DEVELOPMENT OF AN MMPI SUBSCALE AND A 3-ITEM RATING SCALE TO PREDICT OUTCOME OF INPATIENT MULTIDISCIPLINARY TREATMENT OF PAIN

Despite the relative success of multidisciplinary inpatient management of chronic pain, significant numbers of patients either do not improve or fail to maintain improvement. Program referrals continue to exceed capacity. This study addressed the problem of how to maximize successful outcomes and utilize pain center resources more efficiently. A number of patient characteristics and MMPI test data were evaluated as potential predictors of treatment outcome. The data were extracted from the records of 150 patients who underwent a four-week inpatient multidisciplinary chronic pain program at the University of Nebraska Medical Center between 1976 and 1982. The sample consisted of equal numbers of male and female patients with mixed pain complaints. An item analysis of the MMPI was performed to construct a 34-item Nebraska Pain Treatment Success scale. The scale achieved a Cronbach\u27s alpha of .5473, successfully classified nearly 90% of the 47 comparison and 45 criterion subjects, and proved to be the strongest predictor. The scale appears promising but awaits cross-validation. The most discriminating of the non-test independent variables were age, duration of pain history, complexity of pain complaint, employment status, and length of work disability. These findings are compatible with an operant conceptualization of chronic pain behavior. Cluster analysis has been used in previous research to identify and replicate homogeneous subgroups of pain patients using the standard clinical and validity scales of the MMPI. The next step in this research program is to evaluate replicated subgroups for differential response to alternative treatment packages. The present research is the first study known to have clustered the Wiggins content scales of the MMPI in an effort to identify homogeneous subgroups. No male subgroups were identified, but three distinct female subgroups clustered on 9 of 12 Wiggins scales. Future research employing larger more random samples is needed to further evaluate the usefulness of these scales for individualizing treatment and maximizing successful outcomes with this population