Identification of subgroups of persons with chronic pain based on profiles on the pain stages of change questionnaire

Abstract

This study sought to identify reliable subgroups of patients with chronic pain based on profiles of subscale scores on the Pain Stages of Change Questionnaire (PSOCQ), a reliable and valid measure of individuals' readiness to adopt a self-management approach to chronic pain. The PSOCQ was administered to 633 people seeking treatment for chronic pain. Participants were predominantly White, averaged 48 years of age, about half were men, and about half reported back pain as the primary complaint. In a first study, cluster analysis was applied to 250 respondents. Five clusters were identified and named Precontemplation (11.0% of the sample), Contemplation (18.0%), Noncontemplative Action (12.4%), Participation (25%), and Ambivalent (33.6%). Results of a discriminant function analysis (DFA) on this sample, using the solution from the cluster analysis yielded a total error rate of 0.036. In a second study, the results of the first DFA were applied to an independent sample of 383 respondents in order to cross validate the solution from the first study. Cluster assignment proportions were very similar to the first sample and the posterior probability error rate for the second DFA was 13%. As predicted, clusters did not differ on measures of pain, disability, or demographics. Moreover, clusters differed significantly in theoretically consistent directions by scores on the Survey of Pain Attitudes, thus demonstrating criterion related validity for the clusters. Future research should examine the utility of PSOCQ profiles, relative to individual PSOCQ scale scores alone, in predicting response to self-management treatment approaches.