3:41 Patient subgroups identified by cluster analysis of Minnesota Multiphasic Personality Inventory scores and their relationship to spinal surgery outcome

Abstract

Purpose of study: Cluster analysis is a statistical method that identifies subgroups of a sample based on the similarity of subjects' responses to a set of variables. This type of analysis can be used to identify profiles of patients based on empirical groupings of scores. Previous studies have found that elevated scores on the hysteria (HY) and hypochondriasis (HS) scales of the Minnesota Multiphasic Personality Inventory (MMPI) are related to surgical outcome. The purpose of this study was to determine if distinct subgroups of patients could be identified by cluster analysis of MMPI scores and, if so, were such subgroups related to surgical outcome.Methods used: The MMPI-II was given to 204 patients considered as candidates for laminectomy/discectomy or spinal fusion. The group consisted of 87 men and 117 women. The mean age was 41.7 years (range, 18 to 72 years). A cluster analysis was performed on the MMPI scale scores. After subgroups of patients were identified, operative outcomes in the three groups were compared based on Oswestry Low Back Pain Disability Questionnaires, visual analog pain scales and medication use. These data were collected before surgery and at follow-up of minimum 6 months.of findings: Three distinct subgroups were identified by the cluster analysis. The largest group (n=114), had no significant elevations on any of the MMPI scales and thus were labeled the “normal” group. A second group (n=86) had elevated scores on the HY and HS scales, with lower but still significant elevations on the depression (D) and psychasthenia (PT; Obsessive-Compulsive) scales. This group was labeled as “pain sensitive.” The third group of patients (n=22) had elevations on 8 of the MMPI scales and were therefore labeled as “pathological.” These three groups identified by the cluster analysis had significantly different surgical outcomes. The normal and pain-sensitive groups both had statistically significant improvements in pain and Oswestry scores (p<.05, paired t test). The pathological group did not. With respect to Oswestry scores, the normal group improved a mean of 25.7%, the pain-sensitive group 20.8% and the pathological group had no improvement (worsened by 1%). With respect to medication use, the pathological group had a less favorable outcome than did the other two groups.Relationship between findings and existing knowledge: These results support previous findings that MMPI scale scores are related to surgical outcome. However, prior studies have primarily focused on the HY and HS scales, with some also including the D scale. The cluster analysis used in the current study supported the role of these traditionally investigated scales as well as an influence of the PT scale. Also, the analysis identified a subgroup at even greater risk of a poor surgical outcome, that is, the group of patients with elevations on eight of the scales.Overall significance of findings: The cluster analysis identified three distinct subgroups of patients based on MMPI scale scores. There were significant differences in the operative outcome in these three groups, indicating that the results of the cluster analysis were clinically relevant. The results supported earlier reports of the HY and HS scales being related to outcome, but inclusion of the D and PT scales, as well as the identifying of the “pathological” group may further refine the use of the MMPI in predicting outcome. Overall spine surgery outcome may be significantly improved by screening out “pathological” patients who may have a strong psychological tendency to be nonresponsive to surgery.Disclosures: No disclosures.Conflict of interest: No conflicts.