Study design: A prospective, interventional case series design.
Objectives: To determine the degree to which preintervention measures of anticipated pain and fear avoidance beliefs predict outcome after intervention for patients with delayed recovery from low back pain (LBP) for which they are not receiving workers' compensation.
Summary of background data: Anticipated pain and fear avoidance beliefs have been suggested as important factors for the classification and treatment of patients with LBP. However, the degree to which they are associated with outcome after intervention is uncertain.
Methods: There were 42 subjects with activity limiting LBP for more than 2 months enrolled in an exercise based, physical therapy program. A multidimensional test battery was completed before and after a 10-week program of lumbar extensor muscle strengthening. Correlational analyses, independent t tests, and validity indexes were used to determine relationships of preintervention measures of anticipated pain and the fear avoidance beliefs to clinically meaningful improvements in the Roland-Morris score. Intention-to-treat strategies were used to account for study dropouts.
Results: A total of 36 subjects completed the 10-week intervention. The lack of clinically meaningful outcome, as determined by a failure-to-report a minimum of 16% decrease in the Roland-Morris score, was associated with high preintervention scores on the physical activity subscale of the Fear Avoidance Beliefs Questionnaire (FABQP). Subjects with a preintervention score > or =29 on the FABQP had a likelihood ratio of 3.78 (95% confidence interval [CI] 2.4-5.16) for an increased probability of negative outcome after initial testing when compared to those subjects with low (<20) scores. The sensitivity and specificity of low scores to predict clinically meaningful outcomes compared to those with high scores were moderate (sensitivity = 0.87 and specificity 0.77); however, the likelihood ratio was inconclusive. Anticipated pain was significantly correlated with but was higher than the reported pain during activity both before and after intervention but not predictive of outcome.
Conclusions: In a sample of people from a Middle Eastern culture undergoing exercise intervention for LBP for which they are not receiving workers' compensation, the preintervention physical activity subscale of the FABQ is predictive of negative outcome when the observed scores are > or =29. Despite significant improvements in all variables after intervention, anticipated pain remained significantly higher than reported pain during physical performance testing but did not predict outcome.