PubMed1994
OBJECTIVE: The aim of this study was to assess the ability of specific and clinically relevant Minnesota Multiphasic Personality Inventory (MMPI) profile types to predict outcomes in a structured interdisciplinary pain-management program for patients with low back pain. DESIGN: Subjects were divided into clusters representing MMPI profiles yielding similar clinical interpretation. Analyses of variance and chi-square testing assessed the effect of cluster group on a variety of outcome measures at pretreatment, posttreatment, and 6- to 66-month follow-up. Fisher's Least Significant Difference Test assessed the significance of differences between pairs of cluster groups. SETTING: A university-based comprehensive interdisciplinary pain-management program serving both inpatients and outpatients. PATIENTS: 122 subjects with chronic low back pain who completed the program, provided follow-up data, and fit into the definition of one of seven clusters. MAJOR OUTCOME MEASURES: Self-reports of subjective pain intensity, pain-related medication intake, and activity level at pretreatment, posttreatment, and follow-up; employment status at pretreatment and follow-up. RESULTS: Cluster groups did not differ significantly at any time on activity level and medication intake and differed on employment status only at pretreatment. There was a significant (p < 0.05) effect of cluster group on subjective pain intensity, but only two pairwise group comparisons were significant: subjects with a normal MMPI profile and those with no elevations except T = 71-80 on Hypochondriasis and Hysteria reported less pain at follow-up than did subjects with extreme elevations (T > 80) on both Hypochondriasis and Hysteria. CONCLUSION: Even when subjects with chronic pain are divided into cluster groups associated with highly similar clinical interpretations, the MMPI for the most part fails to predict self-reported outcomes in an interdisciplinary pain-management program.