The qualitative approach was chosen due to its suitability for exploring contextual and implementation-related factors that cannot be fully captured through quantitative methods. We conducted semistructured interviews with experts in the field of treatment programs for knee and hip OA as well as back pain.

To identify relevant programs, we first carried out a rapid review focusing on conservative treatment programs for knee and hip OA [7]. Based on this rapid review, we assembled a list of existing approaches and consulted international experts to gain insights into national characteristics and contextual influences.

In addition to OA programs, the review identified one back pain program that had been developed through adaptation of an established OA program. We included this program because it represented a unique example of how implementation strategies and program structures can be transferred across musculoskeletal conditions. Its inclusion enabled exploration of implementation processes beyond a single diagnosis while maintaining a focus on comparable nonsurgical, exercise-based care models.

The selected interventions additionally provided insights into how treatment concepts have been adapted across countries and how care models differ between analog, digital, and blended care [9].

The included programs differed in their form of delivery, implementation strategies, and national contexts. Better Management of Patients with OsteoArthritis served as the foundation for the subsequent development of GLA:D (Good Life with osteoArthritis in Denmark) [10,11], while Joint Academy emerged as a digital adaptation designed to improve accessibility and overcome barriers to participation in conventional OA care [12]. Studies accompanying the GLA:D program showed a significant reduction in the risk of being unable to work due to knee or hip OA, as well as improvements in pain intensity, physical resilience, and everyday activity [6,11].

Although differing in form of delivery and organizational structure, the programs shared a common evidence-based treatment philosophy.

JB and TSB designed the recruitment strategy to target participants with specific expertise in implementing OA treatment programs for knee and hip. Eligible participants were professionals who had been directly involved in the development, coordination, implementation, or scientific evaluation of one of the identified programs. The interviewers (JB and TSB) had no prior professional relationship with the participants before recruitment except for one participant, with whom they had had occasional contact during a joint project in the past. This did not affect the interview situation.

The participants were informed that both researchers are currently working on the facilitating and inhibiting factors in implementation research and about the insights they had gained from the rapid review mentioned above.

In the beginning, JB and TSB compiled a list of all OA treatment programs for knee and hip that had been identified worldwide in the previously mentioned review. JB and TSB then identified and contacted the program coordinators responsible for the development and implementation of these programs. In addition, they expanded recruitment through expert recommendations provided by interview participants (snowball sampling). Recruitment continued until additional interviews were considered unlikely to provide substantially new insights regarding the implementation processes under investigation.

The research team invited all identified candidates via email to participate in the study. Candidates were asked to complete a short questionnaire regarding sociodemographic information.

In total, the research team contacted 30 experts and one additional contact email address for a specific treatment program. Twelve experts agreed to participate in the study. Recruitment and interviews were conducted between August and October 2023. The final sample was considered appropriate because it captured experts from multiple countries, health care settings, and implementation roles, providing sufficient variation to address the study aim.

Interviews were conducted in English or German, depending on the interviewee’s language preference. Accordingly, we developed the interview guides in both languages and ensured full content equivalence across both versions.

The interview guides were reviewed by an orthopedic surgeon for technical adequacy and by a team of research assistants from the Institute of General Practice and Family Medicine at Ruhr University Bochum during a qualitative research workshop. This review process focused on content relevance, comprehensibility, and alignment with the study objectives. Revisions were made iteratively based on the feedback received and discussed within the group of authors.

The semistructured interview guides (see Multimedia Appendix 1) focused on 4 key thematic areas:

Interviews were conducted by 2 female researchers (TSB: doctoral degree in health sciences; JB: MSc in health economics), both trained in qualitative research and employed at the Institute of General Practice and Family Medicine at Ruhr University Bochum at the time of the study. In most cases, both researchers attended the interviews—one leading the interview and the other monitoring the process and taking notes. TSB and JB took turns in this role to ensure consistent quality.

Neither interviewer was involved in the development or implementation of any of the included OA treatment programs. This helped to reduce potential conflicts of interest and enabled participants to discuss implementation experiences from an external perspective. Throughout data collection, the interviewers documented contextual observations and reflected on potential influences on the interview process in field notes.

All interviews were conducted via videoconferencing (using Zoom by Zoom Communications), audio-recorded, pseudonymized, and transcribed by a professional transcription service. German interviews were transcribed verbatim in German, just as English interviews were transcribed verbatim in English. The interviews were always analyzed in the original language. German quotations selected for publication were translated into English by the research team, and the translations were checked by 2 different authors.

In one case, the formal interview was preceded by a detailed preliminary conversation, which the participant had requested before agreeing to the interview. During this conversation, the participant described his role in the program. The participant consented to audio recording only after this preliminary exchange and had limited time available. To avoid redundancy and respect these constraints, the thematic first block of the interview guide was omitted during the recording. This interview was still included in the analysis, as 3 researchers (JB, JN, and TSB) independently agreed that the data provided sufficient contextual richness. The preliminary conversation was not included as primary interview data.

We analyzed the interview transcripts using qualitative content analysis following the approach described by Kuckartz [13]. Interviews were transcribed shortly after completion of data collection and prepared for analysis. The formal coding and category development were conducted between September and November 2024 using MAXQDA software by VERBI GmbH (TSB: MAXQDA 2024; JN: MAXQDA 2022). The time lapse between data collection and analysis is attributable to the authors’ new professional positions and the subsequent conclusion of the project, which prevented them from working on the project earlier, and is not due to changes in the dataset. Coding was performed in English. Two researchers (TSB and JN), both holding doctoral degrees in health care science and trained in qualitative research, independently coded all interviews. Discrepancies were discussed during regular analytical meetings. Where differences in interpretation occurred, both researchers revisited the original transcript segments and agreed on a final coding decision through discussion and consensus.

We applied a structured content-analytic approach combining deductive and inductive category development. Initial main categories were derived deductively from the interview guide, while additional categories were generated inductively from the data. Throughout the analysis, the coding framework was iteratively refined and continuously compared across interviews to ensure analytical consistency.

To enhance credibility, a third researcher (JB) with experience in qualitative content analysis reviewed the final category system and code tree for plausibility, coherence, and completeness. Feedback from this review informed the final interpretation and refinement of the category system, which was finalized in March 2025.

Methodological rigor was supported through independent coding, regular consensus meetings, and review of the final category system by a third researcher. Field notes were used to contextualize interview data and support interpretation.

The study received a favorable ethics vote from the Ethics Committee of Ruhr University Bochum (register number 21-7438) in February 2022.

Interested individuals received detailed study information outlining the study’s objectives, scope, and participation requirements. To formally participate, candidates were required to sign a declaration of consent. Participants returned the signed consent forms via email.

Participation in the interviews was voluntary. Participants could at any time—even without giving a reason—decline to participate or withdraw their consent and leave the study without suffering any adverse consequences as a result.

Participants did not receive compensation for the interviews.

The recruitment process resulted in 12 completed interviews with experts involved in 8 different treatment programs. The interviews ranged from 33.42 to 65.02 minutes, with an average duration of 49.5 (SD 9.42) minutes.

The majority of experts (8 of 12) were affiliated with the GLA:D network, making it the most represented program in the sample. Specifically, experts participated from GLA:D country programs in Germany (n=2), New Zealand (n=2), Switzerland (n=2), Canada (n=1), and Denmark (n=1). Another expert from the Danish GLA:D Back program (n=1) took part in an interview. In addition, experts from the Joint Academy Program (n=2) and the Better Management of Patients with OsteoArthritis program in Sweden (n=1) were interviewed.

Figure 1 provides an overview of the programs in which the experts were involved, detailing their structural characteristics and the national context in which the interviewed experts had worked.

The age of participants as of 2024 ranged from 35 to 81 years, with an average age of 53.4 years (SD: 13.35). Three experts identified as female and 9 as male. To protect the anonymity of the participants, no further sociodemographic data are reported in this publication.

During rollout, responsibilities expanded to include trained therapists (chiropractors, physiotherapists, physicians, and nurses), scientists, software developers, and regional coordinators. Interviewees emphasized the importance of creating local ownership among clinicians, who frequently acted as ambassadors and facilitators within their professional networks:

Rollouts varied from regional pilots scaling nationally to single-insurer collaborations or multiregion launches.

Initial funding during the rollout phase came from associations, grants, or public sources, though high license fees and missing reimbursements hindered progress.

Despite perceived successes, experts identified several ongoing challenges, including insufficient reimbursement structures, continued reliance on surgical care pathways, limited implementation of evidence-based recommendations, and administrative burdens. To address these issues, interviewees emphasized the need for stronger political support, expanded digital services, improved professional education, and sustainable financing models.

Future priorities included increasing public visibility, strengthening quality assurance mechanisms, expanding programs to new settings and populations, and generating further evidence through registry and health services research.

This study provides in-depth insights into international implementation processes for evidence-based care programs for musculoskeletal disorders, with a primary focus on knee and hip OA and 1 related back pain program derived from an OA model. The findings were obtained from 12 expert interviews. Beyond documenting persistent challenges across preparation, initial implementation, and consolidation phases, our findings reveal a critical disconnect: while international networks and evidence-based models like GLA:D are widely valued, systematic knowledge exchange and joint solutions remain rare. In some instances, international cooperation even required additional coordination and administrative effort. Nonetheless, the possibility of adapting and translating existing program materials, rather than developing entirely new ones, was seen as a major advantage. Westerlund et al [16] highlight that implementation science encompasses a growing range of frameworks, models, and theories aimed at bridging the gap between knowledge generation and its application in health care practice [9]. Evidence suggests differing perceptions between researchers (knowledge producers) and policymakers (knowledge users) [8]. Knowledge producers tended to underestimate their influence on policy while overemphasizing barriers to implementation [8]. While opinions differ on whether this gap can be bridged by theoretical approaches in implementation science, it can be assumed that the application of theory is explicit and verifiable but might lead to problems being overlooked that are not addressed in theory. Applying common sense for implementations is often based on implicit assumptions and beliefs [13]. Nilsen underlines that implementation needs to be connected to “the real world of implementation practice” [13].

This study represents an important exploratory step toward understanding the implementation of evidence-based OA treatment programs in different international contexts. However, some limitations must be considered.

By focusing on experts, we were able to extract subject-specific expertise regarding the practical implementation of OA programs. On the other hand, the exclusive reliance on expert interviews restricts the range of perspectives, particularly as patient views are therefore not included.

Regarding the selection of interviewees, the overrepresentation of experts affiliated with GLA:D-related programs may have influenced the thematic emphasis of the findings. In addition, the international sample was recruited through purposive and snowball sampling, which may have favored well-established professional networks and highly visible programs while excluding less prominent initiatives and perspectives.

Another limitation relates to the fact that most participants were directly involved in the development, implementation, or management of the programs under investigation. Their close engagement may have influenced the way implementation successes and challenges were described, potentially resulting in a more favorable assessment of the programs.

Although the interview guide covered the main aspects of implementation, it was not explicitly developed using a comprehensive implementation framework. Consequently, some contextual aspects, such as regional or federal policy influences, may not have been explored in equal depth.

By focusing exclusively on qualitative interviews, the study did not include methodological triangulation (eg, additional quantitative data). Consequently, the results could not be supported by multiple types of evidence.

Similarly, the findings were not externally validated through participants or additional stakeholder validation in the sense of member checking. While participant validation is not considered essential in all qualitative research traditions, its absence may have limited the opportunities to further enhance the confirmability of the findings.

Although purposive sampling allowed us to maximize variation in implementation experiences across countries and program types, transferability of the findings depends on similarities between health care systems, reimbursement structures, regulatory environments, and organizational contexts. The findings should therefore be interpreted in relation to the specific health care contexts in which the included programs were implemented.

Future research should address specific gaps identified in this study, including the cost-effectiveness of implementation strategies, the contextual adaptation of programs to diverse health system contexts, and implementation in resource-limited settings.

This qualitative study explored facilitators and barriers influencing the international implementation of evidence-based nonsurgical OA programs from the perspective of experts (n=12) involved in established international programs.

Participants emphasized the importance of international collaboration, established program structures, and access to implementation experience. They highlighted the importance of stakeholder engagement, effective communication, and interdisciplinary collaboration throughout the implementation process. Furthermore, participants emphasized that context-appropriate combinations of digital, analog, and blended care models may improve accessibility, scalability, and long-term sustainability of OA programs. Fragmented reimbursement structures and short-term funding were consistently described as key barriers to sustainable implementation.

These insights provide practical guidance for researchers, clinicians, and policymakers seeking to implement evidence-based OA care programs across diverse health care systems while acknowledging the importance of adapting implementation strategies to local contexts.