Prediction of chronic pain clinic outcome using psychometric and demographic predictors

Douglas Robert Lyman, Marquette University

Abstract

Chronic pain, any pain lasting longer than 6 months with an unknown prognosis, presents a considerable challenge to the sufferer. In response, pain clinics have been developed to assist those with chronic pain return back to a profitable and satisfying lifestyle. Research has shown that the pain clinic approach may be beneficial for some, but not all those referred to pain clinics benefit from treatment. This study was conducted to investigate factors that may have a relationship to pain clinic outcome. Subjects were 90 chronic pain patients seen at the Curative Rehabilitation Center and had received treatment in at least physical therapy and psychology. Two groups of variables were used to predict pain clinic outcome. These were demographic variables (age, length of time with pain, time since treatment, the number of treatment sessions and whether the patient received compensation for the painful condition) and psychometric variables (the Beck Depression Inventory, the Multidimensional Health Locus of Control, the McGill Pain Questionnaire). Outcome measures were taken from portions of the Curative Pain Clinic Questionnaire which sampled the current level of pain, pain medication usage, exercise, and, the time spent working, sleeping, socializing, laying down and sitting during the day. These variables were subjected to a alpha factor analysis yielding three factors. Factor scores were derived for each subject and served as the dependent variables in a stepwise multiple regression analysis. The results yielded significant simple and multiple correlations between the three outcome factors and scores on the Beck Depression Inventory and the compensation factor (p $<$.05). A double cross-validation was performed as method of replicating the findings. To further investigate the ability of the 11 aforementioned predictor variables to identify potentially successful and unsuccessful pain clinic patients, a discriminant analysis was conducted. Success was defined by a set of rules used in several previous related research studies. Results of the discriminant analysis indicates that the compensation factor, the Beck Depression Inventory and the McGill Pain Questionnaire (Sensation score) were 81% accurate in assigning patients to the successful or unsuccessful group. The clinical implications of this study and directions for future research were discussed.

This paper has been withdrawn.