Conservative non-pharmacological management of knee osteoarthritis in Switzerland: a survey among medical specialists

Background: International Guidelines recommend self-management, exercise and weight management if appropriate as rst line conservative treatment for patients with knee osteoarthritis (OA). The objective of this study was to survey the conservative non-pharmacological management of patients with knee OA in Switzerland and to explore perceived barriers and facilitators for the application of guideline recommendations. Methods: Based on the results of semi-structured interviews with selected general practitioners (GPs), rheumatologists and orthopaedic surgeons, a survey was performed across their scientic societies. Questions addressed diagnostic measures, treatment options, reasons for referral to exercise as well as barriers and facilitators. Results: In total, 234 members responded. They indicated that usually patients presented due to pain (n=222, 98.2%) and functional limitations of the knee (n=151, 66.8%). Additionally to clinical assessment X-ray (n=214, 95.5%) and MRI (n=70, 31.3%) were used as diagnostic measures. The referral to exercise, was driven by the patients’ expectation/high level of suffering (n=73, 37.1%), as well as by their own clinical experience (n=49, 24.9%). They estimated to refer 54% of their patients to exercise. Further, the specialists rated as the most important barriers for the referral to exercise ‘disinterest of patient’ (n=88, 46.3%) and ‘physically active patient’ (n=59, 31.1%). As most important facilitators, they rated ‘priority to mention exercise during short time of consultation’ (n=170, 89.4%) and ‘insuciently physically active patient’ (n=165, 86.9%). Conclusion: A substantial evidence-performance gap in the management of patients with knee OA seems to be present. To successfully support systematic referral to exercise as rst line intervention, it may be important to not simply suggest exercise in general, but a specic best-practice exercise and education programme for knee OA.

The online survey was developed based on literature and semi-structured interviews with GPs, rheumatologists and orthopaedic surgeons selected with regard to gender, region and language area. Eleven telephone interviews, lasting between 5 and 15 minutes, were conducted. The interview questions were related to the domains 1) consultation reasons of patients and diagnostic measures additional to the clinical assessment; 2) conservative treatment options; 3) reasons for referral to exercise, and 4) barriers and facilitators for referral to exercise. The analysis was performed using a directed content analysis (23), i.e. with regard to the recommendations of clinical guidelines, and the results informed the survey questions. The nal version of the survey contained 14 questions within the above mentioned four domains: 1) consultation reasons and diagnostic measures, two questions on main symptoms presented by patients and the use of diagnostic measures in addition to clinical assessment; 2) treatment options in conservative knee OA treatment, assessed on a Likert scale from 3 (always) to 0 (never); 3) reasons for referral to exercise, where participants were asked to prioritize ve answer options from 1 to 5; and 4) barriers and facilitators for referral to exercise, as recommended by the guidelines, assessed on a Likert scale from 5 (I fully agree) to 1 (I disagree). Additionally, the rate of referring their patients (%) to exercise and indication criteria for referring to surgery were inquired. Six questions on the survey participants' characteristics and their work with knee OA patients were asked initially. After pilot testing with three specialists and three researchers, the survey was slightly modi ed to improve its clarity. The survey was kept as short as possible, i.e. less than 10 minutes for completion, taking into account the limited time of physicians busy with clinical work (24).

Recruitment of survey participants and data collection
GPs, rheumatologists and orthopaedic surgeons in the German, French and Italian language areas in Switzerland were invited to participate in the online survey by their societies, i.e. the Swiss family physicians and paediatricians association (mfe, 4764 members), the Swiss Society of Rheumatology (SGR, 570 members) and the Swiss society of orthopaedics and traumatology (swiss orthopaedics, 759 members). The surveys were sent as links to access a German, French or English version, using Survey Monkey®. The SGR sent a reminder to their members after three weeks.

Data analysis
Demographics and work characteristics of survey participants are presented as frequencies and percentages or means and standard deviations where appropriate. To test the signi cance of differences between the subgroups, variance analysis (one-factor ANOVA) was applied where appropriate and the Likert scales were dichotomized for group comparison, i.e. the answer options always/often and the answer options seldom/never, and additional the answer option "I don't know".
Descriptive statistics were performed using the SPSS software, version 25 (SPSS, Chicago, IL.).

Characteristics of participants
A total of 5980 specialists of the three medical societies were invited to participate in the survey and 234 (3.91%) responded. Responders were included for analysis if they answered, additional to the demographic questions, at least the three questions concerning consultation reasons, diagnostic measures and treatment options in conservative knee OA management. This resulted in 226 specialists for analysis, among them 72 GPs, 84 rheumatologists and 70 orthopaedic surgeons. Their characteristics are reported in Table 1. Items are in a ranked order only in the analysis regarding 'All participants'.

Rheum. Rheumatologists
Orthop. Orthopaedic surgeons The use of multiple additional diagnostic measures when clinical signs indicated knee OA are displayed in Fig. 1. Irrespective of the medical discipline, X-ray (n = 214; 95.5%) was the most used diagnostic measure to con rm the clinical diagnosis. MRI was substantially less used (n = 70; 31.3%), and of these, mainly by orthopaedic surgeons (n = 40; 57.1%).

Conservative treatment options
The conservative treatment options most mentioned by all specialists were 'informing the patients about the diagnosis' (n = 223; 98.6%), 'recommending suitable activities or sports' (n = 217; 96%), 'pharmacological treatment' (n = 203; 89.8%) and 'referral to physiotherapy' (n = 188; 83.2%). There were no differences among the subgroups in all these treatment options (p = 0.056). The only difference across the subgroups was in 'referral to other medical health specialist', what 26.6% of the GPs 'always' or 'often' did, compared to 6% of the rheumatologists and 10% of the orthopaedic surgeons (p = 0.000) ( Table 3). Notes: Values are absolute and relative frequencies. Items are in a ranked order only in the analysis regarding 'All participants'.

Rheum. Rheumatologists. Orthop. Orthopaedic surgeons
Reasons for referral to exercise A total of 226 specialists estimated to refer 53.95% (SD 27.80) of their patients to exercise. The subgroup analysis revealed no signi cant differences in the estimated referral to exercise (p = 0.058). Figure 2 displays the prioritization of the reasons for referring patients to exercise. The subgroup analysis only showed signi cant (p = 0.008) differences in prioritizing the reason "degree of osteoarthritis".
Barriers and facilitators for referral to exercise Barriers and facilitators for referring patients with knee OA to exercise as recommended by the guidelines are displayed in Table 4. and 'applicability of the guidelines to suggest exercise' respectively. Most rated facilitators were 'priority to mention exercise in the short time of a consultation (n = 170; 89.4%) and 'insu ciently physically active patient' (n = 165; 86.9%). Furthermore, the 'guideline recommendations' (n = 121; 63.7%) and 'anticipated/perceived interest of patients' (n = 146; 76.9%) were stated as a facilitating factor to suggest exercise to the patients.
Indication criteria for referral to surgery (n = 226) The most mentioned criteria for referral to surgery were 'high level of pain and suffering' (n = 142; 62.8%), 'exhaustion of conservative treatment strategies' (n = 106; 46.9%) and 'limitation of functioning in ADL' (n = 70; 31.0%). There was often a combination of two or more criteria for referral to surgery. The subgroup analysis showed differences in the main criteria for the referral to surgery, i.e. 'high level of pain and suffering' was the main criteria for the GPs (65.3%) and rheumatologists (67.9%), whereas 'exhaustion of the conservative treatment strategies' was the main criteria for the orthopaedic surgeons (57.1%).

Discussion
To our knowledge, this is the rst study that surveys the conservative non-pharmacological management of patients with knee OA in Switzerland. The main nding was that the international clinical guidelines for the management of knee OA were not systematically applied in Switzerland. The (non-) adherence to the guidelines was related to the diagnosis as well as the referral to exercise as recommended in the guidelines of knee OA. Some important barriers to and facilitators for the use of the guidelines have been detected and evaluated. As most important barriers were rated patients who were either disinterested or already physically active, whilst as most important facilitators were rated the importance of the topic exercise and patients who were insu ciently physically active.
The surveyed GPs, rheumatologists and orthopaedic surgeons reported that they most used X-ray and MRI in addition to their clinical assessment. Guideline recommendations suggest that a careful clinical examination is su cient, unless there is any additional bene t to imaging patients as part of the diagnostic pathway or to con rm a differential diagnosis (11,25,26). The fact that orthopaedic surgeons showed a substantially higher use of MRI could be due to referrals from other specialists, thus of possibly more severe cases with knee OA to evaluate the surgical option.
There is a gap between the specialists' ratings for the treatment options, especially regarding 'referral to physiotherapy', and their estimated rate of referrals to exercise. More than 80% of the specialists chose 'referral to physiotherapy' as a treatment option, whereas the estimated rate of referrals to exercise across the subgroups was around 54%. In all subgroups, the patients' expectations or level of suffering as well as their own experience and the clinical picture, drove the decision-making for referral to exercise. Interestingly, even though the orthopaedic surgeons are the last specialists when a surgery is indicated, they did not show a lower rate of referral to exercise. They also prioritized "degree of OA" as reason for referral to exercise higher than GPs or rheumatologists. It can be assumed that orthopaedic surgeons may refer patients with lower degrees of OA to exercise and patients with higher degree of OA to surgery.
There seemed to be no systematic use of the guidelines among all specialists and therefore no systematic suggestion or referral to exercise as rst line intervention. Interestingly knowledge and adherence to guidelines was comparable between the GPs and the specialists for the musculoskeletal system, i.e.
rheumatologists and orthopaedic surgeons, even though GPs are more often challenged with multimorbid patients where guideline recommendations are often not systematically applicable. The only signi cant difference between the GPs and the specialists for the musculoskeletal system was in the referral pattern. GPs are usually primarily consulted and may refer the patients to rheumatologists and orthopaedic surgeons, who in turn see more referred patients than by direct access.
The evidence for the effectiveness of exercise in people with knee OA to reduce pain, improve physical function and quality of life in short-and long-term has been con rmed over and over in meta-analysis (5,9,19,27). Already in 2015, the Cochrane Collaboration stated that the evidence for the effects of exercise were so convincing that further studies were unlikely to change this strong and high-quality evidence (9). Previous studies showed that suboptimal use of exercise could be due to patients' preferences or lack of information about conservative treatment options (28,29). The surveyed specialists have an important impact by their own attitude and how they communicate possible treatment options. They should therefore be aware that they enhance the patients' motivation towards exercise by explaining them the positive outcomes of exercise, and support the shared decision-making towards exercise (9,28,29).
Overall, we conclude that there is a substantial gap between the guideline recommendations and clinical practice in the management of knee OA in Switzerland. To facilitate the guideline application and referral to exercise, it is important to translate the recommendations into a best-practice exercise programme that is of high quality and applicable in clinical practice and provide easy guidance for patients and health care providers alike. There are structured exercise and education programmes for knee OA that have been successfully established across the world, i.e. "Osteoarthritis Chronic Care Program (OACCP) Australia", "Better management of patients with osteoarthritis (BOA) Sweden", "Good Life with osteoarthritis in Denmark (GLA:D)", "Osteoarthritis Healthy Weight For Life (OA HWFL) Australia", "Amsterdam osteoarthritis cohort (AMSOA) The Netherlands" or "Joint Implementation of Osteoarthritis guidelines in the West Midlands (JIGSAW) UK" (12). All those programmes translate the guideline recommendation into practice with the goal to enhance selfmanagement and are endorsed by OARSI. The programmes deliver the rst line treatment exercise and education with different degrees of intensity and standardization. Some programmes include weight management support. The programmes have many similar contents, but differences in targeted groups of patients and health care professionals. Most programmes assess pain, function and quality of life. Such a structured exercise and education programme Figure 1 Diagnostic measures used after clinical assessment (n=226) Figure 2 Highest prioritization of reasons for the referral to exercise (n=1)