- Top of page
- SUBJECTS AND METHODS
Because of its prevalence, knee osteoarthritis (OA) is the most common cause of chronic disability in community-dwelling elderly adults (1). Limitations in walking, stair climbing, and transferring contribute to this disability (1), and understanding the determinants of these mobility limitations could aid in the development of strategies to prevent disability. One such determinant is self efficacy (2–4). Self efficacy refers to a person's belief in their capability to organize and execute actions required to achieve goals (5). Self efficacy is task specific; for example, self efficacy can be applied to academic, social, and mobility tasks. In social cognitive theory, Bandura contends that how a person performs is better predicted by that person's beliefs about his or her capabilities than by what he or she is actually capable of accomplishing (5).
A number of studies have revealed relationships between self efficacy and performance. Self efficacy predicts athletic performance in adults and children (6). Similarly, self efficacy is important to the performance of physical tasks in individuals with knee OA. Self efficacy was moderately related (r = −0.56, P = 0.02) to a woman's self-reported mobility (2) and, together with a balance measure, explained 42% of performance scores in walking, stair climbing, and transferring in 50 women with knee OA (3). In men and women with knee OA, high baseline self efficacy decreased the odds of self reporting poor physical functioning and demonstrating poor performance of a sit-to-stand activity 3 years later (7). Also, self efficacy explained at least 45% of the variance in walking, stair climbing, and transferring in 54 persons with knee OA (4). These studies demonstrate a strong relationship between self efficacy and performance in this population. Some research also suggests that self efficacy may have a causative effect on performance. Participants enrolled in an educational program shown to improve self efficacy experienced a 15–20% decline in pain and reduced number of physician visits over a 4-year period (8, 9).
Little is known about the factors that make up self efficacy in the knee OA population. Because self efficacy is strongly related to performance in individuals with knee OA, understanding what factors relate to self efficacy for physical tasks could help clinicians and researchers develop strategies to improve physical performance, thereby preventing disability. Our goal was to explore what personal factors, disease pathophysiology, and impairment factors were determinants of self efficacy for physical tasks in this population.
Although no studies have been directed toward uncovering the determinants of self efficacy, a number of personal (age, education, depression, and anxiety), pathophysiologic (joint space narrowing, alignment, obesity, and strength), and impairment (pain, stiffness) factors might be involved. In terms of personal factors, knee OA is more common after age 50 (10). It is possible that older individuals have poorer confidence for physical tasks, perhaps because of an expectation of physical decline or reduced activity. Persons with knee OA are more likely to report depression and anxiety (11), which could influence the perception of capability. Educated individuals tend to have better health outcomes; it is possible that these health outcomes are related to improved levels of confidence. Persons with knee OA share common pathophysiologic factors that may influence self efficacy. For example, knee OA has been related to altered alignment because of joint space narrowing in the medial compartment (12). Altered alignment is easily perceived by the individual and may influence confidence in his or her knee. Mechanical joint loading can be influenced by strength, which is thought to have a protective effect (13–15). Obesity has been extensively studied in terms of its mechanical effects: obesity in knee OA has been associated with increased varus alignment (16–18), more severe radiographic OA (16), greater joint space narrowing (16), and deviation from the ideal mechanical axis of the lower extremity (17). However, no studies were found that related obesity to a psychosocial factor even though obesity is a variable easily perceived by the individual. Finally, the impairments of pain and stiffness could influence an individual's confidence to perform physical tasks. Currently, no literature examines which of these factors relate to self efficacy in the knee OA population.
In examining self efficacy and its potential to influence performance, it is desirable to study community-dwelling adults who would derive the maximum benefit from an intervention, should this strategy become possible. However, much of the OA disability research conducted involves clinical populations, particularly from surgical waiting lists. Surgical populations likely require more invasive treatments than improvements in self efficacy to manage knee OA, and it is unclear whether results from studies of these clinical populations can be extrapolated to those with knee OA living in the community. Therefore, there is a need to conduct self-efficacy research relevant to disability in community-dwelling adults with knee OA.
The purpose of this study was to determine what personal, pathophysiologic, and impairment factors relate to self efficacy for common physical tasks in community-dwelling adults with knee OA. The results will provide insight into the meaning of self efficacy scores in individuals with knee OA. Because no previous research into the determinants of self efficacy has been conducted in the knee OA population, we chose a cross-sectional design that was capable of examining more than 1 factor.
SUBJECTS AND METHODS
- Top of page
- SUBJECTS AND METHODS
A convenient sample of community-dwelling adults participated. Participants were recruited using a free community newspaper circulated to >55,000 homes. Inclusion criteria required individuals to be ≥50 years of age with physician-diagnosed medial-compartment knee OA (n = 54). In all but 2 cases that required an orthopedic surgeon, family physicians confirmed the diagnosis based on criteria outlined by the American College of Rheumatology (19). Radiographs, another inclusion criterion, were obtained to confirm that joint space narrowing was greater in the medial than lateral compartment because it is theorized that medial and lateral knee OA are the result of different mechanical pathologies (20, 21). In cases of bilateral knee OA, the more painful knee, determined by self report, was examined. Exclusion criteria included a hip or ankle condition or any corrective surgery on the ipsilateral limb, history of a fall over the past year, and any medical conditions that could be exacerbated by the protocol such as unstable heart disease. The Queen's University Health Sciences Research Ethics Board approved the study and all participants provided written informed consent.
Recruitment from the community was chosen to avoid a bias toward only severe, clinical, or surgical populations. Recruitment continued for 1 year. This lengthy process resulted because we confirmed diagnosis with physicians to avoid problems associated with inaccurate self report, and the inclusion and exclusion criteria led to the exclusion of 60 potential participants.
Functional Self Efficacy.
The Functional Self-Efficacy subscale (FSE) of the Arthritis Self-Efficacy Scale was the measurement used to determine self efficacy for physical tasks (22). This 9-item questionnaire (Figure 1), designed specifically for arthritic populations, uses a visual analog scale in which a higher score indicates greater self efficacy, a positive result. The total possible range is 0–100. The test-retest correlation coefficient for data derived from the FSE subscale was reported as r = 0.85 (22).
Independent measures reflected personal, pathophysiologic, and impairment factors relevant in knee OA. The same tester (MRM) completed all measurements and collected and scored all questionnaires.
Participants reported their date of birth (age, years) and educational level (years). Education was chosen instead of income because it is less likely to be changed by disease. Education was recorded as the number of years of formal education.
Depressive symptomatology was assessed with the Center for Epidemiologic Studies Depression (CES-D) Scale. The CES-D is a 20-item self-report Likert scale developed to identify depression in the general population (23). A high score indicates more depressive symptoms. Emphasis of the scale is on mood, guilt, worthlessness, helplessness, loss of appetite, and sleep disorders. In terms of reliability, correlation coefficients between 0.85 and 0.90 have been reported in general and patient populations (23), and data from the scale have been shown to be valid in arthritic populations (24, 25). A score ≥16 indicates that the individual likely experienced some depression over the past week (23). Scores of 7 have been reported in the general population (23).
The State-Trait Anxiety Inventory (STAI) includes 2 self-administered subscales where each response is marked on a 4-point scale (26). A high score indicates more anxious symptoms. The State Anxiety subscale contains 20 questions addressing how an individual feels at a given moment. A second subscale measures Trait Anxiety using 20 questions that ask how an individual generally feels. The validity and test-retest reliability (r = 0.73–0.86) for data from the STAI have been established and normative data have been published (26, 27). Older adults with generalized anxiety score ∼93 points, whereas healthy older adults score ∼58 points on the STAI (27).
First, medial compartment joint space width (MJSW; millimeters) was measured from standardized, scaled, and parallax corrected radiographs to represent the radiographic severity of OA. Participants were asked to stand barefoot, with equal weight on the left and right legs, in a position that placed the femoral condyles in the coronal plane. The MJSW, known to be the best measure of radiographic progression (28, 29), was the minimum distance between the tibial plateau and femoral condyle in the medial compartment. This measurement has demonstrated superior accuracy, sensitivity, and interrater reliability (intraclass correlation coefficient [ICC] 0.79–0.82) to the Kellgren/Lawrence scale (ICC 0.57–0.66) (30) and allows parametric statistical analysis. Second, the varus-valgus angle (degrees) was measured as the angle between the tibial shaft and femoral shaft in the coronal plane. A varus angle was designated positive.
Body mass index (BMI; kg/m2) was calculated from measured height and weight. BMI is a widely used measure of obesity where “overweight” is classified as a BMI ≥25 kg/m2. In this study, the BMI score was used as a continuous variable. Height and weight were measured while the participant was barefoot, wearing shorts and a t-shirt.
Strength of the quadriceps (Nm) and hamstrings (Nm) muscles was measured using the Biodex System 3 isokinetic dynamometer. Each participant completed 5 submaximal practice trials before recording. Then, 5 trials of maximum effort concentric knee flexion and extension at 60°/second were performed with verbal encouragement. To maximize reliability and validity, the data were “windowed” to remove acceleration and deceleration, thereby eliminating impact artifact (31, 32). Only data recorded within 0.1°/second of 60°/second were included. The reliability of windowed data on concentric knee flexion and extension has been reported as r = 0.90–0.96 on Biodex systems (32). For each muscle group, the 5 peak windowed values for flexion and extension were averaged for each participant.
The Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) Pain and Stiffness subscales were used to collect information on pain (millimeters) and stiffness (millimeters). The WOMAC is a self-administered questionnaire for individuals with hip or knee OA consisting of 24 questions on pain, stiffness, and physical function (33). Using a visual analog format, WOMAC scores vary between 0 and 100 mm, best to worst scores, respectively. The reliability and validity of data from the WOMAC and its subscales have been established (33).
A repeated-measures analysis of variance (ANOVA) was calculated to investigate fatigue over 5 trials of quadriceps and hamstrings strength testing. Pearson's correlation coefficients were calculated to evaluate relationships between variables. A stepwise linear regression was performed using the FSE as the dependent measure. The personal factors (age, education level, CES-D, STAI), pathophysiologic factors (MJSW, varus-valgus angle, BMI, quadriceps strength, hamstrings strength), and impairments (pain, stiffness) were independent variables. In the stepwise regression, the stepping method criteria required an F value ≥0.05 to be included in the model and an F value ≤0.10 to be removed from the model. The SPSS version 11 (SPSS, Chicago, IL) was used.
- Top of page
- SUBJECTS AND METHODS
Our goal was to investigate what personal, pathophysiologic, and impairment factors relate to self efficacy for physical tasks in community-dwelling adults with knee OA. It is important to note that the FSE scores of these participants were considerably higher than the FSE scores reported in other studies (3, 7), which is perhaps a reflection of recruiting community-dwelling adults instead of surgical populations. Therefore, the generalizability of these results must be limited to community-dwelling adults with knee OA and relatively high self efficacy. The results highlight the importance of stiffness in determining the degree of self efficacy individuals with knee OA have for physical tasks. Age and depression were also significant factors explaining the variance of the FSE; however, the contribution of age and depression was smaller. Hamstrings strength also contributed a relatively small amount to the variance of the FSE.
Stiffness was the most important variable, explaining >30% of the variance in the FSE. The greater stiffness, the poorer the individual rated their self efficacy for physical tasks. A literature search did not reveal another knee OA study that demonstrated the impact of joint stiffness, the sensation of restriction or slowness in the ease of movement, on self efficacy for physical tasks. Furthermore, other than its use as an outcome measure, no study was found that identified why self-reported stiffness is important in this population. Our finding that stiffness relates strongly to self efficacy is consistent with social cognitive theory. In addition to past performance, vicarious experience, motivation, and persuasion, negative physical feedback is thought to be a source of self efficacy (5). Joint stiffness likely provides negative feedback to individuals with knee OA during physical activity and therefore influences self efficacy.
Our results demonstrated that hamstrings strength contributed 6% to the variance in self-efficacy scores, whereas quadriceps strength did not enter our regression model. Muscular weakness around the knee joint, particularly of the quadriceps, has been implicated in the development of knee OA (14, 34). There are 2 possible explanations for the greater role of hamstrings than quadriceps in self efficacy for physical tasks. First, it is possible that hamstrings strength plays a greater role than quadriceps in functional activities because hamstrings muscles have the multiple roles of controlling knee flexion and hip extension; balancing the large head, arms, and trunk over the legs; and maintaining stability during mobility tasks (35). Second, because hamstrings and quadriceps strength were highly correlated, the collinearity of these variables prevented quadriceps strength from entering the model. To investigate this second possibility, we performed a stepwise linear regression excluding only hamstrings strength. The resultant model of stiffness, age, and depression did not include any strength measure, yet explained approximately the same amount as our original model. These findings are surprising and difficult to interpret; however, they suggest that hamstrings strength in the context of self efficacy and performance may be more important than quadriceps strength. Much research supports the theoretical importance of quadriceps strength supporting the knee joint, however (13, 14), and as such we acknowledge the need for more research into the role of these muscles in performance. Also, we must acknowledge the possibility that strength may not be necessary to explain an individual's level of confidence for physical tasks. Perhaps work in a different cohort can help elucidate the role of quadriceps and hamstrings strength in self efficacy and physical performance in community-dwelling adults with knee OA.
The remaining 2 variables that were significant factors explaining the FSE scores were age and depression. Individuals are more likely to develop knee OA as they age (36), and our findings suggest that age is related to self efficacy: the older the individual, the poorer their self efficacy for physical activities. Perhaps the expectation of capabilities for physical tasks diminishes as individuals age or as additional comorbidities associated with aging hinder physical activity. Depression contributed ∼5% to the FSE such that higher levels of depressive symptoms related to poorer self efficacy. Depression, therefore, might be a consideration when addressing physical functioning in persons with knee OA. Depression has been shown to be an important predictor of physical functioning in the healthy elderly and those with knee OA (37, 38). In social cognitive theory, Bandura proposes that an individual's affect will influence self efficacy (5); our findings are consistent with Bandura's theory.
Pain was not a significant factor contributing to our model explaining FSE, perhaps because pain was strongly related to stiffness. Although our findings do not highlight pain as an explanatory factor for the confidence an individual has to perform a physical task, the moderate correlation coefficient between pain and the FSE (r = −0.44, P = 0.001) suggests that pain should not be ruled out as a consideration in the treatment of knee OA. Furthermore, previous work suggests that pain is the most important determinant of self-reported mobility and health in this population (14); thus, pain likely has a stronger relationship with a person's perception of his or her own mobility than with self efficacy for performance.
Anxiety did not enter the model as an explanatory variable of functional self efficacy. Anxiety was closely correlated with depression but subsequent regression analyses demonstrated that even when depression was not included as an independent variable, anxiety did not contribute to an explanation of self efficacy. Previous studies have clearly shown that individuals with knee OA have higher levels of anxiety, which likely contribute to pain intensity (37–39). Our participants demonstrated a level of anxiety (mean ± SD 63 ± 16) that was not appreciably different from the general population (mean ± SD 58 ± 6.8) and was much lower than older adults with generalized anxiety (mean ± SD 93 ± 13.1) (26, 27), which could explain why anxiety was not found to be an important variable. Furthermore, it may be possible that the method of recruitment and the high level of education in our group (mean ± SD years of formal education 14.9 ± 4.3) (40) may have been related to the lower level of anxiety in our cohort. Anxiety may have been included in our regression model if the level of anxiety in our participants was higher.
Radiographic findings of joint space width in the medial compartment and alignment were not related to functional self efficacy. Radiographic severity has not been related to pain or function in individuals with knee OA (41–43). It was not surprising that medial joint space did not relate to self efficacy (r = 0.17, P = 0.23) and did not enter the regression model. With regard to alignment, we found a slightly valgus (−0.81 ± 4.5°) position between the femoral and tibial shafts despite the greater narrowing in the medial than lateral compartment (an inclusion criterion) in our study. These findings suggest that the participants likely had mild to moderate disease severity. It is important to note, though, that medial joint space narrowing does not always result in varus alignment but rather a change in the angle between the femoral condyles and the femoral shaft and/or between the tibial plateau and tibial shaft (44).
Some limitations of this study must be considered. In terms of generalizability, our participants appeared to have a mild form of the disease, likely a reflection of our recruitment of community-dwelling rather than clinical or surgical wait list populations. Unlike other studies, our participants did not appear to have scores on the CES-D or STAI that were appreciably greater than the general population. The participants were overweight based on measurements of BMI, although to a lesser degree compared with other studies (45–47). The WOMAC scores indicate that the participants had relatively lower levels of pain. The scores on the self-efficacy questionnaire were considerably higher than those reported in other studies of individuals with knee OA. Together, these characteristics suggest that this sample likely had mild to moderate OA. The results are also likely biased toward educated white persons. Different results may be found in those with more severe scores on these variables and less self efficacy for physical tasks. Finally, it must be acknowledged that participants did have comorbid conditions and may have been unable to separate out the effects of knee OA and their comorbidities. However, it is unrealistic to recruit individuals with knee OA but no other health conditions.
No information about the duration of illness was recorded; therefore, the participants may have been in acute or chronic stages of the disease. The recruitment process of obtaining a diagnosis from the family doctor was lengthy (mean of 3 months between initial contact and participation), suggesting that the participants had knee OA for at least 3 months. Nevertheless, it must be considered that the duration of illness may have had an influence on the outcome. No attempt was made to control for medication usage; we acknowledge that medications may have influenced performance of physical tasks and therefore may have confounded the results. The causal influence of stiffness, hamstrings strength, age, and depression on Functional Self-Efficacy scores cannot be determined from this study. Finally, investigators are recommended to include at least 5 subjects for every independent variable included in regression analyses (48). In this study, 11 independent variables were used, suggesting that at least 55 subjects were required. Our study included 54 subjects, which may have resulted in an enhancement of type I errors.
Because our findings demonstrate that self efficacy for physical tasks is related to stiffness, strength, age, and depression, we suggest that further research should examine whether self efficacy mediates the relationship between these variables and physical performance in persons with knee OA. Also, future studies could aim to characterize the sensation of joint stiffness beyond intensity at 2 times during the day. Examining the mechanical aspects of stiffness on mobility and the relationship between mechanical stiffness and sensation would also add to the body of knowledge relevant to physical performance in persons with knee OA.
In conclusion, self efficacy for physical tasks is related to the sensation of stiffness, hamstrings strength, age, and level of depressive symptoms in individuals with knee OA. The results highlight the importance of self-reported stiffness in this group. Age and depression were also significant factors explaining the variance of FSE; however, the contribution of age and depression was relatively smaller. Hamstrings strength also contributed a relatively small amount to the variance of FSE. Clinicians and researchers could consider these variables when assessing the level of self efficacy for physical tasks in individuals with knee OA.