Introduction: In recent years most industrialized nations have been confronted with a dramatic increase in cases dealing with back pain which, because of the associated costs, has put a strain on health care systems. Because of the tremendous costs caused by patients with chronic illnesses, there is a common interest in identifying determinants responsible for the pain becoming chronic. According to the fear-avoidance model, chronic back pain is the cause of specific patient beliefs according to their perception and appraisal of back pain and their assumptions about the connection between pain and work activities. These fear-avoidance beliefs are assumed as the main cognitive factors of adapting to chronic invalidism. In this study a German version of a self-reporting instrument that assesses fear-avoidance beliefs (FABQ, Waddell et al. 1993) is presented.
Methods: A translated version of the "Fear-Avoidance Beliefs Questionnaire" was given to 87 patients with back pain, who took part in a multimodal treatment program in an outpatient pain unit. The questionnaire was administered at different times in the therapeutic process. We examined the factor structure of the questionnaire and conducted the first examinations concerning its psychometric properties.
Results: The results of our data set concerning different aspects of reliability and validity conform the good psychometric properties of the questionnaire. The factor analysis revealed a three-factor solution, which accounted for nearly 60% of the variance. The first factor made up 34% of the variance. The six corresponding items dealt with "back pain caused by work". The second factor made up 13% of the variance. The content of the 5 items is focused on the patients' view about his/her probable return to work. Beliefs about physical activity in back pain (factor 3, 5 items) accounted for 12% of the variance. Correlation analysis demonstrated that all three factors are of independent value, which showed satisfactory reliability. Furthermore, we analyzed the prognostic value of the three factor dimensions. Patients completed multidisciplinary treatment according to the concept of "functional restoration". At the end of treatment, factor 1 (causation by work) was not diminished, but both other factors showed significant reductions. In a discriminant factor analysis, back-to-work could be significantly predicted by factor 3 (beliefs about physical activity) and factor 2 (patients' prognosis about their probable return to work at the beginning of treatment). Furthermore, length of time off work and age accounted for an additive classification rate of 77%. The FABfactor 1 (causation by work) was of no predictive value.
Conclusions: The results suggest that primary target points for further investigation include an analysis of the patients' beliefs about their pain, with special emphasis on fear-avoidance beliefs. It is also necessary to analyze the behavioral assumptions of the fear-avoidance model in an experimental design. Furthermore, our results concerning the psychometric properties of the questionnaire should be proved in a far larger sample of back pain patients.
