Development of the recommendations
After the first meeting, a total of 168 propositions were suggested by the experts. Propositions that were identical were merged and propositions containing one word only were excluded. The second Delphi round comprised 140 propositions, with topics being very broad and including far more non-pharmacological interventions than currently included in these recommendations. After five Delphi rounds, consensus on 11 recommendations was achieved, which are presented with complete formulation in table 2 with the accompanying level of evidence (LOE) and LOA. The 11 recommendations are ordered in a logical sequence or procedural and chronological hierarchy rather than by any considered importance.
EULAR recommendations for the non-pharmacological core management of hip and knee OA, with levels of evidence (LOE) and level of agreement (LOA). The propositions are ordered by topic
The terms ‘non-pharmacological’ and ‘non-surgical’ management were discussed by the expert group. The terms were considered to be negative owing to their prefix ‘non’ and were therefore not considered optimal; finding a new terminology was included in the research agenda (table 3). In addition, research evidence specifically for hip OA was sparse and, in general, recommendations for the management of people with hip OA were derived largely from trials including people with both hip and knee OA or with knee OA only.
Research and educational agenda for non-pharmacological management of hip and knee OA
Research data on how a comprehensive assessment of people with hip or knee OA should best be carried out are scarce. Since initial assessment will always be a part of the management in any person with hip or knee OA, controlled trials evaluating assessment will have difficulties in selecting the most appropriate comparator. One randomised, controlled trial (RCT) comparing a comprehensive assessment and management approach with usual care showed no difference in pain or physical function.13 However, in that study, both approaches included initial assessments, but with different content and were executed by different professionals.13
The group considered a comprehensive initial assessment to be a prerequisite for the individualised management strategy described in recommendation 2. The recommendation on the initial assessment included the following elements: the person’s physical status, activities of daily living, participation, mood and health education needs, health beliefs and motivation to self-manage. In the absence of evidence from studies on the effectiveness of various forms of assessment, the group based the recommended content of the initial assessment on the main areas of disease consequences, including potentially interacting personal and environmental factors described in the literature.14–22 Evaluation of cardiovascular disease, people’s expectations and self-efficacy were also discussed as important aspects in a biopsychosocial approach.14 ,17 Moreover, the group found that a comprehensive assessment, which is applicable to the initial consultation, should also be repeated during regular follow-up of the person.
The task force agreed unanimously that the overarching principle for treatment of a person with hip or knee OA should be individualised, which is in line with previous guidelines.7–9 ,23 Individualised treatment does not imply that every treatment should be individually provided, it means rather that treatment is personalised, or tailored. RCTs on individualised non-pharmacological management are scant. The available studies showed reduced pain (mean difference, 95% CI (0–20 point scale): −1.19, −2.1 to −0.3 and −1.10, −1.84 to −0.19; and (0–100 scale): −17.0, −23.6 to −10.4) and improved physical function (mean difference, 95% CI (0–68 point scale): 3.65, 1.0 to 6.3 and 3.33, 0.78 to 5.88) compared with usual care,24–26 but not compared with group-based rehabilitation25 ,27 ,28 or information on healthy lifestyle.29 ,30 Follow-ups at 9, 18 or 30 months showed no effect on pain.31 ,32
As the data underpinning this recommendation are limited the factors to be considered for the tailoring of management were mainly based on prognostic factors shown in the literature. An important and modifiable risk factor for knee OA is weight,20 ,33 ,34 implying individualised targeting at weight reduction in people who are overweight or obese.
Moreover, individualised treatment being the standard of care in OA and chronic disease in general7 ,35 ,36 was considered to imply informed, shared decision-making, taking into account the person’s wishes and preferences. The group noted that with the conduct of an RCT to study the impact of individualisation, the patient’s view cannot be wholly taken into account and that some element of individualisation will always be incorporated in any treatment. To better understand individualised treatment, the group found that future research should focus on factors that affect outcome—that is, moderators, not individualisation as such.
Comprehensive package of care
This recommendation deals with the provision of an integrated package of care rather than single treatments alone or in succession. The group recommended five core interventions to be considered comprehensively in every patient with hip or knee OA. The recommendation specifically implies that a person with hip or knee OA should receive education about her/his condition (3a), and be managed accordingly (3b–e).
With the exception of walking aids and assistive technology and dealing with adverse mechanical factors, the literature supports the delivery of combined interventions including information and education, exercise and/or weight reduction.
In people with hip and/or knee OA the combination of patient education or self-management intervention plus exercise was found to have a significant effect on pain, but a less marked effect on function.26 ,31 ,37–40 In people with hip OA the effect of such combinations was mainly seen on function (0–100 point scale) at 3 and 6 months after intervention (mean difference, 95% CI −7.5, −13.9 to −1.0; and −8.4, −15.1 to −1.7).41 ,42 In people with knee OA effects on pain and/or function were seen in eight studies,24 ,25 ,43–48 whereas no effect was seen in four studies.32 ,49–51 The addition of advice from a dietician for overweight or obese patients to the combination of patient education or self-management intervention plus exercise was found to improve both pain and function in patients with hip or knee OA.52–55
Principles of lifestyle changes
Recommendation 4 deals with key elements of the delivery of interventions aimed to initiate and maintain lifestyle changes. It is known that behavioural changes are difficult to achieve and maintain, and the effect of advice and counselling by healthcare providers is disappointing.56 The literature search for this recommendation was limited to lifestyle changes considered most relevant for hip and knee OA—that is, exercise and weight loss.
The common feature in the trials supporting this recommendation was to teach and encourage behavioural change strategies through goal setting of physical activity and weight changes, action plans to maintain changes and regular follow-up over at least 1 year to re-evaluate and discuss goals and action plans.28 ,39 ,40 ,53 ,57–62
Reports examining the effectiveness of specific elements to be included in interventions aiming to change behaviour are scarce. The literature suggests that the following factors improve adherence to exercise or physical activity: individual exercise, graded activity, individualisation according to the person’s exercise goals, feedback on progress made towards the goals, iterative problem solving with emphasis on skills that will improve adherence, reinforcements of maintaining exercise such as additional motivational programmes, exercise plans and log books, written information and audiotape or videotape, and booster sessions.28 ,39 ,40 ,61–63 In addition, some studies found an effect on pain39 ,40 or function59 from lifestyle interventions that integrate such elements. A systematic review including a mixed population of people with OA and/or rheumatoid arthritis found effect sizes of 0.21 (95% CI 0.08 to 0.34) for pain and 0.69 (95% CI 0.49 to 0.88) for increased physical activity from lifestyle interventions aiming at increasing physical activity.64 Over 40% of the included lifestyle interventions prompted problem solving, self-monitoring, goal setting and regular feedback.64
For people with knee OA or knee pain, improvements were seen in pain, function and weight loss from diet interventions that included individual weight-loss goals, problem solving on how to reach these goals and follow-up visits to re-evaluate and discuss goals in combination with exercise.53 ,60 In obese patients, weight-loss programmes with explicit weight-loss goals showed a higher mean change in weight than programmes without explicit goals.65 This indicates that the elements in recommendation 4 are important for the change and long-term maintenance of behaviour. The group discussed the importance of regular follow-up that includes feedback on the progress towards explicit goals and extends over a long time to achieve long-term effects of a healthy lifestyle.
Principles of information and education
Recommendation 5 is concerned with the content and method of delivery of various forms of educational programmes to best benefit the person with hip or knee OA. It is grounded in the general recognition that appropriate information and education are indispensable in prompting adequate self-management in chronic diseases. The recommendation is underpinned by the majority of studies on education interventions provided to patients with hip and/or knee OA. In general, small, but statistically significant effect sizes on pain (0.06, 95% CI 0.02 to 0.10) and physical function (0.06, 95% CI 0.02 to 0.10) have been reported from attending education or self-management programmes.6 ,66 Lower costs of community-based care and medication up to 12 months has been achieved from attending a combined self-management and exercise programme, and a reduced number of medical consultations from attending self-management programmes in patients with hip and/or knee OA have been reported.32 ,67 ,68
The literature review included trials that compared education or self-management programmes with usual care, attention controls or no intervention. These trials described one or several elements from 5c to f (table 2) in their interventions.69–85 The literature did not support the additional value of spouse-assisted coping skills training,79 and no trials were found for individualisation according to illness perception and educational capability, or for inclusion of education in every aspect of management. The group, however, considered the inclusion of spouses in the intervention to be a question of individualisation and appropriate in some cases. One systematic review found that, in people with OA, effective self-management interventions followed a protocol, included elements of cognitive behavioural theory or social cognitive theory and were led by trained health professionals.86 These elements are not specifically dealt with in the recommendation, yet they were supported by the group.
Principles of exercise education
Recommendation 6 deals with the principles of the delivery of education about exercise and physical activity. There is convincing evidence for the overall effectiveness of exercise on pain (ES, 95% CI: 0.40, 0.30 to 0.50) and function (ES, 95% CI: 0.37, 0.25 to 0.49) in people with knee OA,87 and to a lesser extent in people with hip OA (ES, 95% CI, pain 0.38, 0.08 to 0.68).88
Few studies have directly compared different exercise ‘dosage’ (frequency, intensity and duration) and progression approaches in people with OA.87 ,89 ,90 One RCT reported reduced pain from attending a progressive functional strengthening programme compared with a non-progressive programme in people with knee OA,90 but two trials could not show any differences from attending various intensity levels of aerobic or resistance-exercise programmes.89 ,91 Hence, the optimal exercise ‘dosage’ and rate of progression remain uncertain.
In patients with knee OA different delivery modes (individual, group-based or home programmes) have all been shown to effectively reduce pain (individual, ES, 95% CI 0.55, 0.29 to 0.81; group-based, ES, 95% CI 0.37, 0.24 to 0.51; and, home, ES, 95% CI 0.28, 0.16 to 0.39) and improve function (individual, ES, 95% CI 0.52, 0.19 to 0.86; group-based, ES, 95% CI 0.35, 0.19 to 0.50; and, home, ES, 95% CI 0.28, 0.17 to 0.38) compared with education, telephone calls, waiting list, relaxation, ultrasound, hot-packs or no treatment.87 In patients with hip and/or knee OA, water-based exercise was found to significantly reduce pain (ES, 95% CI 0.19, 0.04 to 0.35) and improve function (ES, 95% CI 0.26, 0.11 to 0.42) compared with education, telephone calls or no intervention.92 Home-based exercise was found to be as effective as water-based exercise in one small RCT in people with hip OA.93 Water-based exercise can include swimming and/or different types of exercise programmes. Since the different modes of delivery are equally effective, the person’s preference, findings of the initial assessment and local availability should determine the choice of mode of delivery in clinical practice.
The literature suggests that pacing of activity and/or integrating physical activity into daily living as part of a comprehensive exercise regimen is more effective in people with hip or knee OA or with knee pain than usual care or written information, but not compared with standardised exercise or a pharmacy review.24–26 ,29 ,31 ,38 ,46 ,57 ,58 ,77–79 ,94
This recommendation suggests the need for an increase in the intensity and/or duration of exercise over time. This is based on the literature, where most strength training exercise programmes evaluated in people with knee OA included dynamic exercises with progression over time.95 Moreover, in one study comparing progressive and non-progressive approaches in people with knee OA, the former was found to reduce pain more effectively.90 General recommendations for dosage and progression of exercise in older people and people with chronic disease are aerobic moderate-intensity training for at least 30 min/day or up to 60 min for greater benefit, and progressive strength training involving the major muscle groups at least 2 days/week at a level of moderate to vigorous intensity (60–80% of one repetition maximum) for 8–12 repetitions.96 ,97 These recommendations emphasise that in people with chronic disease who do not reach the recommended level, they should be as physically active as their abilities and condition allow.97
Before considering the evidence for specific exercises in hip and knee OA, it should be noted that although exercise has been shown to reduce pain in patients with hip OA,88 overall there is a lack of information to support treatment effects of exercise in hip OA.8 ,88 ,98–103 The LOE for the recommendation of different types of exercise in people with hip OA therefore could not be graded. For knee OA, however, high-quality research evidence has reported that exercise reduces pain and improves physical function.6 ,87 ,104 Results for the effect of exercise on quality of life are inconsistent.90 ,92 ,95 ,99 ,102 ,104 ,105
Research about strengthening exercises in knee OA shows that both specific quadriceps strengthening exercises or strength training for the lower limb reduce pain effectively (ES, 95% CI 0.29, 0.06 to 0.51 and 0.53, 0.27 to 0.79, respectively) and improve physical function (ES, 95% CI 0.24, 0.06 to 0.42 and 0.58, 0.27 to 0.88, respectively).87 The literature on strength training in people with knee OA in most cases describes dynamic exercises, whereas research on isometric exercises is sparse.95 Hip strengthening exercises have been poorly evaluated in people with hip OA.103 However, in people with medial tibiofemoral knee OA, hip strengthening exercises reduced knee pain and improved physical function.106
Aerobic training (walking) is effective in reducing pain (ES, 95% CI 0.48, 0.13 to 0.43) and improving physical function (ES, 95% CI 0.35, 0.11 to 0.58) in patients with knee OA.87
The evidence for mixed exercise programmes, including strengthening, aerobic and flexibility components, in patients with knee OA is conflicting.107 ,108 One type of exercise has not been shown to be better than another (strength, aerobic or mixed exercises).87 ,107 ,108
The group reached consensus that mixed programmes should be recommended. However, it was noted that with mixed programmes the minimal requirements to improve or maintain muscle strength, aerobic capacity and/or joint range of motion need to be met,97 as some reports suggest that mixed programmes may be less effective than focused programmes.108
This recommendation states that initial instruction is required, but that in the longer term the person should integrate exercise into daily life. This part of the recommendation is substantiated by studies showing that the number of supervised sessions influences outcome in people with knee OA.87 Twelve or more directly supervised sessions have been shown to be more effective than a smaller number on pain (ES 0.46, 95% CI 0.32 to 0.60 vs ES 0.28, 95% CI 0.16 to 0.40, p=0.03) and physical function (ES 0.45, 95% CI 0.29 to 0.62 vs ES 0.23, 95% CI 0.09 to 0.37, p=0.02).87
In addition, it was noted that research evidence is growing for tai chi and yoga. Though not included in the literature review, tai chi has been found to be effective for the reduction of pain in patients with hip or knee OA, with ES ranging from 0.28 to 1.67.108
Education on weight loss
In recommendation 8, the principles of education about weight management are included. The recommendation is mainly supported by the literature in knee OA, as no evidence to support the effect of weight loss in patients with hip OA is available. However, being overweight or obese has been shown to be associated with hip OA (OR=1.11, 95% CI 1.07 to 1.16).33
In patients with knee OA, the effectiveness of weight-loss programmes on body weight, pain and/or physical function was demonstrated in programmes delivered as weekly supervised sessions for a range of 8 weeks to 2 years.54 ,60 ,109–113 The effects on pain, function and weight loss from attending weight-loss programmes were small but significant (ES, 95% CI, pain 0.20, 0.00 to 0.39; physical function 0.23, 0.04 to 0.42; mean weight loss, 95% CI, 6.1 kg, 4.7 to 7.6).109 The interventions included strategies on how to reduce calorie intake by meal plans, reduce fat and sugar, reduce portion size, meal replacements, and comprised behavioural modifications, self-monitoring, weight-loss goals and maintaining body weight in participants who had reached their goals and/or exercises for some of them.54 ,60 ,109–112 Overall, the evidence from RCTs for the maintenance of achieved weight loss after the interventions have ended is absent in people with hip and knee OA.
In general, in overweight or obese populations, healthy eating, limiting fat and salt intake, eating at least five portions of fruit and vegetables a day, being physically active for at least 30 min/day and elements such as self-monitoring, explicit weight-loss goals, and motivational interviewing have all been suggested to promote weight loss and that regular follow-up over 4 years helps in maintenance of the weight loss.65 ,114–118 Weight-loss programmes in older obese people that included explicit weight-loss goals showed mean changes in weight of −4.0 kg (95% CI −7.3 to −0.7), which was significantly more than programmes without explicit weight-loss goals (mean change, 95% CI, −1.3 kg, −2.9 to 0.3).65 To achieve a structured meal plan with a balanced combinations of low calorie and sufficient vitamin and mineral intake, meal replacement bars or powders can be an addition to healthy eating.54 ,60 ,109 ,110 Though not included in the literature review, it has been suggested that bariatric surgery should be part of comprehensive weight management in people with hip or knee OA who are morbidly obese, and could help reduce weight and joint pain.119 ,120
Although research evidence is scant, the group was unanimous in its view that the use of appropriate footwear should be recommended in patients with hip or knee OA. Shoes may help through different mechanisms, such as acting as shock absorbers or controlling foot pronation.121 ,122 Appropriate shoes implies no raised heel, thick, shock-absorbing soles, support for the arches of the foot and a shoe size big enough to give a comfortable space for the toes.121–123
In patients with hip OA there is no evidence to support the effect of specific shoes or insoles on pain or function. In patients with knee OA, the use of shoes with shock-absorbing insoles for 1 month reduced pain and improved physical function in a pre–post test design.124 No differences in knee pain from the use of specialised shoes (unstable Masai technology shoe or variable-stiffness shoe) compared with conventional athletic shoes have been seen, but reduced pain was seen in both groups over time.125 ,126 In addition, decreased knee joint loads were found when specialised mobility shoes were used.121
The literature on the effectiveness of the use of lateral wedged insoles in patients with medial knee OA found no significant effect on pain or function.121 ,127 ,128 There is no support for whether one type of insole would be better than another,129 and adverse effects including foot-sole pain, low-back pain and popliteal pain have been reported.121 ,128 ,129 In light of evidence for no clinical effects of the use of lateral wedged insoles and the report of adverse effects, the group rejected the recommendation (table 2, 9b).
Assistive technology and adaptations at home and/or at work
The frequent use of assistive technology and the high satisfaction rates with its use indicate that walking aids, assistive technology and adaptations are important and useful for people with hip or knee OA.130–133 There are, however, no clinical trials to substantiate elements in this proposition, except for the use of a cane in patients with knee OA.134 However, the group was unanimous in its view that in all patients with hip or knee OA walking aids, assistive technology and adaptations at home and/or at work should be considered systematically and recurrently. The group noted that the value of some of these interventions is so obvious and has an immediate effect in individual cases that further research into the effectiveness of specific devices or adaptations can hardly be expected. Cross-sectional studies show that walking aids, assistive technology and adaptations at home and/or work are important and often used by people with hip or knee OA. Most people with severe hip (63%) or knee pain (90%) reported the use of walking aids.130 ,131 In people with arthritis, a mean of 9.9–10.8 devices has been reported to be in use and the satisfaction rate for all categories of device was more than 87%.132 Moreover, unmet needs for new assistive technology to help perform activities that individuals could not do were identified.132 Having access to a walking aid or other assistive technologies can be a help and provide security for individuals with constant or fluctuating symptoms. The group found that future observational studies on the use, satisfaction from and suggestions for new technology or improvements of existing technology are needed.
Management of work ability
Recommendation 11 deals with the effectiveness of work-related interventions. The proportion of employed people who have work disability due to OA is substantial. Although there are known occupational risk factors for knee OA and its progression—for example, heavy work, knee squatting or bending, lifting and specific sports,18 there are no studies to support the effect of vocational rehabilitation on pain, physical function or quality of life specifically in patients with hip or knee OA. One study in patients with peripheral OA found that a specialist-run, protocol-based early intervention significantly reduced the number of days of sick leave compared with standard primary care.135 The intervention was administered by a rheumatologist and comprised three main elements: education, protocol-based clinical management and administrative duties. The educational part included information about the condition, reassurance that serious disease was not present, self-management, exercises, ergonomic care, booklets, optimal level of physical activity and early return to work. Descriptive studies have found that environmental factors, such as having access to public transport or a car for mobility outside home are facilitators and that the absence of these is associated with limitations to daily activity.136 ,137 Some elements in this recommendation may have to be adapted to the country in which they are executed, since availability and accessibility of services in the healthcare and social security system may vary greatly. The group concluded that there is a clear paucity of research evidence for work-related interventions in people with hip and knee OA.