list of patients. Lack of time prevented preparation and reflection when using a BM approach.

Of course, we may miss something here, some goals along the way. This is something you should have caught and might catch if you are more prepared for it and think of it at that moment, so that it does not just become so improvised. Then, you might have been able to get more out of it. (PT 3)

One important change is that we need to make time in the schedule for evaluation and reflection. Today, the pace of work is too high to achieve this in a constructive way. (Business plan, case 1)

The organizations’ expectations

Expectations of clinical competence were found both from the managers and in the local directives and regulations. Professional development for the PTs was planned jointly by the PT and the manager and was guided by the clinic’s needs. Some clinics had prioritized certain courses that all PTs should have (e.g., courses in motivational interviewing, BM, and pedagogy). There was no structure regarding how the PTs would keep up with the latest research; that was each PT’s own responsibility.

But, what will happen now is that we want our physiotherapists to have a behavioral medicine orientation, too. We highlight that because it fits so well with what we do at the clinic. And pedagogy is also important, to develop that part, because we teach a lot. Our role is more coaching, much more than it was before. (Manager 2)

The local directives and regulations indicated explicit expectations for the use of evidence-based methods, but the biopsychosocial focus varied. Clinical guidelines demonstrated a biopsychosocial focus, while local directives focused on physical exercise:

Treatment, instructions regarding exercise and exercise as a controlled medical treatment shall be conducted by physiotherapist/equivalent with financial compensation from the County Council. Different medical interventions, advice and information can here be included. (Rulebook for Rehabilitation, case 3)

In documents regarding local directives, so-called multimodal rehabilitation was advocated as a biopsychosocial treatment model that should be used in primary care. The county councils used reimbursements to promote selected treatment methods, and multimodal rehabilitation was one of these. However, according to the manager, the downside of the reimbursement model was that recording the number of patient visits became more important than quality.

Discussion

The results demonstrate the complexity of integrating the BM approach into physiotherapy clinical practice. This complexity consists of multiple determinants related not only to the individual PT but also to contextual factors at the micro- and meso-levels. This complexity became obvious when mapping the identified determinants to the domains in the Implementation of Change Model. The domains include explored determinants that could function as both facilitators and barriers. The cases were selected to represent diverse environmental factors present in the context of physiotherapy for patients with persistent pain. The results did not show any differences between the cases that could be explained by the diversity in the sample. In fact, the results primarily revealed similarities between the cases.

Although clinical studies show evidence for a BM approach and that the participating PTs had received a BM education, a biomedical focus remained among the PTs. A biomedical focus was also dominant among the patients and in documents regarding local directives and regulations. Sanders, Foster, Bishop, and Ong (2013) found that the integration of psychosocial factors into the PTs’ traditional biomedical approach complicated clinical practice, implying that the BM approach itself contributes to complexity. As in the review by Synnott et al, (2015), the patients perceived that their movement disorders could be explained as a purely physical problem. It is therefore possible that the patients did not perceive the PT’s biopsychosocial message. Overmeer and Boersma (2016) found that catastrophizing and depression correlated with “not perceiving the biopsychosocial message.” This indicates that the patients who most need the biopsychosocial approach are those who have the greatest difficulty perceiving this message. Overmeer and Boersma (2016) recommend providing clear information and asking the patient to repeat the message to facilitate the patient’s perception of a biopsychosocial approach. The patients’ biomedical expectations influenced the PTs’ attitude and inhibited their use of a biopsychosocial approach. Identifying psychosocial factors remains an uncommon skill for PTs, which makes it difficult for PTs to provide a biopsychosocial explanation for behavioral problems that is plausible and understandable by the patient. Focusing on biomedical aspects is an easier solution and may feel more comfortable for the PTs, as other studies have indicated (Gray and Howe, 2013; Sanders, Foster, Bishop, and Ong, 2013; Synnott et al, 2015). Additionally, the local organization’s expectations of physiotherapy showed a varied biopsychosocial focus. However, these expectations were identified when reviewing documents of local directives and regulations and when interviewing the managers. Because they were not mentioned by the PTs themselves, whether these expectations actually function as determinants is debatable.

The PTs in the current study had been educated in the BM approach; nonetheless, the biomedical focus dominated. Changes in PTs’ attitudes and knowledge as a result of education have been demonstrated in earlier studies, but changes in practice behavior have been less forthcoming (Overmeer, Boersma, Main, and Linton, 2009; Sandborgh, Åsenlöf, Lindberg, and Denison, 2010; Stevenson, Lewis, and Hay, 2006). In our study, the PTs conveyed their knowledge of the BM approach in the interviews, but their skills for applying the approach were incomplete, as found in the observations. Although the participating PTs were educated according to a process model that promoted the systematic use of a BM approach in physiotherapy clinical practice (Åsenlöf, Denison, and Lindberg, 2005), the application was problematic. Some core components of the process model, such as individual functional behavior analyses, supporting applied skills acquisition and maintenance, and relapse prevention, were infrequently used, although the PTs thought that they used the BM approach to a greater extent than they actually did, as shown by the observations. Several studies have shown that a lack of skills in clinical practice is a common barrier to using a biopsychosocial approach in physiotherapy (Foster and Delitto, 2011; Sanders, Foster, Bishop, and Ong, 2013; Synnott et al, 2015). Foster and Delitto (2011) highlight the challenge of embedding a psychosocial approach into educational programs to promote changes in professional skills. Sanders, Foster, Bishop, and Ong (2013) state that a lack of skills could contribute to clinicians’ insecurity regarding asking patients about psychosocial factors. In our study, the PTs said it was embarrassing to ask about psychosocial factors, which prevented application of the BM approach in practice. There were no data from other data sources that illuminated why the PTs felt this embarrassment. It is possible that questions about colleagues’ influences could provide an explanation. However, the PT in case 2 did not mention that it was embarrassing to ask about psychosocial factors and also applied a biopsychosocial approach. This was possibly due to this PT possessing the necessary clinical skills, as suggested by Sanders, Foster, Bishop, and Ong (2013). Thus, continuous support, including feedback and reinforcement regarding the PTs’ clinical behavior, is vital to foster the clinical skills needed to apply and actually use a BM approach.

The patients’ role expectations of the PT as a “coach” and a “doer” were not found among the determinants reported in previous studies. Foster and Delitto (2011) noted that PTs’ treatment method choices are affected by the patient’s expectations. If the patient expects a certain treatment approach, the PT tends to comply with these expectations. In our study, the patients expected the PT to be both a coach and a doer. The BM approach implies active patient involvement (Åsenlöf, Denison, and Lindberg, 2005), which is hindered when the PT acts only as a doer who will fix the patient’s problem. In clinical situations, the challenge for the PT will be to balance the patient’s expectations that the PT is responsible for solving the problem with transferring responsibility for the treatment to the patient.

There is an expectation among the general population that individuals should take individual responsibility for the management of musculoskeletal disorders (Larsson and Nordholm, 2008). Such attitudes may influence patients’ attitudes toward taking responsibility, either by encouraging the patients to take an active role in treatment or by causing them to overestimate their capacity or willingness to be actively involved because that is what was expected of them. Nevertheless, the expectation regarding taking responsibility for one’s own rehabilitation could facilitate the use of a BM approach and act as an incentive for the patient to be an active agent in the treatment. Contrary to what the patients expressed, the PTs and managers viewed the patients as passive receivers of treatment. Synnott et al, (2015) noted that PTs’ perceptions of patients as passive receivers may be modifiable and are therefore important to identify. It is also possible that the PTs in our study generalized this view of patients as passive receivers of treatment to all patients. As the results show conflicting information regarding views on the patient as an active or passive agent in the treatment, this should be further studied. It is important to clarify whether this passivity is an expectation that the PTs attribute to the patients or if this reflects the patients’ actual expectation to determine how to direct implementation strategies.

Lack of time was highlighted as a determinant that created a barrier to a BM approach. Flottorp et al, (2013) described the availability of necessary resources as a determinant for implementation in clinical practice. Time is a necessary resource, and the lack of it affected the PTs’ attitudes regarding the feasibility of the BM approach. Although the patients understood that the rehabilitation process takes time, that attitude was not a strong enough facilitator to influence the PTs’ clinical practice. Components of the BM approach that the PTs perceived as quick to apply were used, and components that required some preparation were rejected. It is likely that regarding the BM approach, reluctance to perform time-consuming tasks influences PTs’ clinical practice more than the long-term benefits for the patients (Friedrich, Gittler, Arendasy, and Friedrich, 2005; Åsenlöf, Denison, and Lindberg, 2009). Time limitations have been expressed as a barrier against including a psychosocial approach in PT clinical practice. Furthermore, time pressure further encourages the use of a biomedical approach because it is the approach that is most familiar to PTs (Foster and Delitto, 2011). Support from peers and management was mentioned as a possible facilitator to a BM approach. It is important that the support, especially from the manager, also includes sufficient preparation time.

No potential determinant was identified in this study in the domain of social, political, and legal factors (Grol, Wensing, Eccles, and Davis, 2013). The parts of the formal local directives and regulations that concerned PT’s clinical practice were often written in general terms without a clear connection to a BM approach. Thus, determinants specific to the BM approach could not be identified at the macro-level. According to previous research, implementation is a multilevel phenomenon (Pope et al, 2006). However, in our study, only determinants at the micro- and meso-levels could be identified. It remains unclear which determinants should be targeted at the macro-level when implementing a BM approach in physiotherapy.

Strengths and limitations

The use of a case study design made it possible to explore determinants of using a BM approach in physiotherapy linked to a specific treatment session, which is influenced by the actual situation and context. This was considered important because the Implementation of Change Model claims that possible determinants could also be found in the context (Grol, Wensing, Eccles, and Davis, 2013).

The case study design was inspired by the Implementation of Change Model (Grol, Wensing, Eccles, and Davis, 2013) and the domains of determinants added a systematic approach to the data collection by guiding the composition of the cases. The model proposes various domains of possible determinants at the micro-, meso-, and macro-levels that influenced the inclusion of not only the PT but also the patient and the management in each case. The use of different sources enabled multifaceted and rich data and led to a comprehensive understanding of the determinants of applying a BM approach in these clinical settings. The supportive function of the PTs’ colleagues when a BM approach is implemented was mentioned by the managers. Including colleagues in the study might have added some valuable information and should be taken into account in future studies.

Using multiple cases helped to identify similarities between cases. Thirteen of the fourteen categories were found in at least three of the four cases. This indicates a high validity of the findings (Yin, 2014). The information from the four cases was considered to have sufficient information power (Malterud, Siersma, and Guassora, 2016). The aim was narrow and concerned a very specific phenomenon that is not widespread among PTs. The sample specificity was dense and was based on the purposive sampling of PTs and patients with specific experiences pertinent to the study aim, which generated abundant data. The sample also represented various clinical settings and included both primary and hospital care. Because the interviewer has comprehensive knowledge of the BM approach, the quality of the interview dialog was strong and several test interviews and observations were conducted in advance, which further supports the quality of the interviews.

There are no standard methods for identifying determinants of professional change. Interviews, direct observations, and standardized questionnaires are commonly used (Grimshaw et al, 2012; Grol, Wensing, Eccles, and Davis, 2013). Interviews with PTs, patients, and managers; observations of video-recorded treatment sessions; and analyses of documents regarding local directives and regulations in this study gave a deeper understanding of the participants’ experiences that could not have been captured using questionnaires alone.

The joint viewing of the video-recorded treatment sequences by the PT and the first author during the interview contributed to the trustworthiness of the findings. This shared viewing helped to focus the interview on the specific patient and situation. During the interview, the PT was able to watch the video-recorded sequences twice to support the strength of the stimulus. Two treatment sessions were recorded in each case to allow the PT and the patient to get used to the camera and to reduce the influence of recording on the participants’ behaviors. The patients seemed untroubled during the recorded sessions, and the PTs said that they forgot about the camera after 5 min. A limitation of the current study was that the video recordings of the treatment sessions were limited to the middle of the treatment period. Some core components of the BM process model are associated with sessions at the beginning or the end of the treatment period and occurred less frequently among the identified video sequences. This means that determinants of some of the core components were discussed to a lesser extent or not at all during the interviews. In hindsight, it would have been preferable to video-record one initial session and one session in the middle or at the end of the treatment period.

Because of their BM education, the participating PTs were well informed about the BM approach, and they could likely guess which BM components the study focused on. It is possible that this knowledge influenced their clinical practices in the video-recorded treatment sessions. Because the observations revealed that the PTs primarily focused on biomedical aspects, it was clear that this knowledge did not motivate the PTs to include the psychosocial approach in their clinical practice.

The careful description of the data collection process and the illustration of the results with quotations and observation narratives increase the validity of the study. The authors were aware that three of them had a previous understanding of the BM approach, which could influence the analysis. The trustworthiness was strengthened by regular discussions among the four authors during the analyses, as well as through triangulation. Triangulating the data from the interviews with the PTs, patients, and managers with the observations from the video-recorded treatment sessions and the documents regarding local directives and regulations led to a comprehensive understanding of the determinants of using a BM approach in PT clinical setting. Several of the determinants were identified in more than one source and all determinants were replicated in most of the cases. Yin (2014) emphasized that triangulation of data and replication in multiple cases increase the robustness and trustworthiness of the results, which is an important strength of the current study.

The replication of the determinants indicates that these were the same across cases despite the differences between the cases in terms of county council affiliations, primary healthcare or hospital care, private or county council practice, and differences in PTs’ former BM education. This indicates a wider transferability to other PT contexts. The results can also be used as examples of determinants when implementing a BM approach in PT clinical practice.

Conclusion

Our study shows that traditional education alone is not sufficient to translate a BM approach into physiotherapy clinical practice. The complexity of integrating a BM approach into physiotherapy clinical practice arises from multiple determinants that could function as both facilitators and barriers. The results concretize relevant determinants at the micro- and meso-levels for integrating a BM approach into physiotherapy clinical practice. Determinants at the micro-level occurred most frequently regarding the individual PT and the patient but also the BM approach itself. Determinants at the meso-level were identified regarding the professional interactions and incentives and resources. However, no determinants were identified at the macro-level. Despite contextual differences between the cases, the identified determinants were replicated across the cases. To optimize the implementation of a BM approach in similar contexts, it is important to address the determinants found in the current study. In other contexts, these results can be used as examples of aspects to investigate when seeking to understand the determinants of the implementation in question. This study can also serve as a model for exploring determinants of the use of complex interventions in physiotherapy.

Declaration of interest

The authors report no declarations of interest.

Additional information

Funding

Funding was provided by AFA Insurance [12169].