Knowledge and appropriateness of care of family physicians and physiotherapists in the management of shoulder pain: a survey study in the province of Quebec, Canada

This descriptive study used a cross-sectional survey design that follows the guidelines for reporting survey-based research [29]. The study was approved by the Health Research Ethics Committee of the CIUSSS-de-l’Est-de-l’Île de Montréal (2021–2224) in Montreal, Quebec, Canada.

The survey was sent to physiotherapists via the email list and social media accounts of the Ordre professionnel de la physiothérapie du Québec (OPPQ) (Quebec’s Physiotherapy Professional College) and to family physicians through the Réseau-1 Quebec newsletter. Réseau-1 Quebec is a primary care knowledge and research network for clinicians and researchers aiming at facilitating research and uptake of evidence in primary care. The link of the survey was also sent via the email lists of selected physiotherapy clinics and university family medicine groups where the research team has ongoing collaborations (n = 6). The survey was active from February 18^th to June 11^th, 2021. Based on our previous surveys sent out via professional associations, we expected a participation rate of 3 to 5% [30, 31]. Considering that approximately 5200 physiotherapists were licensed in 2020[32], we expected that 156 to 260 physiotherapists would answer the survey. Since approximatively 4700 family physicians are working in family medicine groups in the province of Quebec[32], we expected that 141 to 235 family physicians would answer the survey, [30, 31].

Socio-demographic and clinical characteristics of the respondents were collected. The survey included four clinical vignettes presenting patients with the following shoulder pain conditions: rotator cuff (RC) tendinopathy, acute full-thickness (FT) RC tear, adhesive capsulitis or traumatic anterior glenohumeral instability. The vignettes were selected and adapted by our research team based on published survey-based studies on shoulder pain management conducted in the United States, Australia and the United Kingdom [22, 23, 33]. The four vignettes are presented in Additional file 1. Each vignette was followed by a questionnaire on initial shoulder pain management and treatment that the respondent would recommend. This section of the survey was developed by the research team based on previous studies evaluating family physicians and physiotherapists’ evidence-practice gap for general MSK disorders and shoulder pain management [22, 23, 33‐36]. We also questioned family physicians and physiotherapists on their overall confidence level (not confident at all, slightly confident, somewhat confident, fairly confident, completely confident) in making an appropriate diagnosis, selecting appropriate investigations, adequately referring patients to a medical MSK specialist, and selecting appropriate treatments when they are managing shoulder pain in their everyday practice.

To determine the appropriateness of the physiotherapists’ and family physicians’ care offered for all vignettes, recommended management was based on previous survey studies using the same clinical vignettes [22, 23, 33] as well as on the evidence-based recommendations of a recent CPG covering the initial management, medical and rehabilitation treatments of RC disorders in the context of the province of Quebec developed by our team [17]. We also systematically reviewed the literature to identify CPGs covering the management of RC disorders and other common shoulder disorders such as adhesive capsulitis and glenohumeral instability (PROSPERO: CRD42022325614) [37]. In the systematic review, we used a combination of keywords and Medical Subject Headings (MesH) terms including “shoulder”, “rotator cuff”, “adhesive capsulitis”, “GH osteoarthritis”, “GH instability”, “acromioclavicular” and “guidelines”. The search of CPGs published between 2008 and August 2022 was performed in four databases (Medline, Embase, Physiotherapy Evidence Database [PEDro], Google Scholar) and in international CPG databases. In the systematic review process, two reviewers assessed the methodological quality of the CPGs with the AGREE (Appraisal of Guidelines Research and Evaluation) II checklist and extracted the recommendations [38]. We identified five high-quality CPGs covering the management of RC disorders [6, 13, 15, 17, 18], two high-quality CPGs covering the management and indications for diagnosis imaging of adhesive capsulitis [13, 14] and two high-quality CPGs including indications for diagnosis imaging of traumatic anterior glenohumeral instability [6, 13]. We considered which CPGs were of high quality by using a frequently reported method in which domain three and at least two other domains of the AGREE II checklist had a score equal or over 60% [39]. We also included the CPG covering the medical and rehabilitation management of traumatic anterior glenohumeral instability with the highest overall score, since according to our assessment, no high quality CPGs covered the management of this condition [16]. Recommendations on shoulder pain management that were used to asses appropriateness of care are summarized in Table 1. Appropriateness of care was determined by one author (VL) and revised by a second author (FD).

Descriptive statistics were summarized for demographics and clinical characteristics of respondents and for results on shoulder pain management. For analysis purposes, we recoded results regarding the level of confidence into not confident (not confident at all or slightly confident), moderately confident (somewhat confident) and highly confident (fairly confident or completely confident). We also recoded the level of priority of rehabilitation interventions selected by physiotherapists as not a priority or low priority (0 and 1 on the 6-point scale), moderate priority (2 and 3 on the 6-point scale) and high priority (4 and 5 on the 6-point scale). Results regarding the initial management of shoulder disorders, medical treatments selected and confidence in shoulder pain management were compared between physiotherapists and family physicians using Fisher’s exact tests. We used Excel Version 16 to summarize data and RStudio Version 1.4.1106 for all statistical analysis. The alpha level was set at 0.05.

Socio-demographic and clinical characteristics of respondents are presented in Table 2. More than 71% of the physiotherapists and family physicians were women. Most physiotherapists (62%) and family physicians (65%) had ten years of experience or less. Seventy percent of physiotherapists were working in private practice and 93% of family physicians practiced in a public setting, mostly in family medicine groups (96%).

When comparing the confidence level of family physicians and physiotherapists in shoulder pain management (Table 3), there were statistically significant differences (p < 0.001) favoring physiotherapists in confidence in making an appropriate diagnosis and selecting adequate treatments. More physiotherapists reported being highly confident for these type of management (respectively 64% and 80%) compared to the majority of family physicians that reported being only moderately confident (respectively 55% and 50%). No significant differences were observed between providers for selecting appropriate investigations and adequately referring patients to medical MSK specialists.

Selected diagnosis, indication for investigations and indication for medical MSK specialist referrals by family physicians and physiotherapists are presented in Table 4.

Medical treatments selected by family physicians and physiotherapists in the management of shoulder pain patients described in the clinical vignettes are presented in Fig. 1 (a to d).

The rehabilitation treatments prescribed by family physicians are presented in Table 5. The rehabilitation treatments recommended by physiotherapists are presented in Figs. 2, 3, 4 and 5.

The vast majority of physiotherapists were highly confident in diagnosing and managing shoulder pain but only a minority of physicians were. Physiotherapists do have extensive training in the diagnosis and conservative management of shoulder pain. At least a third of physiotherapists’ training in Canada is in the management of MSK disorders [42], while around 3% of the Canadian undergraduate family physician curriculum is dedicated to MSK management [43]. This may explain the significant differences between confidence of family physicians and physiotherapists in selecting a diagnosis and treatments for shoulder pain patients. Moreover, in our sample, experience with treating shoulder pain was not equivalent between providers. Physiotherapists reported offering care to shoulder pain patients more frequently.

A very high proportion of family physicians and physiotherapists selected the appropriate diagnoses for the presented vignettes. These results are encouraging since family physicians in our study reported only being moderately confident in selecting a diagnosis and treatments for shoulder pain patients, while most physiotherapists reported being more confident. The family physicians’ performance here is comparable to a United Kingdom study in which 82% and 92% of respondents adequately identified a RC tendinopathy and an adhesive capsulitis, based on the same vignettes as those in our study [22]. However, these survey results may not represent clinical practice as shoulder pain history and signs and symptoms described in the clinical vignettes were relatively clear and uncomplicated [44]. Also, respondents from the survey may not be totally representative of the general population of family physicians and physiotherapists because participants may be more likely to have an interest in shoulder pain or MSK disorders management.

For all vignettes, family physicians recommended significantly more investigations (30 to 88%) than physiotherapists (13 to 67%). In RC tendinopathy or adhesive capsulitis vignettes, the performance of family physicians was not optimal as imaging tests are not recommended for the initial management of these two disorders. The proportions observed here are however lower than in studies conducted in the United Kingdom and Australia in which up to 82% of family physicians did recommend imaging for these shoulder cases [22, 23]. It has been reported that family physicians often overprescribe investigations because of the fear of missing a serious pathology [45]. The results from our survey study tend to confirm this since up to 87% of family physicians that ordered an imaging test, did so to exclude other pathologies.

Other studies looking at the ability of physiotherapists to adequately refer for diagnostic imaging report a referral rate of up to 31% for tendinopathy [27, 46] and 54% for adhesive capsulitis management [47], which is higher than in our study (13% and 22% respectively). Thus, in cases where diagnostic imaging is not recommended, physiotherapists here were more likely to manage shoulder pain as recommended by CPGs. It remains unclear if this behaviour is related to the scope of practice of physiotherapists in the province of Quebec where they do not autonomously prescribe imaging tests in most situations.

As for the management of acute FT RC tear and traumatic anterior glenohumeral instability, referring the patient for an investigation and a surgical opinion are recommended, but physiotherapists less often recommended such care. More family physicians appropriately managed these disorders. Physiotherapists may be more confident in the efficacy of non-surgical treatment [48], but only at a later stage would they recommend a diagnostic imaging if the patient’s condition is not improving. However, even though a rehabilitation program can be initiated for these pathologies [16, 17], standard of care is to refer these patients to a medical MSK specialist [16, 17]. Only 27 to 46% of family physicians and 30 to 37% of physiotherapists did initially refer the patient to an orthopaedic surgeon for a surgical opinion for acute FT RC tear and traumatic anterior glenohumeral instability. This result is concerning since delays in surgery are associated with poorer outcomes for these disorders [49‐53]. The low levels of referral to medical MSK specialists in our study may be explained however by the administrative requirement of several orthopaedic departments in the Province of Quebec to include results of a diagnostic ultrasound or an MRI when referring the patient for a surgical consultation [54]. Since the survey questions were on shoulder pain management at the initial consultation, several respondents may have not indicated that they would refer the patient to a medical MSK specialist at that moment since they would wait confirmation of the diagnosis with the imaging results.

Most physiotherapists (65 to 95%) did not recommend any medical treatments for the four clinical vignettes, although oral NSAIDs and acetaminophen are recommended in painful shoulder conditions [14, 17]. Potential medical modalities that can reduce patients’ level of pain may be underused by physiotherapists, which may be explained by their scope of practice not allowing them to autonomously prescribe medication. However, physiotherapists are able to use effective active modalities and exercises to reduce pain and improve patients’ function [55]. In the four clinical vignettes, 39% to 87% of family physicians have indicated that they would prescribe oral NSAIDs or acetaminophen, as recommended by CPGs for short-term pain reduction [14, 17]. In the management of RC tendinopathy, there was a significantly higher reliance on the use of corticosteroids infiltrations by family physicians (27%), compared to physiotherapists (2%), even though this modality should not be used as an initial treatment for this pathology [17]. Using corticosteroids infiltrations in the management of RC tendinopathy by family physicians was also too often recommended in the study by Buchbinder et al. (24%) [23] and the study by Artus el al. (48%) [22]. The over-reliance on corticosteroids infiltration in the management of shoulder pain observed in our survey may be partially explained by the lack of other treatment options, such as poor access to free of charge physiotherapy in the Province of Québec or also patient preferences where some could prefer a quick reduction of symptoms [56, 57]. In the management of adhesive capsulitis however, an intra-articular corticosteroids infiltration is recommended, preferably combined with physiotherapy treatments including mobility exercises [14]. It appears that there is a need for education and other guideline implementation strategies [58] on the appropriate management of adhesive capsulitis among clinicians since 24% of family physicians and 38% of physiotherapists did not recommend an intra-articular corticosteroids infiltration or an arthrographic distension.

Regarding rehabilitation care, most family physicians reported that they would refer the patients presented in the vignettes to a physiotherapist (85–95%) and provide advice and education (78–89%), as recommended [14, 16, 17]. These proportions are higher than in previously published survey studies (57–77%) [22, 23]. However, in clinical practice, the actual referral rate to physiotherapists may be lower, because of the lack of access to publicly funded physiotherapy in our health care system, as already mentioned [56]. Several observational studies reporting on shoulder pain management by family physicians have demonstrated an actual referral rate to physiotherapy of 13–53% for patients with RC disorders or adhesive capsulitis [59‐62]. Less family physicians indicated that they would give the patient a home exercise program (50–65%), despite this modality being an essential component of rehabilitation [63]. The development of self-management modules including exercises, general advice and education that are accessible to family physicians and patients could benefit patients’ condition but are not commonly available in clinical practice actually [64].

Almost all physiotherapists adequately indicated that they would provide advice and education to the patients, as recommended. In the management of RC tendinopathy, RC tear and glenohumeral instability, active exercises are recommended [16, 17]. However, for up to 43% of physiotherapists, strengthening and active mobility exercises were not indicated as being an important priority. These results are comparable to two studies that reported that only 54% to 67% would prescribe strengthening exercises in shoulder pain management [25, 26]. Passive modalities such as electrotherapy and thermotherapy are either not recommended or there are no recommendations in the management of most shoulder disorders [14, 16, 17]. Nonetheless, up to 65% and 45% of physiotherapists considered thermotherapy and electrotherapy as being a moderate or a high priority treatment. Active modalities should be prioritized in shoulder pain management since they can promote patients’ self-efficacy, and patients with higher levels of self-efficacy have a lower risk to develop chronic MSK pain [65, 66]. These results indicate a need to implement recommendations on evidence-based treatments that should be used for shoulder pain patients in clinical practice with strategies such as training and education, support of clinicians and development of relationships [4, 58].

This is the first study to evaluate and compare the management of shoulder pain between family physicians and physiotherapists since 2002 [67] and the only study to evaluate shoulder pain management in the Canadian context in the last twenty years [68]. The clinical vignettes were used in previous family physicians survey studies in other countries, which allowed us to compare management across settings [22, 23, 33]. To evaluate appropriateness of care, several high quality CPGs were selected using a systematic review process with a methodological quality assessment of their content [37, 38].

However, our study presents some limitations. Using high quality CPGs to assess quality of care of respondents involve that evidence from recent individual studies or systematic review may have not been considered in the evaluation. However, high quality CPGs are the highest level of evidence to help clinician manage patients’ conditions [69]. The principal limitation of our study is that the sample size of the study was relatively small. In the context of the COVID-19 pandemic, healthcare providers were very busy in their clinical practice, which may have limited their time to participate in this survey study. Another limitation of survey studies is that clinicians that decided to participate in the survey on shoulder pain management potentially have a greater interest in managing these disorders. Also, most respondents had less than 10 years of experience. Therefore, they may not be representative of all clinicians, physicians or physiotherapists. This bias may be especially present for family physicians in our study since our sample size is small and the management of family physicians was more concordant with recommendations than in previous studies using the same vignettes [22, 23]. Finally, survey studies with clinical vignettes do not entirely represent real clinical decision making and practice of family physicians or physiotherapists. Surveys may be easier to answer and does not take into account the clinical ability of providers in performing a valid questionnaire and physical examination. Yet using clinical vignettes has been reported as a valid and cost-effective option to evaluate health providers practice variations [70, 71].