Background
In a quasi-experimental trial, we used facilitation as the main method to support the implementation of a behavioral medicine (BM) approach in primary healthcare physiotherapy. A large effect size was found (
r = .72) regarding changes in the physiotherapists’ clinical behavior immediately after the implementation period, but the changes were not maintained at follow-ups [
1]. In contrast, self-efficacy in applying the BM approach increased and was maintained at follow-up. To increase understanding of the successful and unsuccessful parts of the intervention, the implementation process warranted further exploration.
Process evaluation has been suggested as an essential part of designing implementation interventions [
2,
3]. In the current study, the implementation intervention comprised the methods used to support the implementation of a BM approach. There is growing interest in the use of frameworks and models to make implementation efforts easier to plan and replicate and more likely to succeed by offering a structure and highlighting causal assumptions [
4]. The Medical Research Council has provided guidance on how to perform process evaluations of complex interventions [
5]. Process evaluations of the implementation of BM interventions in physiotherapy have focused on the fidelity [
6,
7] and feasibility [
8] of delivery. Process evaluations focusing on impact mechanisms (i.e., how the delivered intervention produces change [
5]) are sparse [
9‐
11] and are non-existent in physiotherapy. Thus, there is a need for process evaluations in the physiotherapy context.
A BM approach in physiotherapy is recommended in the treatment of patients with persistent musculoskeletal pain to increase their ability to participate in daily life activities [
12‐
14]. In this study, a BM approach implies a focus on health-related behavior changes in the assessment, analysis, and management of important biopsychosocial factors for behavior change. Behavior change techniques, such as the patient’s goal-setting, self-monitoring of behavior, and feedback on the patient’s behaviors, are important tools [
15,
16]. However, the implementation of a BM approach in a real-world setting is complex because of the multiplicity of clinical behaviors that must be adopted [
1,
17]. The implementation often results in some changes in knowledge and attitudes, but change in physiotherapists’ traditional biomedical approach is less common [
15,
17‐
22]. Forming new habits also requires considerable time, from 18 to 254 days (median 66 days), for the repetition of behaviors [
23]. Further research is needed on how to support the implementation of a BM approach.
Facilitation is a promising strategy to support the implementation of evidence-based guidelines in primary health care [
24]. Facilitation involves both the role of a person who facilitates and the process of practices to support the development of new knowledge and skills [
25,
26]. The current study applied basic assumptions for behavioral change in social cognitive theory in the facilitation intervention. Social cognitive theory emphasizes that a behavior is reciprocally influenced by personal and contextual factors [
27]. Self-regulation is the capability to control and manage these factors [
28]. Forethought capability (i.e., the capability for intentional actions), self-efficacy beliefs, self-monitoring, social support, and observational learning are important sources of self-regulation capability for behavior change [
27,
29,
30]. By addressing theory-based assumptions in the facilitation intervention, positive outcomes were expected in terms of the physiotherapists’ clinical behavior changes [
31].
To explain the findings in the quasi-experimental trial, a deeper understanding of the implementation process was necessary. Thus, the aim of this study was to explore the impact mechanisms in the implementation of a behavioral medicine approach in physiotherapy by examining dose, reach, and participant experiences.
Discussion
Because of high workload, the physiotherapists reported difficulties in prioritizing time for the implementation. Although the physiotherapists knew that they were allowed to allocate time for the implementation, they needed support for this. The outreach visits scheduled by the facilitator contributed to a structure that supported the allocation of time for implementation. A lack of time was mentioned in this and other studies as a barrier to implementation [
34,
37,
38] that hindered the repetition required to establish habits [
23]. The physiotherapists perceived that support from the manager in prioritizing their daily work was a prerequisite for implementation. All managers were encouraged to actively support the physiotherapists during the implementation, but only one physiotherapist perceived that this happened. Tistad et al. [
10] found that managers needed support to develop leadership behaviors in operationalizing the implementation plans. Aarons et al. [
39] noted that paying attention to implementation and allocating resources are important for managers’ facilitation of strategic climates for implementation. The challenge for many managers is to find the time to coach. Managers with a small number of employees (which was not the case in the clinics included in our study) appear to have better opportunities for coaching [
40]. When implementing new methods, time management is important. An external facilitator can contribute to this support during the implementation intervention period, but an engaged manager contributes to more sustainable support.
Role models provided by the video-recorded role-plays contributed to support for practice through observational learning. Most of the video-recorded role-plays concerned the same BM components as the areas in which a change occurred in the physiotherapists’ clinical behavior [
1]. If video recordings of the other components of the BM approach had been provided, increases in these clinical behaviors might have been found. Observational learning can be a shortcut when learning new behaviors [
27] and can be helpful in a time-pressed work situation.
Social influences such as peer coaching can increase physiotherapists’ capacity to initiate and maintain behavioral change [
27,
41]. Peer coaching contributed to support for practice through feedback, reflection and problem solving, and emotional support. These results are consistent with previous research suggesting that interventions focusing on action, experience, and peer support are more likely to lead to professional behavior change in health care [
11]. Learning new behaviors is linked to feelings of anxiety and frustration that require emotional support [
42]. Emotional support can also be important to overcome feelings of embarrassment when asking about psychosocial factors [
34]. However, there are barriers to making peer coaching work in reality. The physiotherapists had difficulty arranging peer-coaching situations and needed the facilitator to schedule a time for these situations.
The use of individual goals and behavioral contracts functioned as self-management support to structure the skills training as part of the forethought capability [
29]. The physiotherapists did not rate their own individual goal-setting as an important implementation method, which was contradicted by their experiences of goal-setting as a cue for skills training. The majority of the goals that were set during the implementation period corresponded to the same BM components for which a change in clinical behavior was found. It seems that these components were practiced to a larger extent than the components for which no change in clinical behavior was found. Locke and Latham [
43] claim that goal-setting affects motivation and persistence in achieving the goal. Given that the goal achievement in this study was only 59%, this phenomenon was not observed. The physiotherapists’ motivation may have been extrinsically driven by a willingness to please the facilitator. Behavior change is more likely to be maintained if the person perceives intrinsic motivation, such as satisfaction in performing the activity itself [
41,
44]. Most activities performed by physiotherapists in the clinic are not intrinsically motivated but rather are performed to achieve patient outcomes or to comply with guidelines and regulations. Nevertheless, self-rewards and self-control can contribute to a sense of competence and autonomy that is important for enhancing intrinsic motivation [
44]. Thus, the stimulation of self-reinforcement through satisfaction with goal achievement can be a successful method to increase intrinsic motivation for the maintenance of behavior change.
Different combinations of facilitation methods were preferred by different physiotherapists, revealing variation in their preferred ways of learning to acquire knowledge and skills. In higher education, multiple learning methods that integrate web-based and face-to-face learning activities have positive effects on students’ learning [
45]. An intervention including both practical tools and the ability to ask questions and receive feedback from a facilitator has a positive impact on learning outcomes [
18]. Thus, facilitation should be tailored to the physiotherapist’s personal preferences. In our study, the facilitation was tailored in relation to adaptations of the action plan, problem-solving, and reflections based on the physiotherapists’ needs. It is possible that an even more tailored intervention would have had a greater impact on the outcomes. The challenge is to balance adaptations of the implementation intervention for both the individual physiotherapist and the group of physiotherapists working at the same clinic.
Self-monitoring by video or diary was not widely used by the physiotherapists, thus excluding self-monitoring as support for the practice in our study. These self-monitoring methods aim to stimulate attention to one’s own performance as a self-diagnostic function prior to goal-setting and to stimulate the self-motivating function through reinforcement. Although self-monitoring is important for supporting behavior change [
29,
46], it was of little prominence as the physiotherapists did not use it. Thus, there is a need to identify other feasible methods for self-monitoring. It is possible that to be able to manage the barriers to using video recordings, stronger emotional support is required than was offered in this study.
The physiotherapists in our study asked for keywords in the patient records for psychosocial and behavioral factors that could function as reminders to use the BM approach. Computer reminders in electronic patient records have been used as memory support to prompt new behaviors [
47]. Small to modest improvements were found, but these improvements were larger when a response from the user was required to proceed. According to these results, computer reminders alone would probably not change the physiotherapists’ behavior. However, as part of a multifaceted implementation strategy, a computer reminder requiring an answer from the physiotherapist may contribute to forming habits and thus to the maintenance of the behavior change [
41].
The decision to initiate a behavior change depends on expectations regarding future favorable outcomes [
48]. The physiotherapists had high expectations regarding patient outcomes when using the BM approach, but the perceived importance of using each BM component was moderate [
1]. The core components “patients’ goal-setting,” “promoting patients’ self-monitoring,” and “functional behavior analysis” were perceived as less important. The physiotherapists’ attitudes towards the BM approach likely affected their intention to use it.
According to social cognitive theory [
27,
30], self-efficacy is a crucial determinant of the initiation and maintenance of behavior change. We previously reported increased self-efficacy for using the BM approach as an intermediate effect of the facilitation intervention [
1]. Although self-efficacy increased and was maintained, the behavioral changes were not maintained [
1]. According to Rusk et al. [
49], an intervention needs to address multiple domains, helping the system to “tip over” and change. In addition, multiple pathways for change contribute to maintenance through synergistic effects. Both the outreach visits and the individual goal-setting prompted skills training of the BM approach, which likely contributed to synergistic effects due to mastery experience and increased self-efficacy. However, when the external support for implementation ceased, the synergistic effects ended as well. The physiotherapists’ self-efficacy alone seemed insufficient to provide the tip-over effect for clinical behavior change to be maintained. It is therefore important to ensure that synergistic effects can continue after the implementation intervention.
The facilitation intervention addressed several behavioral change techniques [
35] described in Additional file
2. The results showed that the physiotherapists perceived most of these techniques as present in the intervention. However, the balance between the facilitation methods is worth considering to maintain clinical behavior change over time. To initiate clinical behavior change in the current study, a variation of externally initiated facilitation methods seemed important. According to Clark and Zimmerman [
28], external support should gradually decrease as the self-regulation capability increases. The balance between external support for clinical behavior change and support to increase self-regulation capability is an important factor to consider in future studies.
Well-known theoretical approaches to behavior change [
27,
46] do not formally distinguish between how to initiate and how to maintain behavior change. However, our results indicate that different processes guide the initiation and maintenance of behavior change. Theoretical explanations for the maintenance of behavior change focus on motives, self-regulation, habits, resources, and environmental and social influences [
41], which correspond to the physiotherapists’ experiences in our study. Therefore, future process evaluation studies should include important factors for maintenance.
Strengths and limitations
A particular strength of our study lies in the theoretical base. The Medical Research Council guidance for process evaluation [
5] provided a structure for exploring the implementation process, and social cognitive theory [
27,
29,
30] guided the understanding of the process evaluation findings. This study also concretized and discussed how the components of the social cognitive theory were addressed, which strengthens the transparency of the study. To the best of our knowledge, this is the first study to use social cognitive theory to seek explanations for the underlying processes that make implementation interventions effective. The Behaviour Change Taxonomy [
35] was used to clarify and provide further transparency regarding which behavior change techniques were used in the implementation intervention.
Qualitative data about the physiotherapists’ experiences of the facilitation methods complemented the quantitative data about the dose and reach of each method. This information enhanced the understanding of the successful and unsuccessful parts of the implementation intervention. However, the mapping to the Behavior Change Taxonomy [
35] was somewhat problematic. There is some overlap between behavior change techniques. Some techniques are described as processes (e.g., feedback on behavior, prompt/cues) and others as strategies to obtain these processes (e.g., social support, goal-setting). To address these overlaps, the physiotherapists’ experiences were categorized in relation to the behavioral change technique that they primarily addressed. The qualitative analysis was continuously discussed and confirmed among the researchers in the study to strengthen the trustworthiness. Quotes from participating physiotherapists are presented to add transparency and trustworthiness to the findings [
50].
The characteristics of the sample in this study are likely similar to physiotherapists in primary health care in Sweden, although descriptive studies of primary health care physiotherapy in Sweden or other countries are sparse [
51]. The sample in the current study had a wide span in age and work experience and represented both cities and smaller towns. The self-selecting nature of the sample could imply that these physiotherapists were more motivated towards behavior change. However, the physiotherapists’ moderate expectations of changing their clinical behavior by participating in the study suggests the opposite. The contextual factors, such as a high workload and lack of time, are probably valid for most physiotherapists in primary health care. The characteristics and contextual factors taken together thus support the transferability of the findings to physiotherapy in primary health care.
Our results indicate that different processes guide the initiation and maintenance of behavior change [
52]. In this study, data were only collected during and immediately after the implementation intervention period. We recommend that future studies extend the duration of the process evaluation to focus on mechanisms for the maintenance of changed behavior.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.