Le but de cette étude fut d’établir des sous-groupes de profils psychopathologiques de patients douloureux chroniques à partir du Minnesota Multiphasic Personality Inventory-2 (MMPI-2), et de rechercher des variations sur les mesures d’intensité et de répercussions de la douleur à l’aide d’échelles visuelles analogiques. Les recommandations pour la pratique discutées pour les sous-groupes décrits ont pour but d’améliorer les décisions thérapeutiques. Une méthode hiérarchique en clusters a montré une première solution en trois sous-groupes : le premier dans les limites de la normale, le deuxième en « V de conversion », et le troisième avec une élévation générale du profil. Des différences significatives ont été retrouvées entre les sous-groupes d’élévation générale et des limites de la normale sur les échelles d’activité, de tristesse et de dépression. La seule différence significative entre le sous-groupe dans les limites de la normale et de « V de conversion » concernait l’échelle d’activité. Deux autres solutions en quatre puis en cinq clusters ont individualisé au total six sous-groupes psychopathologiques qui seront discutés dans les recommandations. Ces résultats montraient tout d’abord l’importance de différencier des profils psychopathologiques, qu’il n’existait aucune différence significative d’intensité ou de durée de la douleur entre les sous-groupes, et que les patients rapportant plus de répercussions sur les échelles visuelles analogiques avaient des profils plus pathologiques. De plus, nous avons insisté sur les recommandations thérapeutiques pour chacun des six profils psychopathologiques décrits.

The first aim of this paper was to analyze the Minnesota Multiphasic Personality Inventory (MMPI)-2 profiles of chronic pain patients using a hierarchical cluster method, in order to find subgroups of patients with differing psychopathological profiles and then to assess variations in the intensity and repercussions of pain among the subgroups. The Visual Analogue Scale (VAS) is a measurement technique commonly used in the assessment of intensity and repercussions of pain in chronic pain patients. The second aim was to examine the relationship between VAS measures of repercussions and scale scores on the MMPI-2. The final goal was to give recommendations to improve therapeutic decision-making for each patient. The best-known classification, the P.A.I.N. cluster typology, is based on the first version of the Minnesota Multiphasic Personality Inventory (MMPI). Keller and Butcher reviewed the data on patients with chronic pain obtained from the first version of the MMPI. In the same publication, they used a hierarchical cluster analysis, basing their process of agglomeration on the MMPI clinical scales of the second version (MMPI-2) to classify the patients into different groups whose members shared characteristics on the MMPI clinical scales. In Table 1, we have summarized and classified cluster labels used in the literature with the MMPI-2 since Keller and Butcher's MMPI-2 work.

Patients (148) went into hospital for a one-day assessment. The assessment included the validated French version of the MMPI-2 and Visual Analogue Scale for pain intensity and repercussions. The local ethics committee approved this study, conducted in agreement with the declaration of Helsinki. We used a hierarchical cluster method to establish subgroups of patients with different psychopathological profiles and evaluated variations between the subgroups on measures of intensity and repercussions of pain.

Figs. 1 to 2 show scale profiles for three- and four-cluster hierarchical analyses. A three-cluster solution best fitted the data (Fig. 1 and Table 2). The three clusters, which were of approximately equal size, were: (1) “Within Normal Limits”; (2) “Conversion-V”; and (3) “General Elevation”. We found significant differences between the “General Elevation” and “Within Normal Limits” clusters on the VAS measures of activity, sadness and depression. The only significant difference between the “Within Normal Limits” and “Conversion-V” clusters was on the activity scale. Patients in the “General Elevation” cluster were in the most urgent need of psychiatric help – it could be called a high distress, high disturbance cluster as Baker has described it. The four-cluster analysis of the MMPI-2 clinical scales (Fig. 2 and Table 2) gave four patterns showing the same “Conversion-V” and “Within Normal Limits”, but the “General Elevation” divided into a “Hypochondriac Depressive” cluster with 2-D scale elevated but with 1-Hs at the same level as the depressive scale, and the rest of “General Elevation” cluster could be split in a five-cluster solution into two very small subgroups, a “Psychotic V” cluster with a true V-shaped psychotic pattern with scores above those of the V-shaped neurotic triad scales, and a “Depressive General Elevation” cluster where the 2-D scale was the highest score in the profile but with all other scales also elevated except 9-Ma

These results showed that, firstly, different profiles of psychopathology were found even though there were no significant differences between the clusters in the level of pain intensity and pain duration; secondly, patients reporting more repercussions on the VAS measures showed profiles with more psychopathology. They were the patients most in need of psychological treatment. Finally, these results and the recommendations given could improve the therapeutic decision for each patient's subgroup.