Depression among injured workers: measurement and prediction

Abstract

Work-related injuries or disabilities result in significant negative consequences to physical, economic, social, and psychological well-being. Depression has been shown to increase post-injury and to contribute to poor return to work outcomes. Several authors have raised concerns about the use of instruments such as the Beck Depression Inventory-II (BDI-II; Beck, Steer, & Brown, 1996) with medically ill individuals, given the potential confounding of somatic items of depression and somatic items of medical illness/treatment. Furthermore, few studies have identified predictors of depression among injured workers. The goals of this study were to examine the reliability and factor structure of the BDI-II and to identify significant predictors of depression in an archived sample of 253 injured workers receiving vocational rehabilitation. Item reliability analysis, factor analysis, and MLR analyses were conducted to answer the research questions and test the hypotheses of the study. The BDI-II demonstrated excellent psychometric properties, with none of the somatic items worthy of removal. Factor analysis resulted in two correlated factors (Cognitive and Somatic-Affective), consistent with the factor structure found in previous factor analyses of "unhealthy" (i.e., mentally/medically ill) individuals. Several significant predictors of depression emerged in MLR analyses: pain, psychosocial stress, Accepting work value, and the interaction between pain and stress. In a stepwise MLR analysis, the four predictors explained 38% of the variance of depression scores, representing a "large" effect size (f2 = .62; Cohen, 1988). Several recommendations were offered based on the results. The BDI-II is recommended for use, without modification, when assessing depression among injured workers. Nonetheless, some clinicians and researchers may prefer to use the subscales to identify depressive symptomatology in greater detail. The significant predictors of depression should be routinely assessed. An equation was developed from the stepwise MLR that can be used to predict depression with relatively good accuracy. Future research with injured workers should further examine the validity and reliability of the BDI-II scores, identify other predictors of depression, identify the causal relationships among predictors and depression, use well-operationalized, standardized instruments, examine theoretical models, control for symptom exaggeration, and develop effective treatments that reduce depression among injured workers.