Disability is proposed to be an important outcome in pain research (Deyo et al., 1994), and 30% of persons with neck, shoulder, or back pain may be expected to report limitations in daily life (Picavet and Schouten, 2003). Psychological factors are related to both the onset and development of spinal pain and disability (Linton, 2000). Self-efficacy, i.e. one's confidence in performing a particular behavior and in overcoming barriers to that behavior (Bandura, 1977 and Bandura, 1997), is believed to be an important mediator of disability related to pain. Self-efficacy was found to influence adjustment to a pain condition (Jensen et al., 1991), and pain-related disability (Estlander et al., 1994 and Lackner et al., 1996), to mediate the relationship between pain intensity, disability, and depression (Arnstein, 2000 and Arnstein et al., 1999), to predict lifting capacity (Lackner and Carosella, 1999), and pain behaviour and avoidance (Asghari and Nicholas, 2001) in chronic pain patients.
During the last decade, fear avoidance (Kori et al., 1990 and Vlaeyen et al., 1995) has gained increased empirical support as a mediator of disability in chronic pain (Vlaeyen and Linton, 2000). Empirical support for fear avoidance in relation to disability comes from several studies (Al-Oubadi et al., 2000, Buer and Linton, 2002, Crombez et al., 1999, Fritz and George, 2002, Fritz et al., 2001, Geisser et al., 2000 and Picavet et al., 2002). In a primary health care setting, however, van den Hout et al. (2001) showed that pain intensity and pain catastrophizing were better predictors of disability than pain-related fear. When prediction of disability by both self-efficacy and fear avoidance was examined simultaneously, self-efficacy was found to be the more powerful predictor (Ayre and Tyson, 2001).
Most studies concerning self-efficacy, fear avoidance, and disability have been conducted in secondary or tertiary health care settings where patients are highly selected due to the referral filtering process (Turk and Rudy, 1990). However, most MSP patients are managed in primary health care, and results from secondary or tertiary settings may not necessarily generalise to primary health care patients. Specifically, patients who remain in primary health care may be expected to be less disabled than patients who are referred to specialised pain clinics or rehabilitation clinics. Since self-efficacy may explain why patients persist in confronting daily activities in the face of obstacles such as pain, we argue that it is a more important predictive factor than fear avoidance in primary health care clients. Thus, the purpose of this study was (1) to test the hypothesis that self-efficacy is a better predictor of disability than fear avoidance variables and pain intensity in a primary health care sample of patients with subacute, chronic or recurring MSP, and (2) to replicate the findings in a second sample.
The results of this study confirmed our hypothesis that self-efficacy is a better predictor of disability than fear avoidance variables and pain intensity in a primary health care sample of patients with subacute, chronic, or recurring musculoskeletal pain. The results were replicated in a second sample. Gender, age, and pain duration were not significantly correlated to any of the variables in the regression model.
Bivariate correlation analyses showed that self-efficacy was significantly, and negatively, associated with disability, which is in accordance with the results reported by Arnstein et al., 1999 and Arnstein, 2000, and Lackner et al. (1996). In both samples self-efficacy showed the highest correlations with disability, as compared to pain catastrophizing and kinesiophobia. Self-efficacy correlated (negative association) with pain catastrophizing (r=-0.44, P<0.001), and kinesiophobia (negative association) in both samples (r=-0.32 and -0.38, P<0.001). The latter finding is consistent with the results of Ayre and Tyson (2001) who found a significant negative correlation between self-efficacy and fear avoidance in a sample of patients with chronic low back pain. However, the squared correlation coefficients, representing 10 and 15% of shared variance, respectively, in the two samples in the present study indicate that these two constructs were not overlapping to a great extent.
The bivariate analyses also showed positive and significant associations of fear avoidance variables with disability. This is in accordance with other studies reporting significant bivariate correlations (positive associations) between pain catastrophizing or pain-related fear, and disability (Crombez et al., 1999, Fritz and George, 2002, Fritz et al., 2001, Koho et al., 2001 and van den Hout et al., 2001).
The fear avoidance variables did predict a unique proportion of the variation in PDI scores in both samples, albeit considerably smaller than did self-efficacy. One explanation may be that fear avoidance is a more important construct in patients who are more dysfunctional and therefore managed in secondary and tertiary health care settings. Much of the work regarding fear avoidance and disability comes from pain clinic or rehabilitation program samples, e.g. Waddell et al., 1993, Vlaeyen et al., 1995 and Crombez et al., 1999, where patients are highly selected. The primary health care samples in the present study are likely to be more functional and better adjusted than the samples used to develop the fear avoidance construct. van den Hout et al. (2001), using a primary health care sample found, for example, that pain-related fear was a less important predictor of disability than pain intensity and pain catastrophizing. Fear avoidance has, however, been shown to predict disability (Picavet et al., 2002) and activities of daily living (Buer and Linton, 2002) in population-based samples, and to be present in acute stages of low back pain (Fritz et al., 2001). Thus, fear avoidance seems to be present in patients in different stages of MSP and at different levels of health care. Further research involving both self-efficacy and fear avoidance in different types of samples and settings will clarify this matter.
Pain intensity did not emerge as a consistently-significant predictor of disability in the two samples, which is contrary to the results reported by van den Hout et al. (2001). Because van den Hout et al. measured pain by the McGill Pain Questionnaire, which is a measure of both pain intensity and pain quality, the different modes of pain measurement may explain the differing results. Another possible explanation is that all subjects in the study of van den Hout et al. (2001) were sick-listed at entry of the study, as compared to about 37% of the subjects in both our samples (Table 1), indicating that our samples may have been less influenced by pain intensity.
Pain duration did not correlate significantly with any of the variables in the model, although pain duration ranged from 1 month to several years in both samples. Patients who are able to cope with their pain are likely to remain in primary health care (Turk and Rudy, 1990), and for those patients, pain duration may not be of great importance.
The clinical implications of the results in this study involve the need for primary health care professionals to focus on pain-related beliefs rather than on pain intensity reports in these patients. By relying on knowledge of pain duration and assessment of pain intensity alone to guide management, clinicians are likely to overlook important aspects of disability, and subsequently to engage in ineffective treatment strategies. As an alternative, systematic assessment of self-efficacy beliefs and fear avoidance beliefs regarding activities relevant for daily living would make a better starting point in the management process. Treatment strategies should focus on improving functional abilities related to specific and prioritised activities, using a small-steps approach to ensure success, thus enhancing self-efficacy and reducing fear.