UK-based physical therapists' attitudes and beliefs regarding exercise and knee osteoarthritis: Findings from a mixed-methods study




Within the UK, differences exist between physical therapists' use of exercise for patients with knee osteoarthritis (OA) and recent exercise recommendations. This may be explained by their underlying attitudes and beliefs. We aimed to describe UK physical therapists' attitudes and beliefs regarding exercise and knee OA, and understand and explain them.


A survey was mailed to 2,000 UK-based chartered physical therapists that included 23 attitude statements derived from recently published recommendations. Semistructured telephone interviews were conducted with a purposeful sample of questionnaire respondents (n = 24), and were recorded and analyzed thematically.


The questionnaire response rate was 58% (n = 1,152); 538 respondents reported treating a patient with knee OA in the last 6 months. The survey highlighted uncertainty about potential benefits of exercise for knee OA: only 56% largely/totally agreed that knee problems are improved by local exercise. Although exercise adherence was deemed important, it was seen as the patient's, not the therapist's, responsibility. Interviews revealed an underlying biomedical model of care of knee pain, with knee OA viewed as a progressive degenerative condition. A paternalistic treatment approach was evident. Health care systems presented a number of barriers to best practice, including limited opportunity to provide followup.


Although the attitudes and beliefs of physical therapists may help to explain differences between current practice and recent exercise recommendations, the wider health care system also plays a part. Further research is needed to support meaningful shifts in physical therapy care in line with the best practice recommendations.


Clinical guidelines and systematic reviews consistently recommend exercise as a core treatment for patients with knee osteoarthritis (OA) (1–7), although there is a lack of evidence around the practical aspects of exercise prescription, including the appropriate dose and how best to support individuals to continue exercise over the long term. Recent recommendations incorporating research-based evidence and expert opinion address specific questions about the role of exercise in lower extremity OA (1), including the benefits and safety of strengthening and aerobic exercise, the effectiveness of exercise completed in different settings, and the importance of exercise adherence (Table 1).

Table 1. Summary of the 10 propositions included in the recent exercise recommendations for hip and knee OA (adapted, with permission, from ref.1)*
PropositionCategory of evidence (1–4)Strength of recommendation (A–D)
  • *

    OA = osteoarthritis; RCT = randomized controlled trial.

  • 1A = meta-analysis of RCT; 1B = at least 1 RCT; 2A = at least 1 controlled trial without randomization; 2B = at least 1 type of quasi-experimental study; 3 = descriptive studies (comparative, correlation, case–control); 4 = expert committee reports/opinions and/or clinical opinion of respected authorities.

  • A = directly based on category 1 evidence; B = directly based on category 2 evidence or extrapolated recommendation from category 1 evidence; C = directly based on category 3 evidence or extrapolated recommendation from category 1 or 2 evidence; D = directly based on category 4 evidence or extrapolated recommendation from category 1, 2, or 3 evidence.

Both strengthening and aerobic exercise can reduce pain and improve function and health status in patients with knee and hip OA.Knee 1BA
Hip 4C (extrapolated from knee OA)
There are few contraindications to the prescription of strengthening or aerobic exercise in patients with hip or knee OA.4C (extrapolated from adverse event data)
Prescription of both general (aerobic fitness training) and local (strengthening) exercises is an essential, core aspect of management for every patient with hip or knee OA.4D
Exercise therapy for OA of the hip or knee should be individualized and patient centered, taking into account factors such as age, comorbidity, and overall mobility.4D
To be effective, exercise programs should include advice and education to promote a positive lifestyle change with an increase in physical activity.4D
Group exercise and home exercise are equally effective and patient preference should be considered.1AA
Adherence is the principal predictor of long-term outcome from exercise in patients with knee or hip OA.4D
Strategies to improve and maintain adherence should be adopted, e.g., long-term monitoring/review and inclusion of spouse/family in exercise.1BA
The effectiveness of exercise is independent of the presence or severity of radiographic findings.4Not recommended
Improvements in muscle strength and proprioception (balance) gained from exercise programs may reduce the progression of knee and hip OA.4D

Physical therapists play a key role in managing patients with knee OA through designing and supervising exercise programs. Within the UK, some differences exist between the physical therapists' use of exercise for patients with knee OA and the recent recommendations: local strengthening exercise is prescribed more frequently than general aerobic exercise, it is delivered over relatively few treatment sessions with limited supervision, and followup is not common (8). These differences may limit the potential benefits of the exercise program.

The attitudes and beliefs held by health care professionals, including physical therapists, have been shown to be associated with their clinical practice behavior (9–11), and may help explain the limited uptake of clinical guidelines and the best practice recommendations (12, 13). At present, little is known about physical therapists' attitudes and beliefs regarding exercise and knee OA. In a mixed-methods study, we aimed to describe UK physical therapists' attitudes and beliefs regarding exercise and knee OA, with a specific focus on recent recommendations (1), and to understand and explain them.


We performed a postal survey and semistructured interviews with chartered (licensed) physical therapists practicing within the UK between January and December 2006. Ethical approval was granted by the North Staffordshire Research Ethics Committee. Data were gathered sequentially: surveys were completed initially, and preliminary quantitative data analyses then directed the focus of the interviews. For clarity, the survey and interview methods are described separately. Quantitative and qualitative results and discussion are reported together to facilitate a comprehensive exploration of the findings (14).

The survey.

A cross-sectional questionnaire was mailed to a simple random sample of physical therapists, identified from the membership list of the professional body, the Chartered Society of Physiotherapy (CSP). A filter question at the beginning of the questionnaire screened for physical therapists that reported treating at least 1 patient with knee OA in the past 6 months. Based on the results of a pilot study (n = 200), 1,800 questionnaires were mailed for the main study and a second questionnaire was sent to all of the nonresponders at 4 weeks in order to achieve approximately 400 applicable responses for analyses. This allowed survey estimates (proportions) to be calculated with a margin of error of less than 5% and with 95% confidence (15). Only very minor changes to the questionnaire were made following the pilot study; thus, these were combined with the main study for analyses.

Survey instrument.

To explore physical therapists' attitudes and beliefs about exercise and knee OA, a list of 23 attitude statements was developed directly from the recent exercise recommendations (1) and tested with 12 physical therapists working in the field of musculoskeletal research and/or knee pain. Respondents were asked to rate their agreement with each statement on a 6-point Likert scale ranging from 1–6, where 1 = totally disagree and 6 = totally agree (for clarity, the statements have been grouped into those relating to benefits of exercise for knee OA and those relating to the delivery of exercise and exercise adherence) (Figures 1 and 2). The questionnaire also included demographic questions, a measure of physical therapists' illness perceptions (16) and treatment orientations (11, 17), and a vignette describing a patient with knee OA and clinical management questions (8).

Figure 1.

Percentage of physical therapists that largely or totally agreed with attitude statements relating to perceived benefits of exercise. OA = osteoarthritis.

Figure 2.

Percentage of physical therapists that largely or totally agreed with attitude statements relating to the delivery of exercise and exercise adherence. OA = osteoarthritis.

Survey data analysis.

Data analyses were carried out using the Statistical Package for Social Scientists, version 14 (SPSS, Chicago, IL). The level of agreement with each attitude statement was determined by the percentage of respondents who largely or totally agreed. Because no standard definition for consensus exists (18), we defined 100% = unanimity, 75–99% = consensus, 51–74% = majority view, and 0–50% = no consensus, which are thresholds used within the guidelines of the CSP for whiplash-associated disorder (18) and persistent low back pain (19).

The interview study.

Semistructured interviews were completed over the telephone with physical therapists that returned the questionnaire, had treated a patient with knee OA in the last 6 months, and provided consent for further contact. Following a review of questionnaire responses, the sample was purposefully selected to include male and female therapists with different levels of clinical experience (i.e., length of time since qualification) and a range of survey responses, including attitudes about exercise for knee OA. We aimed to conduct between 20 and 30 interviews.

The interview schedule.

The interview schedule was developed following preliminary analyses of survey data, allowing areas of interest to be explored in depth. Physical therapists were probed about their responses to a number of the attitude statements, and open-ended questions allowed interviewees to raise and discuss other relevant issues. Qualitative data analyses were conducted in parallel with further data collection, allowing the interview guide to be modified with ongoing analysis, and a thorough exploration of emerging themes. Two researchers completed all of the interviews (MAH and JY), which were audio taped, transcribed verbatim, and anonymized for analysis. Data collection ceased when no new themes were emerging.

Interview data analysis.

Interview data were thematically analyzed, aided by the computer software package NVivo, version 2.0 (QSR International, Analysis was based on the principles of grounded theory (20, 21). Two researchers (MAH and JY) reviewed all of the interview transcripts and by constantly searching for commonalities and contrasts within the data, identified and developed themes that were clustered into categories that shared meanings (21, 22). Categories were reappraised and revised throughout the entire analysis process and links between categories were explored (21, 22). Emerging ideas were discussed and checked for credibility with a third member of the study team (NEF), who independently coded a number of transcripts. In the final stages of analysis, a qualitative researcher independent to the study team analyzed 3 interview transcripts, and her interpretation of the interviews was compared with that of the research team. No changes resulted from this. In contrast to grounded theory, themes and categories did not solely emerge from the raw data, but were also informed by the survey findings. In addition, possible links between categories were not intended to form a formal theory, but to complement, expand, and provide understanding to the survey findings (21, 22). In the Results section, the categories are shown in quotations, the themes are italicized, and in Table 2, verbatim, anonymized quotations are used to exemplify each theme.

Table 2. Categories, themes, and verbatim examples of themes uncovered in the interview study
ThemeVerbatim example
Views of knee osteoarthritis 
 Biomedical perspective“This idea of this arthritis, one of the things I like to do with the patients is to make sure (that) they've got as full a range of motion as possible, with the idea of trying to lubricate the joint and to get the synovial fluid to give its nutrition through to the cartilage, and I think one of the ways of doing that is moving the knee through that full range of motion and I think there's a danger, with patients, if they're not … if they're in pain, or getting stiffness (such) that they tend to avoid the movement and then they just get stiffer and if they're getting stiffer, they're not getting that lubrication and they're not feeding their cartilage to try and keep it as healthy” (participant ID: 79).
 Chronic degenerative condition“My guess is that, regrettably, long term her knee is only going to get worse, assuming that it is, sort of, arthritic and, really, a knee replacement is the answer” (participant ID: 378).
 Surgery“I mean, if they've got a really bad crumbly knee and it's got a big effusion in it and it's hot and everything hurts, then sometimes it's past the point of help, physio-wise and it would need surgery” (participant ID: 536).
 Worsen, irrespective of physical therapy intervention“The arthritic degeneration is going to continue to ‘rumble on’ ‘pretty much’ whatever we do and so, ultimately, if it was reviewed in three years, I feel quite sure that the arthritic changes would be worse, irrelevant of what we do” (participant ID: 1376).
Views of exercise for knee osteoarthritis
 Strengthen and stabilize“The exercises are important because the stronger your quads are, the more chance you have of preventing damage to the knee, because it's not, sort of … the tibia's not ‘drifting around’ on the end of the femur, somewhere. You've got good muscle bulk and stability and hopefully the ligaments don't become too lax. So, you've got good anterior/posterior plane movement” (participant ID: 282).
 Exercise more effective with less damage or pain“I think if you've got mild to even moderate symptoms, an exercise program to build up the quads and the hamstrings and the supporting muscles can have an effect on the pathology of the condition itself. It can help to take some of the stress out of the joint and also, assist with the prevention of a deformity forming. I think once you get severe arthritis and you've got boney changes and joint deformity, I think exercise is as I said, sort of moderately effective because you've got the underlying pathology working against you and the change is going to be irreversible” (participant ID: 343).
 Barriers“I just feel that when it becomes a more severe case of arthritis, that apart, maybe, from hydrotherapy, it's very difficult for people to actually go in a gym and do something and feel quite comfortable to do it without any pain” (participant ID: 1432).
 “You could just go down the leisure route and try and encourage her to do stuff that would help her (to) lose weight, as well as keep her knee strong and keep her flexible, but because she's mildly hypertensive, it would depend, really. I mean, if it's controlled by (being) on drugs …” (participant ID: 633).
 Pain contingent model“I wouldn't be going ‘overboard’ with strenuous exercise. It would have to be moderate exercise; being built up and carefully monitored, so that the patient didn't cause even more damage to the knee that she'd got already” (participant ID: 14).
Exercise adherence 
 Importance of exercise adherence“Your chances of getting any improvement in her condition in the longer term, or a medium term are very dependent on her doing her exercises regularly and the treatment is completely ineffective if she doesn't” (participant ID: 343).
 The dose-response effect“Of course, if they're not complying with that and they're not doing the exercises as you've prescribed them, they won't be as effective in the same way as a doctor prescribes you to take four painkillers a day and you only take two, you're not going to be as comfortable as if you'd taken the four” (participant ID: 998).
 Negative perception of patients' levels of exercise adherence“I think as soon as they leave [physical therapy services], the majority of people would let it go ‘by the by.’ It may be only a small percentage that carry on” (participant ID: 1433).
 Stereotypes“Quite often, you'll find you have two types of patient, well, that's a bit generalistic, but you do. You've got those patients who are very keen to do the exercises and will do them regardless of pain ‘and so on’ and you have the patients who just ‘won’t go there at all'” (participant ID: 14).
 Patient-centered barriers to adherence“Again, it does depend on their mindset. Some are just not receptive to it at all. I mean, we've got to face facts that some are just lazy. Most aren't and most, when you explain why you're teaching them certain exercises, and what the likely outcome will be, most people will make changes, but there are some that won't” (participant ID: 502).
 Blame“I don't think I'd ‘take the blame’ if somebody didn't do something, I wouldn't particularly ‘take the blame’ that I hadn't passed it on…but, I would, if I'd given them something too hard to do, or I hadn't totally examined what…. So, I suppose, in that respect I'm making sure I've gone over all options, as to what they can do and how they can achieve it and then, I'd have to pass, you know, then, I think it's their responsibility” (participant ID: 1254).
Paternalistic approach 
 Distinctive roles“I think it's the Physio's responsibility to motivate the patient and explain why they're doing it, but, ‘in the end,’ it's their responsibility to do the actual exercise. So, once you've presented the case and presented the information, and the reasoning for it, and what you're expecting them to achieve and by what time and when; it's totally up to them to do that, but with your support, initially” (participant ID: 1254).
 “Obviously, I would say to them, ‘it’s your life.' I mean, ‘but we’re trying to improve and help you manage your lifestyle, manage your pain and obviously, this is all part of it. If you feel that you want to come and get on board and help yourself, then these will help you. If you don't, then it's just a waste of time'” (participant ID: 1432).
 Shared responsibility of exercise“I feel it is the responsibility of both. Ultimately its important as Physios that we empower people, that we don't create an environment where people are dependent upon us, otherwise we create victims rather than individuals. So, it's very important that part of our role is to pass on the ownership of the program to the patient and therefore that becomes more and more their responsibility, but during the process, we must also take the responsibility as their therapist to contact that patient. Even if we're not seeing them on a regular basis any more, it's important that they're aware that we are there as a support structure, and there'll be occasions where I don't feel that any further physiotherapy would be of benefit to somebody but I will book them back in, for six weeks or for three months, whatever, simply because it will ensure or it will increase the possibility of that person continuing with her exercises” (participant ID: 1376).
The wider health care system 
 Structure of the system“My frustration is not being able to follow people up sufficiently to actually identify those [nonadherent patients] and this is because of restraints of time, and in the private sector, restraints of people, of the cost for the individual patient” (participant ID: 343).
 Knowledge and skills“Some patients are very…you know, they'll come along; they will listen to the reason why they need to change their activity and will readily change. Some patients are just so resistant to change. I mean, I'm thinking of one lady, at the moment and there's such psychological barriers there and I don't find I've got the time or the resources to deal with that” (participant ID: 502).



Of the 2,000 questionnaires sent, 8 were excluded (due to incorrect addresses) and 1,152 responded (adjusted response rate 58%). Of these, 538 (47%) physical therapists reported treating at least 1 patient with knee OA in the last 6 months, and these applicable survey responses are summarized below. The characteristics of the survey sample are shown in Table 3. From the applicable survey responses, 84 physical therapists were purposefully selected and invited to participate in the interviews, and 24 agreed to take part and were interviewed. The characteristics of the interview sample are shown in Tables 3 and 4.

Table 3. Physical therapists' characteristics*
 Survey respondents (n = 538)Interview participants (n = 24)
  • *

    Values are the number (percentage). Individual items may not add to totals due to missing data. NHS = National Health Service; OA = osteoarthritis.

  • Training varied and included in-house training (e.g., a seminar on current knee OA research), specific short courses (e.g., Pilates instruction), and formal training as part of a degree (e.g., a physical therapy master's degree or bachelor's degree in exercise physiology).

  • Measured by the Short Telephone Physical Activity Recall Questionnaire (39), modified for use within the postal survey.

Women465 (87)16 (67)
Clinical experience, years  
 1–3110 (21)3 (13)
 4–10132 (25)9 (39)
 ≥11278 (54)11 (48)
Work setting  
 Exclusively NHS301 (56)10 (42)
 Combination of NHS and non-NHS113 (21)6 (25)
 Exclusively non-NHS123 (23)8 (33)
Number of patients seen with knee OA  
 Less than 1 per month110 (21)6 (26)
 At least 1 per month249 (47)12 (52)
 At least 1 per week175 (33)5 (22)
Postgraduate training  
 Knee OA127 (24)7 (30)
 Exercise therapy188 (36)10 (42)
Personal experience of knee pain326 (61)15 (63)
Physical activity level  
 Meets current recommendations (38)315 (60)14 (58)
 Active but not sufficient to meet current recommendations208 (40)10 (42)
Table 4. Interview participants' characteristics*
Study no.SexWork settingYears of clinical experiencePostgraduate training
Knee OAExercise therapy
  • *

    OA = osteoarthritis; NHS = National Health Service.

0014FemaleNHSMissing dataNoYes
0343FemaleNHS + non-NHS32NoNo
0378FemaleNHS + non-NHS30NoNo
0502FemaleNHS + non-NHS26YesNo
0633FemaleNHS + non-NHS22Missing dataNo
1261FemaleNHS + non-NHS6YesYes
1523MaleNHS + non-NHS4NoYes

Perceived benefit of exercise for knee OA.

Fifty-five percent of physical therapists largely or totally agreed that “knee problems are improved by local strengthening exercises,” and that “physical therapists should prescribe local strengthening exercises for every patient with knee OA” (Figure 1). No consensus was reached regarding these statements about general exercise (44% and 43%, respectively), the safety of local or general exercise for every patient with knee OA (35% and 24%, respectively), or that increasing knee strength or overall activity level stops the progression of knee OA (19% and 8%, respectively). Although the majority of physical therapists largely or totally agreed that “exercises are effective for patients if an x-ray shows mild knee OA” (72%) or moderate knee OA (63%), there was no consensus that “exercises are effective for patients if an x-ray shows severe knee OA” (30%) or that “exercise works just as well for everybody, regardless of the amount of pain they have” (12%). Themes from the interview study offered some explanation for these findings. Interviewees viewed knee OA from a biomedical perspective, attributing signs and symptoms to local knee joint pathology or wear and tear. OA was seen as a chronic degenerative condition that would progressively worsen over time, the only cure being surgery. Therefore, therapists reported that the knee condition would worsen, irrespective of the physical therapy intervention, including exercise therapy. Interview data showed that an important aim of an exercise program for knee OA from therapists' perspectives was to strengthen and stabilize the joint rather than provide pain relief. This may explain why, within the survey, local strengthening exercise was viewed more favorably than general aerobic exercise. As suggested within the survey, the effectiveness of an exercise program was seen as being related to the severity of joint damage, or pain level, with therapists believing exercise to be more effective the less damage or pain a patient was experiencing.

In accordance with the biomedical view of knee OA, a number of barriers to physical therapists prescribing exercise for knee OA were identified, including: fear of increasing symptoms; causing progression of the disease, particularly through weight-bearing activities; and exacerbating patients' comorbidities. The therapists appeared to be working mostly within a pain contingent model, and talked about feeling reluctant to promote exercise in the presence of pain. This helps to explain the concerns about the safety of exercise highlighted by the survey.

Delivery of exercise and adherence.

There was consensus among survey respondents that “exercise for knee OA is most beneficial when it is tailored to meet individual patient needs” (87%), and “physiotherapists should educate chronic patients with knee OA about how to change their lifestyle for the better” (80%). However, no consensus was reached regarding the statement, “it is important that people with knee OA increase their overall activity levels” (36%) (Figure 2). There was consensus of agreement among physical therapists that “how well a patient complies with their exercise program determines how effective it will be” (81%), and that “it is the patient's own responsibility to continue doing their exercise program” (85%). Only 5% largely or totally agreed that “it is the physiotherapist's responsibility to make sure that the patient will continue doing their exercise program.” The category “exercise adherence” was explored within the interviews. Because knee OA was viewed as a chronic condition, physical therapists expected patients to maintain exercise programs over the long term, and as such, recognized the importance of exercise adherence and the link between the level of adherence and clinical outcomes (the dose-response effect). Overall, interviewees held negative perceptions of patients' levels of exercise adherence, which may contribute to the feeling that knee OA would worsen despite the prescription of exercise therapy. Physical therapists described stereotypes of patients who would or would not adhere to an exercise program. Many barriers to exercise adherence were identified, and these were commonly seen as patient centered, such as lack of motivation or laziness, human nature, pain, fear of harm, and negative treatment expectations. Some physical therapists appeared to blame patients for not completing an exercise program as it had been prescribed. This may be explained by an evident category of “paternalistic approach,” in which the health care professional decides on the treatment approach and the patient is expected to comply despite not being involved in the decision-making process (23). Interview participants talked about distinctive roles of both the physical therapist and patient. The therapist's role was seen as assessment, exercise prescription, and education. It was seen as the patient's role to follow the prescribed exercise program over the long term and get on board with the treatment being offered. The relatively short-term responsibilities of the therapist may help explain why limited treatment sessions are offered to this patient group (8): “Four or five times [number of treatment sessions provided] would be, you know: obviously assessment, setting up your program, a few weeks of monitoring of the program to make sure it's going okay, and progressing it, as needed, but then with the emphasis on the fact that she can carry on that program without my supervision” (participant identification: 191).

Some physical therapists did recognize the potential influence they could have on exercise adherence, and shared the responsibility of exercise adherence with the patient. Interestingly, the ultimate treatment goal was similar: to enable patients to independently complete exercises over the long term.

Service issues.

Although not explored within the survey, an important category within the interview study was the role of “the wider health care system.” The structure of the system in which physical therapists worked influenced their treatment approach toward exercise and knee OA. Limited time to review individual patients reduced opportunities to facilitate behavior change, large caseloads and pressure of waiting lists reduced the number of treatment sessions provided, and limited opportunity to provide followup sessions after discharge coupled with poor links to community facilities such as local leisure centers left some therapists with no option other than to deliver a self-directed exercise program with no long-term support or followup. Some physical therapists also identified gaps in their knowledge and skills, including how to facilitate behavior change, particularly with less motivated patients.


To our knowledge, this is the first national survey within the UK to explore physical therapists' attitudes and beliefs regarding exercise and knee OA, and the mixed-methods approach has allowed a more thorough exploration and deeper understanding than the survey results provided alone (14). Due to the likely association between therapists' attitudes and beliefs and their clinical behavior, better understanding of these attitudes and beliefs is important in explaining clinical behavior and in working toward improved ways of supporting therapists to provide clinical care in line with recent exercise recommendations for knee OA (1).

The study revealed uncertainty among physical therapists about the potential benefits of exercise for knee OA. Only approximately half largely or totally agreed that knee problems are improved by local strengthening exercise, and even fewer agreed that knee problems are improved by general exercise. This contrasts with the recent exercise recommendations that state “both strengthening and aerobic exercise can reduce pain and improve function and health status in patients with knee and hip OA” (1).

Some therapists had an underlying belief that exercise was more effective for patients with less severe knee joint damage. The biomedical approach evident within interview data, particularly that knee OA is a degenerative, progressive condition, may underpin these attitudes and beliefs: if the condition is going to worsen over time, why would exercise help? Although this contrasts with the proposition within the exercise recommendations, “the effectiveness of exercise is independent of the presence or severity of radiographic findings,” there was no evidence supporting this (1). The gap in the literature and uncertainty in the clinical setting about the potential value of exercise for different patient groups highlight the need for further research to address this issue.

The association between knee OA and wear and tear may explain the concerns that physical therapists have about the safety of exercise, although evidence from clinical trials suggests that moderate-intensity aerobic and strengthening exercise does not worsen the disease (24). Within the biomedical model, pain is a consequence of tissue damage. This helps explain why physical therapists tended to adopt a pain contingent approach to exercise for knee OA, and discouraged exercise in the presence of pain. To address the differences between physical therapists' attitudes and beliefs and the recent recommendations, a shift away from the biomedical approach is likely to be needed, perhaps along similar lines to that seen in low back pain (25). Instead of the primary aim of the exercise program being to increase strength and stability, a biopsychosocial model would view exercise as a tool for pain relief, addressing unhelpful attitudes and beliefs, and understand social obstacles to recovery in order to increase personal control, improve function, and maintain independence.

In line with the exercise recommendations, there was consensus that exercise should be individualized, and that physical therapists should deliver advice about positive lifestyle changes. No consensus was reached as to whether exercise is most effective if delivered in a group setting or completed at home alone, which may in fact indicate overall agreement with the proposition that “group exercise and home exercise are equally effective and patient preference should be considered” (1).

There was consensus about the importance of exercise adherence, and interviews highlighted the link physical therapists made between adherence and clinical outcome, which is also shown in clinical trials (26, 27). Therapists perceived that levels of exercise adherence among patients with knee OA were low, which may be a reality because in some randomized controlled trials of exercise and knee OA, adherence rates have been as low as 30% (28), and can reduce over time (29). However, these relatively negative attitudes and beliefs about the likelihood of patients adhering to exercise programs may contribute to the uncertainty of the effectiveness of exercise for knee OA. The focus on patient-related barriers to exercise adherence, and the tendency to stereotype patients as adherers or non-adherers, may reflect limited knowledge and understanding of the complex nature of adherence behavior, including the role of the health care provider in optimizing adherence. Some interviewees identified gaps in their knowledge and skills surrounding behavior change. Providing further targeted training in these areas may improve the confidence and skills of physical therapists in facilitating patients to initiate and maintain an exercise program for knee OA.

Despite the perceived importance of exercise adherence, there was a strong belief among some physical therapists that adherence was the patient's own responsibility. This may reflect a paternalistic approach adopted toward exercise and knee OA, and may explain why, as shown in the interviews, instead of trying to understand the reasons for poor adherence and implement strategies to overcome this, some physical therapists may simply blame patients. The relatively short-term responsibilities of the therapist within this approach may help to explain why limited treatment sessions are provided for this patient group, with a lack of followup (8). This may result in insufficient supervision of exercise and insufficient support and encouragement for the patient, which may negatively impact exercise adherence (30, 31). Shifting the focus of care toward a more equal partnership between the patient and therapist that acknowledges the complex nature of adherence and includes a range of educational and behavioral adherence-enhancing strategies (32), for example pedometers (33), may help physical therapists optimize their potential role in supporting patients to maintain exercise activity.

Although the attitudes and beliefs of physical therapists may offer some explanation of the clinical behaviors reported in the national survey (8), the interviews also revealed that the context of the health care setting and the wider system plays an important part. The current context in the UK, which promotes prescription of self-directed exercise with no long-term support or followup, may have a negative influence on the exercise adherence behavior of patients (30), and contrasts with the current best practice guidance, which promotes timely and ongoing high-quality care to meet the continuing and changing needs of patients with long-term conditions (34).

This study utilized the benefits of mixed-methods research in adding richness and depth of understanding to the survey data alone. It has shown that there are some differences between UK physical therapists' attitudes and beliefs and the recent exercise recommendations regarding the benefits of local and general exercise for knee OA, and the potential role that physical therapists could play in facilitating exercise adherence. The biomedical, paternalistic approach adopted in general, and gaps in knowledge and skills regarding exercise adherence, may help explain why these differences exist and why exercise is currently delivered over relatively few treatment sessions with limited followup after discharge from physical therapy services (8). Providing further training and shifting toward a biopsychosocial, more patient-centered model of care may support physical therapists to change their practice in line with the recent exercise recommendations. However, it is clear that the wider system as a whole must also be targeted, and barriers such as lack of time, limited opportunities to provide followup, and poor links to community services must be addressed. Further research is needed to identify how best to facilitate these changes, as well as to address gaps in the literature to determine the appropriate prescription of exercise for patients with knee OA (for example, for patients with different severity of joint damage), in order to optimize the potential benefits of an exercise program.

Although using mixed methods can add depth and richness to data, due to the sequential design of the study, any sources of error in the survey may have been compounded in the interviews (35). Although the survey response rate was in keeping with other similar surveys of UK physical therapists (36), nonresponse bias may have been present, reducing the generalizability of findings to all of the physical therapists in the UK. In order to interview 24 physical therapists, we invited a total of 84. Although interviewees were purposefully selected to represent physical therapists with a range of personal characteristics and survey responses, they may have held different attitudes and beliefs than physical therapists that declined to be interviewed, thus reducing the transferability of findings. Finally, there are mixed results from other musculoskeletal pain conditions about the strength of association between practitioners' attitudes and beliefs and their behaviors (37). Although the attitudes and beliefs of physical therapists may offer some explanation of clinical behaviors reported in the national survey, this association has not been directly tested.

This study has highlighted key differences between the attitudes and beliefs of physical therapists within the UK, and the recent best practice recommendations regarding exercise and knee OA. An underlying biomedical model of care that views OA as wear and tear and a progressive degenerative condition may explain the uncertainty of physical therapists about the potential benefits of exercise for knee OA and the concerns about safety. The tendency is that a rather paternalistic approach is adopted, in which therapists emphasize the patient's responsibility in maintaining the exercise program over the long term rather than recognizing the potential role they could play in facilitating successful behavior change. A shift away from a biomedical and paternalistic approach and increased efforts to meet the training needs of physical therapists about the complex nature of exercise adherence may help. However, contextual influences such as the structures and processes of the health care system also need to be addressed to support meaningful shifts in care in line with recent recommendations, which identify exercise as a core treatment for patients with knee OA (6).


All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be submitted for publication. Ms Holden had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Holden, Nicholls, Young, Hay, Foster.

Acquisition of data. Holden, Nicholls, Young, Hay, Foster.

Analysis and interpretation of data. Holden, Nicholls, Young, Hay, Foster.


We thank the CSP for assistance with generating the sample and administering the survey; Dr. Elaine Thomas, Professor Mike Doherty, and Dr. Edward Roddy for ongoing support throughout the study; Janet Grime for independently analyzing 3 interview transcripts and sharing her interpretation of these; and finally, all of the physical therapists that participated in the study.