Coping strategies and self-efficacy as predictors of outcome in osteoarthritis: a systematic review
Abstract
Objective: Systematically to review the literature, investigating the prognostic value of self-efficacy and coping strategies used by adults with osteoarthritis (OA) recruited within the community or primary care.
Methods: An online electronic search was performed from inception to August 2009, using EMBASE, CINAHL, PsycINFO and MEDLINE databases. A search of keywords and key authors was performed to find related articles, and experts in the field were contacted to identify additional literature. Three reviewers blindly assessed the quality of the included studies, using pre-determined criteria. Data on coping strategies and self-efficacy were extracted and tabulated.
Results: Eight studies were identified and included in this review. Six of the papers were rated as being of acceptable methodological quality. Strong evidence was identified for the role of self-efficacy in predicting disability, but no evidence was found for the relationship between self-efficacy and pain. Although problem-solving coping had no effect on pain, there was weak evidence that active coping strategies predict increased pain and worsened mood. There was also weak evidence demonstrating that problem avoidance, wishful thinking, social withdrawal, self-criticism and turning to religion are predictors of depressed mood. This review also presents evidence for catastrophizing, self-efficacy and passive coping strategies as predictors of mood.
Conclusion: Coping strategies and self-efficacy are important prognostic factors for people with OA. This review highlights the need for further research to ascertain the predictive values of coping strategies that, to date, have been less well investigated. This may, in turn, result in a better understanding of the role of coping in OA and enable clinicians and patients to manage the condition more effectively. Copyright © 2010 John Wiley & Sons, Ltd.
Introduction
Osteoarthritis (OA) is a common progressive disease, affecting approximately 8.5 million people in the UK (Arthritis Care, 2004). While pain and disability are the most commonly reported symptoms, sufferers may also experience high levels of depression, anxiety and fatigue (Somers et al., 2009a). It is thought that, although the degree of degeneration demonstrated on X-ray may correlate to some extent with the disease state, there is some discrepancy between the severity of X-ray findings and the actual pain and disability experienced by the patient (Dijk et al., 2006). Many factors are thought to be associated with this variance, including age (Creamer et al. 1999; Dijk, et al., 2006; Wright et al., 2009), body mass index (Dijk et al., 2006; Juhakoski et al., 2008), muscle strength (Dijk et al., 2006), aerobic exercise (Dijk et al., 2006) and symptom duration (Creamer et al., 1999; Juhakoski et al., 2008). Recently, there has been increasing interest in the role of psychological factors in the association between pain and disability in OA and the capacity of interventions to modify these to improve outcome (National Collaborating Centre for Chronic Conditions, 2008).
An abundance of studies have established a cross-sectional relationship between coping strategies used by people who have OA, and pain, disability and mood (Gaines et al., 2002; Hopman-Rock et al., 1998; Jones et al., 2008; Keefe et al., 1997, 2000; Maly et al., 2005, 2006; Marks, 2007; Perrot et al., 2008; Prior and Bond, 2004; Rapp et al., 2000; Sale et al., 2008; Somers et al., 2009b; Tak and Laffrey, 2003; Tsai et al., 2008; Van Baar et al., 1998). For example, Perrot et al. (2008) demonstrated that passive coping strategies – defined as the surrendering of control and allowing external factors to influence outcome (Brown and Nicassio, 1987) – such as worrying, resting and retreating were associated with higher levels of pain and greater functional impairment. Catastrophizing – defined as focusing on the magnitude of the pain and feeling helpless (Somers et al., 2009b) – has also been shown to be related to pain and disability in people with OA (Keefe et al., 2000; Rapp et al., 2000; Somers et al., 2009b). For example, Keefe et al. (2000) reported that women experience higher levels of pain than men. However, once catastrophizing was adjusted for, the difference in gender-specific pain scores was statistically insignificant; thus, the higher levels of pain were related to catastrophizing. In addition, Rapp et al. (2000) found that higher levels of catastrophizing were related to both higher pain intensity and lower physical ability. While these cross-sectional studies are useful in demonstrating an association between specific coping strategies and OA, they do not enable prediction of outcome. It is, therefore, important to consider alternative methodological designs such as prospective cohort studies to establish whether specific coping strategies are predictive of outcome in people with OA.
To date, the role of psychological variables in back pain patients has been investigated more thoroughly than in OA patients, and a number of reviews on this topic have been published (Linton, 2000; Pincus et al., 2002). These studies reported that catastrophizing, passive coping strategies, self-perceived poor health and somatization were related to pain and/or disability. However, to date there has not been a systematic review of prospective cohort studies investigating the predictive role of coping strategies in people with OA. The aim of this systematic review was to evaluate the available literature on the role of coping strategies among adults with OA in primary care and community settings.
Materials and methods
Literature search strategy
A systematic search was performed using selected electronic databases (EMBASE, PsycINFO, CINAHL and MEDLINE) from inception to August 2009. The general search terms used were: pain, arthritis, arthralgia, osteoarthritis, adaptation, psychological, self-efficacy, fear avoidance, catastrophizing, coping, adaptive behaviour, longitudinal and prospective cohort. Full details of the search are available on request from the first author. The search strategy used a combination of MESH/Thesaurus terms and keywords. Manual searches were also carried out using authors' names from relevant cross-sectional studies, in order to find related cohort studies. Key authors and experts in the field were also contacted for additional unpublished papers.
Inclusion criteria
This review included longitudinal cohort studies investigating the coping strategies used by adults with OA in a primary care or community setting. Only studies presented in the English language were included. Titles and abstracts were screened and studies were excluded if they did not meet the inclusion criteria (Figure 1).
Result of systematic search and selection of studies
Quality assessment
In order to assess the methodological quality of the included articles, a combination of two quality assessment tools (Hayden et al., 2006; Mallen et al., 2006) were used (Table 1) and modified by the first author to meet the specific needs of this review. This 17-item checklist includes details of sample size, length and completeness of follow-up, external validity, method of analysis, the study limitations and adjustment for confounding factors. Each criterion was given equal weighting, scoring 1 point when the response to a question was ‘yes’, 0 when the response was ‘no’ and 0 when the response was uncertain. All relevant papers were checked against the quality criteria by two researchers, with an agreement rate of >85%; when disagreements occurred, the additional input of a third researcher was introduced.
| Criteria | Blalock et al., 1995 | Hampson et al., 1996 | Affleck et al., 1999* | Rejeski et al., 2001 | Steultjens et al., 2001 | Sharma et al., 2003 | Keefe et al., 2004 | Allen et al., 2006* |
|---|---|---|---|---|---|---|---|---|
| Is the population clearly defined? | Y | Y | N | N | Y | Y | Y | ? |
| Is the study design clearly defined? | Y | Y | Y | Y | Y | Y | Y | Y |
| Is the selection process of participants described? | Y | Y | N | N | N | Y | Y | Y |
| Are the objectives of the study defined? | Y | Y | Y | Y | Y | Y | Y | Y |
| Is the sample size appropriate? | Y | ? | Y | Y | Y | Y | ? | N |
| Is the sample representative of the target population? | ? | ? | ? | ? | ? | ? | N | N |
| Is completeness of follow-up adequate? | Y | Y | Y | Y | ? | Y | Y | N |
| Is completeness of follow-up described? | Y | N | ? | Y | N | Y | Y | Y |
| Is the follow-up time of appropriate length? | Y | Y | Y | Y | Y | Y | Y | Y |
| Are the outcomes clearly defined? | Y | Y | Y | Y | Y | Y | Y | ? |
| Are the outcomes measured appropriately? | Y | Y | Y | Y | Y | Y | Y | Y |
| Is the analysis described? | N | Y | Y | Y | Y | Y | Y | Y |
| Is the analysis appropriate? | Y | Y | ? | Y | Y | Y | Y | Y |
| Is there external validation? | Y | ? | ? | N | ? | Y | ? | ? |
| Does the evidence support the conclusions? | Y | ? | Y | Y | Y | Y | Y | Y |
| Are any limitations of the study addressed? | N | Y | N | Y | Y | Y | ? | Y |
| Are the results adjusted for confounding factors? | Y | Y | N | Y | Y | Y | Y | N |
| Quality Score | 14 | 12 | 9 | 13 | 12 | 16 | 13 | 10 |
| (82%) | (71%) | (53%) | (76%) | (71%) | (94%) | (76%) | (59%) |
- * Studies rated to be of poor methodological quality
- b Y = yes; N = no
Using the 17-item quality assessment tool, papers scoring <60% were deemed to be of poor methodological quality and >60% of acceptable methodological quality (Kuijpers et al., 2004) (see Table 1). All of the papers that met the inclusion criteria were included in this review.
Grading of evidence
In order to quantify the strength of evidence for each prognostic indicator, a grading tool was used. This tool was developed from the Oxford Centre for Evidence-Based Medicine (Phillips et al., 2007) for a review investigating the prognostic factors of the progression for hip OA (Wright et al., 2009). The categories of evidence were strong, moderate, weak, inconclusive or conflicting (refer to Table 2).
| Strong evidence | Consistent evidence in ≥2 acceptable quality cohort studies |
| Moderate evidence | Consistent evidence in 1 acceptable quality cohort study and ≥1 low quality cohort study |
| Weak evidence | Findings in 1 acceptable quality study or consistent evidence in ≥3 low quality cohort studies |
| Inconclusive evidence | Findings in 1 low quality cohort study |
| Conflicting evidence | Conflicting evidence in any cohort studies |
Data extraction
The following data were extracted from included studies: author, year of publication, country, setting, sample size, study objectives, method of coping strategy measurement, outcome measure and key results. Statistically significant results were defined as p values <0.05 or odds ratios (ORs) >1 with non-overlapping 95% confidence intervals (CIs).
Results
Study selection
A total of 1,468 titles were identified by the systematic search, of which 16 were duplicates; 1,297 articles were excluded on title alone. The lead reviewer read 171 abstracts, and a further 123 studies were subsequently excluded. The remaining 48 studies were read and six of these met the inclusion criteria for this review. Reasons for exclusion at this stage included: 15 papers for being of cross-sectional design, nine for not reporting data for OA, seven for being based in secondary care, five for not measuring coping strategies, three for being intervention studies and two for being literature reviews. An additional two papers were found from performing a manual search on authors identified from the initial search. This gave a total of eight papers to be included in this systematic review (see Figure 1).
Study characteristics
Of the eight papers included in this review, all investigated OA of the hip, knee or both. None of the included studies considered OA of the joints of the upper limb or spine. Half of the papers reported pain as an outcome, three reported physical disability, three reported mood and one reported depression. Sample size ranged from 36 to 364 (mean = 191.5). The follow-ups varied from daily diaries for 30 days, up to three years; the mean follow-up was 10.88 months.
The instruments used to assess coping were the Coping Strategies Questionnaire (Rosenstiel and Keefe, 1983), Daily Coping Inventory (Stone and Neale, 1984), Arthritis Self-Efficacy Scale (Lorig et al., 1989), Pain Management Inventory (Brown and Nicassio, 1987), Pain Coping Inventory (Kraaimaat et al., 1997), Fear-Avoidance Beliefs Questionnaire (Waddell et al., 1993) and Summary of Arthritis Management Methods Questionnaire (Hampson et al., 1993). The tools used to measure disability ranged from well-recognized, validated tools such as the Western Ontario and McMaster Index (Bellamy et al., 1988a,b) to observational tasks which were not validated (Rejeski et al., 2001; Steultjens et al., 2001). Those measuring mood (Affleck et al., 1999; Hampson et al., 1996; Keefe et al., 2004) used the Profile of Mood States (Lorr, 1988; Lorr and McNair, 1982); in addition to this, one paper (Hampson et al., 1996) used the Arthritis Impact Measurement Scale (Meenan et al., 1992). Three papers measured pain as an outcome; two used a visual analogue scale (Allen et al., 2006; Steultjens et al., 2001) and the other (Affleck et al., 1999) used the Rapid Assessment of Disease Activity in Rheumatology (Mason et al., 1986) (see Table 3 for the complete lists of the assessment tools used in each paper).
| Author, year and country | Setting | Follow-up period | Predictors (measures used) | Outcome (measures used) |
|---|---|---|---|---|
| Blalock et al., 1995, USA | Community | 6 months | Problem solving, problem avoidance, social withdrawal, self-criticism, turning to religion | Affect (Centre for Epidemiologic Studies Depression Scale (Radloff, 1977), Positive and Negative Affect Schedule (Watson et al., 1988), Arthritis Impact Measurement Scale (Meenan et al., 1992) |
| Hampson et al., 1996, USA | Community | 3 months | Active and passive coping [Summary of Arthritis Management Methods (Hampson et al., 1993), Pain Management Inventory (Brown et al., 1987)] | Mood (Profile of Mood States (Lorr, 1988), Arthritis Impact Measurement Scale (Meenan et al., 1992) |
| Affleck et al., 1999, USA | Community and rheumatology practices | 30 days | Emotion-focused coping [Daily Coping Inventory (Stone and Neale, 1984)] | Pain [Rapid Assessment of Disease Activity in Rheumatology (Mason et al., 1986)], mood [abbreviated version of the Profile of Mood States-B (Lorr et al., 1982)] |
| Rejeski et al., 2001, USA | Community | 30 months | Self-efficacy (level of certainty that they could perform and repeat the stair climb) | Disability (time taken to climb up and down five stairs; a 5-point Likert scale) |
| Steultjens et al., 2001, Netherlands | Primary care | 36 weeks | Passive coping: resting and active coping: pain transformation [Pain Coping Inventory (Kraaimaat et al., 1997), Fear-Avoidance Beliefs Questionnaire (Waddell et al., 1993)] | Disability (observation)Pain (visual analogue scale 0–100) |
| Sharma et al., 2003, USA | Community | 3 years | Self-efficacy (function subscale of Arthritis Self-Efficacy Scale (Lorig et al., 1989) | Disability [Western Ontario and McMaster Index (Bellamy et al., 1988a, b), Physical Activity Scale for the Elderly (Washburn et al., 1993)] |
| Keefe et al., 2004, USA | Primary care and community | 30 days | Catastrophizing, coping efficacy, problem-focused coping [Daily Coping Inventory (Stone et al., 1984), Coping Strategies Questionnaire (Rosenstiel et al., 1983)] | Mood [Profile of Mood States-B (Lorr et al., 1982)] |
| Allen et al., 2006, USA | Primary care | 30 days | N/a | Pain (visual analogue scale 0–10) |
Quality assessment
Included studies scored between 9 and 16 points on the 17-point methodological quality assessment scoring system, with a mean score of 12 points. All the studies gained a point for defining their objectives and study design. Few studies gained points for external validity and for using a sample that was representative of the target population. Of the eight papers that met the inclusion criteria, two were considered to be of poor methodological quality and six were of acceptable quality (see Table 1).
Prognostic indicators
Nine coping-related prognostic indicators were identified and shown to be associated with OA outcome in at least one study. Using the pre-specified criteria to quantify the strength of evidence (see Table 2), one prognostic indictor had strong evidence and eight had weak evidence for their predictive value (see Table 4 for a summary of results).
| Author, year and country | Mean age, standard deviation and gender (%) | Initial sample size and baseline response rates (%) | Summary of results |
|---|---|---|---|
| Blalock et al., 1995, USA | 68.1 years; SD = 8.00 | 300* | Problem solving was predictive of positive affect [F(1,239) = 23.85; p < 0.0001], self-criticism [F(1,241) = 5.48; p < 0.03), social withdrawal [F(1,241) = 5.04; p < 0.03], lower problem avoidance [F(1,241) = 6.83; p < 0.01] and turning to religion [F(1,241) = 4.82; p < 0.03] were predictive of negative affect. |
| 83% females | |||
| Hampson et al., 1996, USA | 71 years; SD = 5.7 | 80 | Passive coping was predictive of negative mood (partial r = 0.28; p < 0.01). Active coping was predictive of less negative mood (partial r = −0.34; p < 0.01). |
| 71% female | 87.8% | ||
| Affleck et al., 1999, USA | 62.1 years | 269 | Seeking emotional support was predictive of less next-day pain (b = 0.59; p = 0.05). Emotional-focused coping: redefining (b = 0.332; p = 0.01) and venting emotions (b = 0.377; p = 0.01) increased next-day negative mood. |
| 57.7% female | 54.6% | ||
| Rejeski et al., 2001, USA | 71.8 years; SD = 5.00 | 480 | Self-efficacy was predictive of disability (p < 0.01), particularly when knee strength was low. |
| 51% female | 51.1% | ||
| Steultjens et al., 2001, Netherlands | 68.3 years; SD = 8.7 | 190* | Resting was predictive of increased disability (b = 0.219; p = 0.024). Active coping with pain transformation predicted higher levels of pain (b = 0.206; p = 0.003) |
| 77% female | |||
| Sharma et al., 2003, USA | 68.6 years; SD = 10.8 | 285 | Self-efficacy was a protective factor (adjusted OR 0.80/5 points, 95% CI 0.65–0.98) for disability. |
| 73% female | 89.1% | ||
| Keefe et al., 2004, USA | 64.4 years; SD = 8.26 | 100* | Catastrophizing (b = −0.087, F = 5.80; p < 0.05) and problem-focused coping (b = 0.153, F = 4.66; p < 0.05) predicted more negative mood. Coping efficacy predicted more positive mood (b = 0.217; F = 8.62; p < 0.01). |
| 61% female | |||
| Allen et al., 2006, USA | 62.9 years; SD = 10.4 | 39 | Mean pain score for participants who used more emotion-focused coping was 4.8 compared with 3.4 for those who did not use emotion-focused coping**. |
| 11% female | 78% |
- * Unable to determine response rates;
- ** statistical significance, inferential test statistics and SDs were not reported.
Strong evidence
Higher levels of self-efficacy at baseline were shown to be a significant predictor of reduced disability (Rejeski et al., 2001; Sharma et al., 2003). Sharma et al. (2003) demonstrated that self-efficacy was predictive of less disability at three-year follow-up (OR 0.79, 95% CI 0.67 to 0.93). Rejeski et al. (2001) found that self-efficacy predicted less self-reported disability (p < 0.01) and greater physical performance (p < 0.001) at 30-month follow-up. This relationship was most pronounced when knee strength was low.
Moderate evidence
Studies investigating the association between problem solving and pain severity (Affleck et al., 1999; Allen et al., 2006) revealed that there was no longitudinal relationship between the two entities.
Weak evidence
It was found that lower problem avoidance (p < 0.01), social withdrawal (p < 0.03), self-criticism (p < 0.03) and turning to religion (p < 0.03) were predictive of symptoms of depression (Blalock et al., 1995), as measured by the Centre for Epidemiological Studies Depression Scale (Radloff, 1977). Self-efficacy and catastrophizing were investigated with mood – using the Profile of Mood States (Lorr and McNair, 1982) – as an outcome in one study (Keefe et al., 2004). Higher coping efficacy in the morning was shown to be prognostic of improved mood in the evening (p < 0.01), whereas, catastrophizing was shown to be predictive of more negative mood (p < 0.05) (Keefe et al., 2004). Active coping – adapting to a situation in an attempt to control outcome (Brown and Nicassio, 1987) – was shown to be predictive of an increase in pain (p = 0.003) (Steultjens et al., 2001) and less depressed affect (p < 0.01) (Hampson et al., 1996). Passive coping significantly predicted greater disability (p = 0.024) (Steultjens et al., 2001) and increased negative mood (p < 0.05) (Hampson et al., 1996), using the Profile of Mood States (Lorr and McNair, 1982).
Inconclusive evidence
The prognostic evidence for emotion-focused coping was demonstrated in one study (Affleck et al., 1999) which our assessment tool deemed to be of low quality. This showed that emotion-focused coping strategies such as seeking emotional support resulted in reduced next-day pain, while other emotion-focused strategies, such as redefining and venting emotions, were predictive of increased negative mood using the Profile of Mood States-B (Lorr and McNair, 1982).
Conflicting evidence
Problem-focused coping was shown to be predictive of daily negative mood in one study (Keefe et al., 2004), whereas a further study showed it to be predictive of less depressed affect at six-month follow-up (Blalock et al., 1995).
Discussion
Summary of main findings
The aim of this systematic review was to assess the prognostic value of self-efficacy and the coping strategies used by adults with OA within the community and primary care. It has been demonstrated that self-efficacy and active coping are predictive of improved mood, while passive coping and catastrophizing may be predictive of negative mood. Self-efficacy was shown to be predictive of reduced disability, whereas, passive coping was predictive of increased disability. There was also weak evidence to suggest that active coping was predictive of increased pain. A wide variation of coping strategies and outcomes were investigated by the included studies; as a consequence, only weak evidence was found for the prognostic value in the majority of coping strategies investigated. The studies included in this review show that there is weak evidence for the prognostic value of passive coping strategies in relation to increased levels of disability (Steultjens et al., 2001). This is supported by two cross-sectional studies investigating passive coping in people with OA (Perrot et al., 2008; Prior and Bond, 2004). Furthermore, the prognostic value of these strategies has also been demonstrated in longitudinal studies of back pain (Jones et al., 2006; Mercado et al., 2005) and rheumatoid arthritis (Brown and Nicassio, 1989).
One study identified catastrophizing to be a significant predictor of depressed mood in people with OA (Keefe et al., 2004). Systematic reviews of other musculoskeletal conditions such as low back pain (Pincus et al., 2002), rheumatoid arthritis and fibromyalgia (Edwards et al., 2006) and cross-sectional studies of OA (Keefe et al., 2000; Rapp et al., 2000; Somers et al., 2009b) have demonstrated that there may be an association between catastrophizing and pain/disability as outcomes. One of these papers (Edwards et al., 2006) suggests that catastrophizing may be multidimensional, involving interaction with numerous factors, including depressed mood. Consequently, further research is needed to distinguish whether catastrophizing is a prognostic indicator of mood, or whether the two factors are inter-related and act together to increase pain and disability in OA.
There is a vast literature describing the concepts and theories of coping. Two of the models that have informed many of the measures of coping used by the studies in this review have been based on the common-sense model of self-regulation (Leventhal et al., 1980) and the transactional model of stress (Lazarus and Folkman, 1984). Leventhal and colleagues state that a person's illness perceptions and emotional reactions about a health threat guide the development of action plans for coping with the problem and the emotion. The individual then evaluates the effectiveness of the coping strategy to determine whether to continue or adopt an alternative strategy. Lazarus and Folkman (1984) describe coping as an evaluative process, comprising primary and secondary appraisal. Primary appraisal is the individual's initial judgement of a stressful situation, whereas secondary appraisal occurs as a result of facing the situation and drawing on coping resources in order to deal with the situation. The way in which people cope can be determined by physical or psychological resources, including positive belief, also known as self-efficacy. Self-efficacy is an individual's belief that they can perform in a particular way to achieve specific goals. Individuals with a strong self-efficacy are deemed to view problems as tasks to be mastered and are able to recover quickly from setbacks and disappointments, whereas those with a weak sense of self-efficacy believe that difficult tasks and situations are beyond their capabilities and engage in avoidance coping. Individuals with a strong self-efficacy are better able to cope. Self-efficacy is, therefore, thought not to be a coping strategy per se but a mediator in initiating specific coping behaviours (Bandura, 1977, 1997). Despite this, a number of studies have considered self-efficacy as a prognostic indicator in OA outcome (Keefe et al., 2004; Rejeski et al., 2001; Sharma et al., 2003). There is strong evidence from these studies to suggest that people who have lower self-efficacy will have increased disability more than six months later. This seems to support the theory that self-efficacy is a mediator in initiating specific coping strategies, and, in this situation, coping strategies that are less effective. There is further support for the predictive value of self-efficacy and disability from cross-sectional studies of OA patients (Gaines et al., 2002; Maly et al., 2005; Marks, 2007; Prior and Bond, 2004).
Strengths and limitations
A systematic search was performed using several electronic databases. This, combined with a comprehensive search strategy, provided an expansive and specific search. This systematic review applied a rigorous methodological quality assessment to investigate potential sources of bias. While this approach is useful in determining potential sources of bias, we were dependent on the authors of the original studies fully reporting their methodology. Incomplete reporting could have resulted in the allocation of lower scores during the quality assessment process and, as such, had the potential to impact on our findings.
There were fewer prospective cohort studies investigating coping in adults with OA than initially expected. Having a range of different coping-related prognostic indicators and a number of outcomes (including pain, disability and depression), coupled with the small number of studies, made comparison difficult and resulted in only weak evidence for a number of prognostic indicators. This also prohibited a formal meta-analysis. Along with the diversity of outcomes reported in the studies, there was also variability in the way that coping strategies were quantified. Some studies used umbrella terms such as passive, active, problem-focused or emotional-focused, while others measured specific coping strategies such as catastrophizing, seeking emotional support or self-criticism. Using an evidence stratification tool (see Table 2) reduced this problem to some extent, as it allowed each combination of coping strategies and individual outcomes to be quantified into a category on the basis of the strength of the evidence. All studies used appropriate measurements for assessing both coping- and OA-related outcomes, but there was no standardized method among the studies for measuring coping strategies, disability, mood or pain. This could have led to inconsistencies in the findings due to the inability to make exact comparisons between studies.
Another possible limitation of this review is that the studies included were restricted to participants with lower limb (hip and knee) OA. This could have created bias in relation to the coping strategies utilized and the associated outcome. For example, people may perceive, and therefore cope, with OA of the hand differently from OA of the hip.
As with all literature reviews, publication bias could have influenced our findings. This relates to the general under-reporting of negative study findings and an increased focus on statistically significant findings. In an effort to identify grey literature and other unpublished material, the lead reviewer contacted experts working in the area to ascertain other relevant studies. However, no additional studies were identified using this method. Another potential limitation of this systematic review was the restriction to the English language, although no studies were excluded on this basis.
Implications for research and clinical practice
This review demonstrates the potential importance of self-efficacy as a prognostic indicator for OA. Due to the relatively small number of cohort studies identified by this review, our findings are limited. Further research is needed to investigate the full range of coping strategies used by people with OA and to determine their association with different outcomes. More high-quality studies are needed, using adequate sample sizes and standardized measurements, and measuring a range of outcomes. Future studies should measure pain, disability and depression, thereby allowing comparisons to be made between studies and thus demonstrating the true prognostic value of the coping strategies.
The findings in this review may have implications for clinical practice and the current management of OA. The National Institute for Health and Clinical Excellence (NICE) OA guidelines (National Collaborating Centre for Chronic Conditions, 2008) recommend a holistic approach to the management of OA that considers not only pharmacological and physical therapies, but also includes the assessment of psychological factors, including anxiety and depression. This review would suggest that advice on coping strategies should also be incorporated. However, before this is achieved, further high-quality cohort studies are needed to demonstrate the prognostic effects of each coping strategy. It is only then that randomized controlled trials can investigate fully whether these coping strategies can be manipulated into improving outcome.
Acknowledgements
The first author was funded by the West Midlands Deanery.
Appendix 1
MEDLINE search terms
MESH headings
Pain, arthritis, arthralgia, osteoarthritis, osteoarthritis hip, osteoarthritis knee, joint disease, adaptation psychological, avoidance learning, fear, internal external control, somatoform disorders, self-efficacy, cohort studies, epidemiologic studies, cross-sectional studies, follow-up studies, retrospective studies, longitudinal studies, prospective studies, questionnaires, models, theoretical, review literature as topic, review publication type, editorial publication type, health surveys, pain clinics, primary health care, family practice.
Free text
Musculoskeletal pain, chronic pain, arthropathy, coping, cop*, adaptation, adjustment, confrontation, cognitive restructuration, perceived control, distraction, positive thinking, avoidance, praying, helplessness, hopelessness, helplessness-hopelessness, (avoidance OR protective) AND behaviour*, disuse syndrome, fear-avoidance NEAR model, fear-avoidance, fear NEAR avoidance, catastrophising, catastrophizing, kinesiophobia, adaptive NEAR cogniti*, transactional processes, reinterpreting, causal attribution, attentional bias, self-efficacy, readiness to change, acceptance, passive coping, depression AND coping, negative affect, survey* OR survey, general population, primary care.
PsycINFO search terms
MESH headings
Pain, pain management, pain measurement, pain perception, somatoform pain disorder, arthritis, joint disorders, musculoskeletal disorders, chronic pain, coping behaviour, test anxiety, adjustment, emotional adjustment, fear, somatoform disorders, somatoform pain disorders, adaptive behaviour, cognitive restructuring, distraction, avoidance, avoidance conditioning, helplessness, learnt helplessness, hopelessness, self efficacy, readiness to change, epidemiology, prospective, longitudinal studies, questionnaires, theories, survey, cohort analysis, literature review, retrospective studies, primary health care, general practitioners, family physician.
Free text
Joint disease, osteoarthritis, arthropathy, avoidance learning, internal-external control, coping, cop*, confrontation, perceived control, positive thinking, praying, avoidance behaviour, protective behaviour, catastrophizing, catastrophising, fear-avoidance model, fear avoidance, adaptive cognitions, transactional process*, kinesiophobia, passive coping, depression AND cop*, negative affect, cross-sectional, cohort stud*, editorials, health survey*, primary care, general population.
Embase search terms
MESH headings
Chronic pain, hip pain, knee pain, musculoskeletal pain, pain, osteoarthritis, knee osteoarthritis, hip osteoarthritis, arthritis, chronic arthritis, knee arthritis, arthralgia, arthropathy, somatoform disorder, coping behaviour, fear, adjustment, adaptive behaviour, avoidance behaviour, helplessness, hopelessness, causal attribution, patient attitude, control, retrospective study, prospective study, longitudinal study, coping strategy questionnaire, questionnaire, ‘ways of coping questionnaire’, epidemiology, review, health survey, theory, editorial, primary health care, pain clinic, general practice, primary medical care.
Free text
Pain perception, avoidance learning, cognitive restructuring, distraction, self efficacy, readiness to change, psychological adaptation, confrontation, perceived control, positive thinking, praying, protective behaviour, catastrophizing, catastrophising, kinesiophobia, adaptive cognitions, transactional process*, acceptance, reinterpreting, causal attribution, attentional bias, passive coping, depression AND coping, negative affect, family practice, general population, primary care.
CINAHL search terms
MESH headings
Arthritis, arthralgia, osteoarthritis, osteoarthritis-hip, osteoarthritis-knee, joint disease, coping, adaptive psychological, fear, behavioral changes, hopelessness, cross-sectional studies, prospective studies, retrospective studies, coping strategies questionnaire, ways of coping questionnaire, questionnaires, surveys, survey research, epidemiology, epidemiological research, systematic review.
Free text
Avoidance behaviour, avoidance, adjustment, adaptive behaviour, control, catastrophizing, catastrophising, fear-avoidance, adaptive cognitions, kinesiophobia, passive coping, depression and cop*, negative affect, longitudinal study, editorial.
