Agreement of general practitioners with the guideline-based stepped-care strategy for patients with osteoarthritis of the hip or knee: a cross-sectional study

Osteoarthritis (OA) of the hip or knee is a common joint disorder causing pain and functioning impairment. The core treatment for OA, a combination of pharmacological and non-pharmacological treatment modalities, is mainly performed in primary care. In the Netherlands, once the diagnosis had been established, it has been found that patients with OA are treated in general practice for approximately 82 months before they are referred to an orthopaedic department [1]. Although several clinical practice guidelines (CPGs) to manage hip or knee OA exist [2‐6], diagnostic procedures, referrals, and use of treatment modalities observed in primary care tend to be inadequate [7‐9]. In addition, a recent review depicts a notably low pass rate for quality indicators for OA care (interquartile range 29-41%) [10]. These findings support the conclusion that incorporation of CPGs into clinical practice is not that simple.

We have developed a multidisciplinary, guideline-based stepped-care strategy, known as BART, i.e. Beating Osteoarthritis, to improve the management of hip or knee OA [11]. Experts from each discipline involved in OA care, as well as representatives from patient organizations and professional associations, participated in the development. This stepped-care strategy (SCS) provides a framework for various health care providers and patients to manage hip or knee OA (see Table 1). In addition to the current CPGs that give recommendations about the appropriate non-surgical treatment modalities, the SCS focuses also on the optimal order in which to employ them. This sequence is presented in three steps. At each step, recommendations for diagnostic procedures, treatment modalities, and the length of treatment before evaluation are made. Consequently, more advanced options are only recommended if the options listed in previous steps have failed to produce satisfactory results. Despite a possible risk of delay for more advanced modalities, the SCS is a suitable approach for patients with hip or knee OA.

To implement this strategy, the views about the SCS recommendations and working procedures concerning OA care of general practitioners (GPs) need to be assessed and the barriers that prevent GPs from using the SCS need to be identified [12‐14]. This knowledge can then be used to develop implementation activities tailored for GPs. The importance of such activities, created in response to identified barriers, has been demonstrated in a recent review that reported that tailored activities are more likely to improve professional practice than non-tailored activities [15]. Also, insight into the GPs’ views provides information concerning the strengths and weaknesses of current practice and heightens the perception of the need for change [13].

An approach based on a theoretical model can help to identify these barriers. Barriers are frequently grouped into social factors (including patients’ preferences), the clinicians’ attitudes, the implementation process, accessibility and format of the program, and external barriers of a practical nature [12, 16‐18]. In this study, we assessed the attitudes towards the SCS recommendations expressed by GPs which could be a potential barrier to successful implementation of this strategy [19]: specifically, their attitudes regarding the effectiveness of recommended non-surgical treatment modalities for hip or knee OA and their degree of agreement with recommendations regarding the sequence for care. Considering the fact that each recommendation can be influenced by different factors [20], we will assess the attitudes towards several specific SCS recommendations.

The aim of this study is to describe the GPs’ attitudes with respect to the two key elements of the SCS: 1) their attitudes regarding the effectiveness of the recommended non-surgical treatment modalities and 2) their agreement with specific recommendations regarding the sequence for care. In particular, those factors, that could influence their agreement with these recommendations, will be identified and will be used to develop tailored made implementation activities for this target group.

Out of the 1230 approached GPs, 456 GPs (37%) responded to the survey. Differences between the main characteristics of the participating GPs and the total population of Dutch GPs [21] were limited to the number of working hours and location of practice (Table 2).

In general, the GPs’ attitudes were in concordance with SCS recommendations regarding the non-surgical treatment modalities and the sequence for care; seven of the eleven recommended treatment modalities were considered effective for patients with hip or knee OA by the great majority of the GPs and five of the seven SCS recommendations regarding the sequence for care were consistent with the attitudes of most GPs. However, we found a notably high number of GPs who reported that tramadol, topical NSAIDs, and glucosamine were not effective and who reported that non-recommended modalities were effective in patients with hip or knee OA. Also, many GPs considered an X-ray necessary to diagnose OA and considered NSAIDs the drug of first choice (instead of acetaminophen). GPs’ agreement with SCS recommendations regarding the sequence for care was only weakly associated with a positive attitude towards the effectiveness of recommended modalities, a negative attitude towards the effectiveness of non-recommended modalities, and working in a duo or group practice.

As mentioned above, many GPs reported thought that several types of pain medication that are recommended in the SCS are not effective in patients with hip or knee OA. An explanation for the negative attitude towards tramadol could be the high prevalence of side effects [22, 23], which may result to therapy switching or discontinuation, and thus a negative evaluation of its effectiveness by GPs [24, 25]. Furthermore, the GPs’ negative attitudes towards the effectiveness of topical NSAIDs, glucosamine, and hyaluronic acid intra-articular injections might be explained by the discrepancies between the recommendations of the SCS and the Dutch NHG-standard for non-traumatic knee complaints in adults [6]. As the SCS was developed through a consensus procedure based on several national and international guidelines, the SCS recommendations differed slightly from those of the Dutch NHG-guideline. For example, the NHG-standard recommends GPs not to use glucosamine in patients with OA and gives no recommendations regarding topical NSAIDs, while the SCS does. In general, NHG-standards have a great impact on GPs’ knowledge [20] and thus could have influenced GPs’ attitudes regarding the effectiveness of these modalities. Another possible explanation for the GPs’ negative attitude towards topical NSAIDs, glucosamine, and hyaluronic acid intra-articular injections is that these are not reimbursed in the Netherlands. The GPs’ positive attitude regarding the non-recommend modalities, like manual therapy and other passive modalities, could be explained by the preferences for these modalities by the patients and physical therapists. This explanation is supported by the fact that these modalities are frequently used in patients with OA [26].

The SCS points out that medical history and physical examination are sufficient to diagnose (symptomatic) hip or knee OA, as radiographic confirmation of OA has little impact on the management, particularly in the early stages of the disease [3, 5, 27]. Interestingly, many GPs reported that X-ray is necessary to diagnose hip or knee OA. This finding is concordant with other studies that assessed GPs’ behaviour for ordering X-rays in the management of OA or back pain [28‐30]. The GPs legitimise their use of radiographs by expressing the view that it aids the discussion of management with the OA patient, is required for specialist referral, and can be used to reduce referrals [29]. Moreover, GPs believe that X-rays provide reassurance to patients which can outweigh the risks; furthermore, denying X-rays could adversely affect the doctor-patient relationship [30]. Although patients who have had X-ray seemed to be more satisfied, they reported more pain, lower overall health status, though no difference in disability, and consult their doctor more frequently [30, 31]. In light of this, GPs need to be informed about the limited value of X-rays in early OA.

We found three factors that are associated with the GPs’ attitude regarding the sequence for care. The first two factors, that concern the GPs’ agreement with the effectiveness of recommended and non-recommended treatment modalities, suggest that GPs who are aware of the effectiveness of these modalities agree with the SCS recommendations regarding the sequence for care. This result might be explained by the fact that these recommendations were based on evidence outlined in CPGs. Furthermore, the association between the GPs’ attitude regarding the sequence for care and the type of practice is in concordance with other studies suggesting that the organizational setting of the practice is the most consistent predictor of the GPs’ behaviour and can influence GPs’ performance [32, 33]. Generally, the isolated physician in solo practice provides a more limited range of services and show lower levels of clinical competence [33]. Moreover, GPs in solo practices appear to have a more aggressive treatment style than those physicians in group practices, which might be explained by financial incentives, lack of peer influences and availability of colleagues for informal consultation [32].

This study is not without limitations. First of all, only 10% of the variance in GPs’ agreement could be explained by factors related to the GP, the practice, or the management organization for OA patients. We did not examine the contribution of person-related or situational factors, e.g., the GPs’ experience, the patients’ preferences, local infrastructures, and rules or laws on the sequence for care; these factors have been named in literature as potentially able to influence GP’s attitudes [12, 16]. Secondly, our study does not cover all professionals because the research aim of this study was restricted to GP’s views and working procedures, while implementation of the multidisciplinary SCS should involve different disciplines. However, this study is part of an umbrella project, the BART-project, which aims to implement the SCS in practice and evaluate the implementation process in one region of the Netherlands, in preparation for the nation-wide implementation. Therefore, the views and working procedures of patients and other health care providers will be studied and described at a later date. Thirdly, the self-designed survey was tailored to our target population and not validated. Fourthly, we measured the GP’s attitudes regarding recommendations of the SCS and not the GP’s actual behaviour. Although a positive relation between attitude and behaviour can be assumed [34] our results do not give insight into the extent that the current clinical practice is concordant with the recommendations of the SCS. Consequently, there might be other barriers that impede a successful implementation. Finally, the response rate (37%), although relatively high for these kind of surveys among GPs, can raise some concerns regarding the validity and generalizability of our findings. Although we did not find large differences in several demographic and practice-related characteristics between the responders and the total population of Dutch GPs, we are aware that non-response bias could have affected the results. It has been stated that “serial” non-responders to GP surveys tent to be older, less likely to possess a postgraduate medical qualification, or belong to a practice that is involved with postgraduate or undergraduate training. [35] As we expect that additional education is associated with a more positive attitude to evidence-based practice, it could be hypothesized that our findings are an overestimation of the degree of agreement with the SCS recommendations.

Given the above-mentioned findings, the GPs’ attitude regarding recommendations in the SCS is not an insurmountable barrier for implementing the SCS in general practice. GPs are supportive of the recommendations regarding the effectiveness of treatment modalities and the sequence for care. Potential targets for implementation are improving the GPs’ knowledge regarding the effectiveness and optimal sequence for diagnostic procedures and treatment modalities, particularly in GPs who are working in a solo practice. Therefore, we recommend to include these themes in the GP-guideline and embed these in the program of the (post-graduate) training program and/or post-academic training for GPs. We did not identify any barriers that substantially contribute to GPs’ agreement with the SCS recommendations regarding the sequence for care. Further efforts should be taken to identify barriers that could prevent GPs from using the SCS.