Abstract
Introduction
The Distress and Risk Assessment Method (DRAM) is a tool for measuring the psychological risk factors in low back pain patients. In our previous study, the DRAM scale, its categories as well as its components (Zung Depression Scale (ZDS) and Modified Somatic Perception Questionnaire (MSPQ)) proved to be significant predictors for global treatment outcome in routine lumbar degenerative surgeries. On the other hand, the original DRAM has to be calculated from 45 items making that a long and clinically impractical tool. The aim of the study was to develop and validate a short form of the DRAM (sDRAM) and to determine the prognostic power of the new version in terms of the global treatment outcome.
Material and Methods
Classic Test Theory were applied for item reduction after the exploratory factor analysis of ZDS and MSPQ on the dataset of a large (N = 1837) cohort of surgically treated low back pain patients. Internal consistency and the correlation of the shortened DRAM (sDRAM) with the original scale were analyzed. After that, a two-step cluster analysis of sDRAM was applied to determine the cut-off scores. The prognostic value of the scale and its categories was tested on a prospective cohort of routine lumbar degenerative surgeries (N = 445) where global treatment outcome was assessed two years after the surgery by a Likert scale.
Results
The sDRAM consisted of five ZDS items (1, 6, 7, 12, 18) and three MSPQ items (3, 9, 11) preserving the original item-distribution ratio. sDRAM scale (range: 0–24) was highly correlated with the DRAM scale (range: 0–99) (r=0.88). Cluster analysis of the patients showed a four-cluster structure in the cohort similarly to the original questionnaire. Cut-off scores (sZDS > 6, sMSPQ > 2) were used to formulate the sDRAM categories (Normal, At risk, Distressed). The sDRAM scale showed a significant association with the global treatment outcome in a logistic regression model (p < 0.001, B=0.258, OR=1.3) and in ROC analysis (AUC=0.744, p < 0.001) and it was similar to the original DRAM (p < 0.001, B=0.066, OR=1.07 as well as AUC=0.72, p < 0.001). Risk for poor outcome was significantly associated with the suggested categories of sDRAM (6%, 20%, 39% of poor outcome in “Normal,” “At Risk” and “Distressed” subgroups, Chi-square=38.9, df=2, p < 0.001).
Conclusion
A shorter, clinically applicable method for the assessment of DRAM has been developed. The sDRAM consists of 8 items and correlates with the original questionnaire. sDRAM and its categories proved to be significantly associated with global treatment outcome. The tool can be used for preoperative risk assessment and reduction strategies, however, further validation is needed to show its utility in different, wider patient-groups.