The clinical effectiveness of self-care interventions with an exercise component to manage knee conditions: A systematic review

Objective Treatment of knee conditions should include approaches to support self-care and exercise based interventions. The most effective way to combine self-care and exercise has however not been determined sufficiently. Therefore the aim was to evaluate the clinical effectiveness of self-care programmes with an exercise component for individuals with any type of knee conditions. Methods A keyword search of Medline, CINAHL, Amed, PsycInfo, Web of Science, and Cochrane databases was conducted up until January 2015. Two reviewers independently assessed manuscript eligibility against inclusion/exclusion criteria. Study quality was assessed using the Downs and Black quality assessment tool and the Cochrane Risk of Bias Tool. Data were extracted about self-care and exercise intervention type, control intervention, participants, length of follow-up, outcome measures, and main findings. Results From the 7392 studies identified through the keyword search the title and abstract of 5498 were screened. The full text manuscripts of 106 studies were retrieved to evaluate their eligibility. Twenty-one manuscripts met the inclusion/exclusion criteria. Conclusion The treatment potential of combined self-care and exercise interventions has not been maximised because of limitations in study design and failure to adequately define intervention content. Potentially the most beneficial self-care treatment components are training self-management skills, information delivery, and goal setting. Exercise treatment components could be strengthened by better attention to dose and progression. Modern technology to streamline delivery and support self-care should be considered. More emphasis is required on using self-care and exercise programmes for chronic condition prevention in addition to chronic condition management.


Introduction
Self-care is a concept widely applied across healthcare and can be broadly defined as "what people do for themselves to establish and maintain physical and emotional health and prevent or deal with minor illness, injury, or chronic conditions". This incorporates concepts such as exercise, hygiene, nutrition, medication, and environmental and socioeconomic factors [1,2]. Treatment techniques that have been incorporated into self-care programmes include: collaborative care plans between service users and healthcare professionals; setting goals that are reviewed and modified; helping individuals explore barriers to self-care; aiding people to monitor their symptoms and what action to take; providing advice and education; and coaching and peer support from other service users [3,4]. For musculoskeletal conditions, self-care programmes have been developed and evaluated for knee osteoarthritis, but their effectiveness is considered limited, due to methodological weaknesses in study designs [5,6].
Despite this, current evidence suggests that individuals with knee conditions should be given access to information about their condition and advice on self-management, especially exercise [6,7]. This poses certain challenges to healthcare professionals that deliver exercise based interventions. Firstly, there is evidence within physiotherapy that information provision and exercise are the most widely used treatment modalities for knee rehabilitation [8], but much of the self-care and exercise research has been carried out independent of each other. Therefore, the most successful approaches of combining self-care and exercise are not yet known; it is also not clear whether the same self-care techniques benefit all knee conditions. Secondly, exercise rehabilitation is an effective intervention across all knee conditions, ranging from ruptures of the anterior cruciate ligament to patellofemoral joint pain and postoperative care following knee surgery [7]. Techniques on exercise prescription and progression are however less well developed for individuals with a knee condition. Finally, when modern technology, such as the Internet, is used to deliver supported self-care programmes, it needs to be clarified how effective it is for individuals with knee conditions. Several systematic reviews and meta-analyses have been carried out to evaluate the effectiveness of self-care interventions (with or without exercise components), but these have been confined to patients with osteoarthritis [5,[9][10][11]. Many of these have also included other pathologies or were not knee specific [5]. One study evaluated self-management or exercise programmes but not necessarily studies that combined the two [11]. Overall these studies have concluded that the evidence was low to moderate quality and only demonstrated small benefit to patients. The current systematic review extends these themes to identify specific self-care and exercise approaches and outcome measurements that need to be developed to improve clinical effectiveness in the future. Therefore the aim was to evaluate the clinical effectiveness of self-care programmes with an exercise component for individuals with any type of knee condition. Clinical effectiveness was measured according to patient rated outcome measure scores.

Methods
The systematic review was carried out according to the PRISMA statement [12].

Data sources and search
The search strategy was designed by combining keywords from relevant published literature on self-care programmes and knee conditions [4,7]. The first category of keywords related to knee conditions: knee injury, knee surgery, knee joint, knee osteoarthritis, osteoarthritis, knee dislocation, knee replacement, knee arthroplasty, anterior cruciate ligament, medial collateral ligament patellofemoral pain, and knee pain. The second category of keywords related to self-care: self-care, self-efficacy, internet, social support, social networking, patient education, telemedicine, behaviour therapy, goal setting, self-groups, selfmonitoring, self-management education, and motivational interviewing. This keyword search was carried out on the following electronic databases: Medline, CINAHL, Amed, Embase, PsycInfo, Web of Science, and Cochrane Library Trials Register to search for manuscripts published up until January 2015 (no restriction was placed on the start date of the search). These searches were combined and limited to studies on adult humans and written in the English language. Additional articles were identified by reviewing the reference list of the retrieved manuscripts and checking the citations of all the full text manuscripts. An example of the search strategy is illustrated in Appendix A.

Study selection
Initially, all of the titles and abstracts were screened by reviewer 1 to check if they met the inclusion and exclusion criteria. To be included a study had to: • Have an adult population aged 18+ years • Be conducted in the out-patient/community setting • Use individuals with knee conditions • Be a randomised control trial • Evaluate clinical effectiveness using patient rated outcome measures as a primary or secondary outcome measure • Be an intervention that includes an exercise and self-care component • Be written in the English language.
Studies were excluded if they: • Evaluated economic effectiveness as the primary outcome • Evaluated in-patient care or housebound individuals • Compared lay versus professional intervention delivery • Targeted other family members/partners • Targeted treatment of other joints in the intervention or treatment to other health conditions in addition to the knee.
Full text manuscripts that met the inclusion/exclusion criteria were retrieved. For studies with a knee osteoarthritis population this could be unilateral or bilateral. Two reviewers checked the full text manuscripts against the inclusion/exclusion criteria, any disagreements were resolved through discussion to achieve consensus. The reference lists and citations of all full test articles were checked for relevant articles.

Quality assessment
The methodological quality of the remaining articles was independently rated by two reviewers using the Downs and Black quality assessment tool [13]. Any disagreements in the quality rating were resolved by discussion and reaching a consensus. The 'Risk of Bias' tool evaluated in depth any sources of manuscript bias [14].

Data extraction
Data were extracted about research design, self-care intervention, control intervention, participants, follow-up, main findings, and outcome measures. All data were extracted by reviewer 1 and then checked by reviewer 2, agreement was achieved by consensus.

Study characteristics
From the n = 7392 articles identified through the literature search n = 21 were included in the systematic review. Of these n = 18 were related to self-care interventions and three were long-term follow-ups of these interventions [15][16][17]. The PRISMA flow chart ( Fig. 1) identifies the process of how the final n = 21 articles were included. The numbers of studies evaluating each knee condition were: early osteoarthritis n = 3, mild/moderate osteoarthritis n = 2, severe osteoarthritis n = 1, osteoarthritis (unstated severity) n = 9, chronic knee pain n = 5, and anterior cruciate ligament reconstruction n = 1 (Table 1).

Study quality
The quality of the articles according to the Down and Blacks tool is displayed in Table 2. Seventeen manuscripts were classified as good and four as fair. The 'Risk of Bias' tool therefore provided a detailed assessment of bias (Table 3). In general there was a low or unclear risk of bias for random sequence generation, blinding of outcome assessment, incomplete outcome data, selective reporting, and other bias. As expected, there was a high risk of bias from lack of allocation concealment and blinding of participants and personnel. None of the studies scored well for blinding of participants due to the nature of the interventions. Overall, the evidence presented is from articles with a low or unclear risk of bias.

Duration and frequency of treatment
The frequency of treatment ranged from three home visits [25] or four clinic consultations [22,23] to three times weekly treatment over 12 months [18,19]. For treatment duration, nine of the studies had interventions that lasted a minimum of six weeks and required attendance at least once a week [15,[17][18][19][20][21]26,28,31,33]. The length of these sessions ranged from 20 min [27] to 2 1/2 h [20]. The length of the self-management programmes ranged from four weeks [27] to two years [18,19]. Specific guidance was generally given on the prescription of the self-care programme, but when an exercise programme was prescribed the number of repetitions, sets, and method of progression were not clearly stated. Only a limited number of studies provided specific reproducible guidelines on exercise prescription and these were interventions led by activity trainers and health educators [26,28,33,34] (Table 1).

Control groups
The most common control arm was an alternative intervention that either had a scaled down exercise component combined with self-care [18,19,24,25,28,32], or was a treatment that contained no self-care [23,26,27,33,34]. The other type of control used was a delayed start of the self-care intervention [20]. Several studies used a usual care control group, i.    they were not part of this research [15][16][17]21,22,31]. One study used a no-intervention control [35].

Outcomes
The outcome measures used are listed in Table 1. The WOMAC [37] was most frequently used [15,18,[20][21][22]24,26,27,30,[33][34][35]. This is a valid and reliable osteoarthritis specific patient rated outcome measure that assesses pain, stiffness, and function in patients with knee or hip osteoarthritis. The other frequently used valid and reliable condition specific tool was the AIMS 1/AIMS 2 [18,25,31,38,39], which measures physical, social, and emotional well-being. Three studies have specifically included an outcome measure to assess self-efficacy. This is an important variable to measure, because the aim of a self-care intervention is to promote self-efficacy [19,17,23,36]. The scales used are the Stanford Self-Efficacy Scale, scale K-SES [40], and Arthritis Self-Efficacy Scale [41]. Various other questionnaires and physical tests have been used, but there is no consistency on their use across studies.

Length of follow-up
All the studies carried out a baseline measurement, but several did no follow-up beyond completion of the intervention and were therefore not able to assess maintenance of long-term treatment effects. For those that did a long-term follow-up, there was no standardised time frame used and this varied from two to 12 months post-completion of the intervention (Table 1). For a trial using chronic conditions this may need to be at least six months to understand the course of the condition after treatment and to allow for any immediate improvement that may be due to the Hawthorne or placebo effect [42,43].

Findings
Nine studies measured outcome only up to the completion of the intervention, of these six were found to have a statistically significant improvement in outcome in the self-care intervention compared to the control group for WOMAC [15,19,22,35], SF-36 [19], AIMS/2 [19,31] and self-reported physical disability and performance test [28]. Three studies found no statistically significant difference in primary outcomes compared to the control on completion of the intervention [23,26,33] (Table 1).
Four studies that included long-term follow-up beyond the completion of intervention demonstrated statistically significant improvement in the self-care intervention group compared to the controls [20,27,34,36]. In six studies with long-term follow-up there was no statistically significant difference between the self-care and control groups [16,21,24,25,30,32]. In one study there was no significant improvement for the primary outcome but there was for exercise health beliefs [21]. Two studies demonstrated better outcome for self-care at completion of their intervention [15,31], but this benefit was not maintained at long-term follow-up [16,30].
The data was explored to evaluate if a sub-group analysis could be carried out using pain as an outcome measure across the studies. This evaluation indicated that there was too much heterogeneity (clinical diversity) between the studies based on long and short term clinical effectiveness for types of self-care and exercise interventions, professional delivering and type of control group. Therefore no meta-analysis has been carried out on this data.

Discussion
This systematic review evaluated the clinical effectiveness of selfcare and exercise programmes for individuals with knee conditions. This has been done following the guidance set out in the PRISMA statement. Overall study quality was good, but based on the risk of bias tool there was an 'unclear' risk of bias that introduced some weakness in the evidence presented. As expected none of the studies scored well for blinding of participants due to the nature of the interventions.
When self-management and exercise outcome had been assessed at the post-intervention time point, the majority of studies demonstrated that individuals in the self-care and exercise group had a better outcome than controls [15,18,19,22,28,31,35]. This benefit was not maintained in studies that had a longer time span for follow-up (beyond the intervention), as only four demonstrated a long-term benefit to the patient for self-care and exercise programmes [20,27,34,36]. This is important because it is the long-term success of an intervention that is important to patients and policy makers. Therefore, based on the findings of this review, there is conflicting evidence regarding the long-term effect of self-care and exercise interventions. It is recommended that in the future all studies include a long-term follow-up beyond completion of the intervention. Some of the differences in outcome between studies may be related to study design and this is discussed in the following sections. Table 2 Quality assessment using Downs and Black score for reviewers 1 (R1) and 2 (R2). Total quality score             One reason for inconsistent long-term outcome between studies may be related to the type of control group used. Three of the studies with a positive long-term outcome used a control group that was very different from the self-care intervention group, i.e., medication control, electrotherapy, delayed start, and usual care. In these studies it also appeared that individuals in the control group had less contact with the individual delivering the programme. For example, in the study by Yip et al. [17] patients in the control group received usual care, but this could have been little or no treatment and therefore no professional contact; unfortunately this was not defined in their publication. On the other hand, the studies that demonstrated no difference in outcome at the end of the intervention in the experimental group compared to the control group/s, often used treatment modalities within the different study arms that were common to the experimental and control arms. For example, in the study by Thomee et al. [23], both groups used the same pool of exercises but the self-management intervention had two sessions on self-management. In two other studies there was a distinct self-management and exercise group but both studies contained a further group that was a mixture of both exercise and self-care [26,33]. This may have made it more difficult to demonstrate a statistically significant difference between the groups because the interventions they received were not sufficiently different to have a treatment effect. These potential confounding factors could have reduced the internal validity of the studies. It is essential in future studies to ensure the treatment content of the control group is sufficiently different to that of the experimental group.
A further factor that explains the varied clinical effectiveness of the self-care programmes is the high level of heterogeneity in the study population, research design, and intervention content. The high level of heterogeneity is the main reason that a meta-analysis had not been carried out. This review specifically included self-care interventions with an exercise component for all knee conditions to evaluate effectiveness of the interventions for enabling recovery and chronic condition prevention, as well as chronic condition management. Despite this, the majority of studies had been carried out using either an osteoarthritis or chronic knee pain population. In addition, within these populations there is a high degree of heterogeneity, especially for osteoarthritis, where it is recognised that there are different stages to the disease [44]. This means that an early osteoarthritis population is not automatically comparable to late stage osteoarthritis and heterogeneity within the population should be taken into consideration when designing the study. Adopting standardised criteria for the diagnosis of knee osteoarthritis such as that recommended by the American College of Rheumatologists [45,46] has not routinely been used but may improve study quality and assessment of participant heterogeneity in the future.
It is evident that there is a wide range of treatment approaches available to support self-care and it needs to be established which components are most beneficial and what is the most effective and efficient manner of delivery. This systematic review can provide some insight into this by analysing in detail the group of studies that had a positive long-term effect beyond completion of the intervention. All of these studies had a well-defined information provision component, but the exercise component was either part of the information and education [20,34], or delivered as a practical exercise group [17,27]. Generally, there was insufficient detail on exercise prescription to be reproducible. One of these studies took self-management beyond information provision and also focused on other practical steps to develop selfmanagement skills, such as goal setting and self-efficacy [20]. Recommendations based on this group of studies are that information provision is an essential component, but the best mode of delivery and content could not be specified. Likewise, exercise had been delivered in several ways across these studies, either through practical groups or as part of structured programmes that focus on the development of self-management skills. The optimal method is however yet to be established. Three of these studies did use the same theoretical framework [17,20,34], i.e., social and cognitive theory [47], to underpin their self-care approach, but there is no gold standard as to what selfmanagement techniques this should include. In addition there is not an underlying framework that cohesively brings together the self-care and exercise components. What is reassuring is that these programmes were delivered with relatively few contacts with a healthcare professional (four to six contacts) and therefore had relatively low use of healthcare resources. This does demonstrate the potential for selfmanagement and exercise programmes to be delivered independently to clinic visits.
A range of healthcare and exercise professionals were involved in delivering treatment across these studies. The studies with the best outcome were not delivered by one specific professional group, therefore which professional delivers the intervention does not appear to be a factor influencing clinical effectiveness. What seems essential is that adequate training is available to ensure that the individual has the skills to ensure delivery on both the self-management and exercise components. Of note, the studies that provided most detail on exercise prescription and progression were delivered by activity trainers and  SM-exs: Small groups. Holistic approach addressing: OA, self-management skills (goal setting, problem solving, modelling, positive thinking and improving self-efficacy), medication, fitness and exercise, joint protection, nutrition, fall prevention, environmental risks, and coping with negative emotions (guided imagery, cognitive behavioural therapy). Printed information.
CONT: Delayed start. SM: 6 weeks Group session once a week, 2.5 h

Healthcare professional delivered
Ettinger et al. [28] SM-Aerobic exercise (SM-aerobic): 3 months supervised group walking programme on indoor track. 15 months home-based walk programme: Transition phase (months 3-6) 4 home visits to develop walk programme in home environment and 6 telephone calls with leader. Maintenance Phase: (months 6-9) telephone contacted by telephone every 3 weeks (months 6-9), then monthly 10-18 months. Aerobic session: combined stretching, walking phase with callisthenics walk at 50 -70% heart rate. SM-resistance exercise (SM-resistance): 3 months facility based classes, trained leader, 15 months home-based programme, same contacts as other interventions. There were 2 orientation classes (1 h), 9 exercises, 2 sets 9 repetitions, 3 days per week for 18 months. General muscular fitness designed to strengthen major muscle groups, upper and lower limbs. Resistance progressed in stepwise manner as long as participants could complete 2 sets of 10 reps. During home phase weights exchanged at request, after face-to-face contact or telephone follow-up.   To improve programme quality, exercise prescription needs to be incorporated into future interventions and be embraced by other professional groups that frequently deliver exercise programmes within healthcare settings. Individuals with knee osteoarthritis, who acknowledge the importance of exercise in their management, have reported concerns over how this should be done long-term [48]. Taking into consideration the heterogeneity in populations and different treatment components within the interventions, there is unlikely to be one type of self-management and exercise combination that suits all. There is therefore a need to establish which treatment components are likely to deliver maximum benefit to individual presentations, and future research needs to be directed at phenotyping or stratifying patients accordingly to provide these answers [44]. In addition further exploratory research is required to understand the patient's perspective on the most effective self-care components and if current methods of delivering exercise meets their needs. This fits in the MRC framework for researching complex interventions [49].
The use of modern technology has not been reported by any of these studies, but using the Internet to deliver self-care programmes has been identified as an approach that warrants further research for healthcare delivery [50]. For example, online approaches could provide virtual contacts through tele-rehabilitation, better access to up to date information, virtual support groups, notifications and prompts, self-monitoring, and tracking. Several studies have recently been developed in this field, but did not match the inclusion criteria of this systematic review [51,52]. In addition, technology could facilitate new self-care approaches that incorporate factors that individuals with long-term conditions identify as important. This includes: better support that acknowledges the physical and emotional hard work of self-care, facilitates ongoing care, does not isolate healthcare to one time point, supports individualised strategies, promotes self-efficacy, does not trivialise the condition, provides encouragement and endorsement from clinicians and introduces help to younger populations [48,[53][54][55].
The patient rated outcome measure most frequently used across the studies was the WOMAC, because this is an osteoarthritis specific tool which measures patient rated changes in symptoms and function, which are generally considered to be the outcomes for arthritis research [10]. This outcome measure is not transferable to other knee pathology. Interestingly, only two studies included a measure of self-efficacy, this is surprising as this is what self-care interventions target and therefore ability to self-care would be expected to improve when symptoms may not [56]. These scales [40,41] have not undergone full psychometric testing and tend to be condition specific so they do not translate across all knee conditions. Inclusion of a self-efficacy or empowerment outcome measure needs to be considered in future research to allow comparison across studies. As yet no gold standard for use across patient groups can be recommended [56].

Conclusion
The studies included in this review demonstrated an 'unclear' risk of bias and conflicting evidence regarding the long-term effect of self-care and exercise interventions. Nine of the included studies failed to have a long-term follow-up, which threatens the external validity of their findings. The four studies that did demonstrate long-term clinical effectiveness all used an OA population and had a strong focus on information provision, goal setting, and developing self-management skills. The exercise component of these interventions was poorly developed and could be strengthened by improving the exercise content, prescription, and progression. The evidence on exercise prescription needs to have a higher priority alongside self-care interventions. Further research on how to combine and integrate the self-care and exercise components is required and using better designed studies on other knee conditions. This could be achieved using modern technology which to date has been underutilized in this field. Alongside this, there is a need to ensure that all healthcare professionals working in a rehabilitation environment have the skills to deliver on both the self-care and exercise treatment components. Greater integration of outcomes that measure patient ability to self-manage is required. Little evidence exists on the combined use of self-care and exercise interventions for prevention of chronic knee conditions, which needs to be addressed in the future.
Abbreviations AIMS 1/AIMS 2 Arthritis Impact Measurement Scale K-SES Knee Self-Efficacy Scale WOMAC Western Ontario and McMaster Universities Arthritis Index Table abbreviations ACR American College of Rheumatologists Combined combined intervention of self-management, exercise and control treatment CONT control Exs-only exercises intervention only OA osteoarthritis SM-aerobic exs self-management and aerobic exercise intervention SM-exs self-management and exercise intervention SM-home exs self-management and home exercise intervention SM-home + tel self-management, home exercise and telephone intervention SM-only self-management intervention only SM-resistance exs self-management and resistance exercise intervention SM-tel self-management and telephone intervention VAS visual analogue scale