Les dernières années ont apporté de nombreuses innovations dans le domaine de la remédiation des troubles cognitifs participant à la genèse de l’altération du fonctionnement des patients atteints de schizophrénie. Parmi les programmes ciblant la cognition sociale à disposition des services de soins francophones, ToMRemed peut aujourd’hui bénéficier du retour d’expérience de la part des équipes (nous comptons 146 soignants formés en France) l’employant régulièrement, contribuant à déterminer ses principales propriétés et son impact. Utilisé en groupes de patients, animés par des soignants, ce traitement cible spécifiquement l’attribution d’intentions à autrui et les facultés métacognitives qui lui sont associées. L’harmonisation de la formation à l’usage de ToMRemed a contribué au recueil de données cliniques en soins courants de 110 patients pris en charge dans 11 centres en France. Nous rapportons ici les principales modifications cliniques et cognitives constatées indiquant une amélioration des performances mesurées à une échelle cognitive liée à la thérapie (Versailles – lecture intentionnelle en situation). Les données rapportées sont conformes aux améliorations attendues et rapportées dans la littérature. De plus, en l’absence de facteur prédictif clinique ou sociodémographique de l’amélioration cognitive, nous n’identifions pas de population de patients atteints de schizophrénie et présentant un trouble du fonctionnement social, échappant à l’indication d’un tel entraînement. Les limitations de la présente étude sont cependant à prendre en compte, notamment concernant l’absence de groupe témoin et la présence de biais de sélection (les patients inclus sont ceux qui ont bénéficié de l’ensemble du module ToMRemed et qui ont été réévalués ensuite).

Among the French programs targeting social cognition, ToMRemed benefits of accumulated clinical knowledge about its uses, properties, and impacts. ToMRemed is a group therapy supervised by two clinicians which aims at improving both the attribution of intentions to others and metacognitive skills. Ten therapy sessions plus preliminary and conclusion meetings are administered on a weekly basis. Five to seven patients with schizophrenia are allowed to participate in a group. Cognitive and metacognitive training consists in: 1) determining and discussing the indirect intentions of characters involved in short videos by taking into account context and social-emotional cues, 2) determining the most likely hypothesis about the character's intentions among a set of assertions, 3) updating metacognitive judgment's on hypotheses’ likelihood, while discussing with patients and clinicians. During the therapy, these three steps are referred to as “the ToMRemed strategy”. In the final part of each session, patients are asked to use this strategy while reporting a recent social situation in which they were involved and that was associated with another one's ambiguous intention.

In the present report, standardized practice among 11 psychiatric groups, in France, allowed gathering clinical data with an open before/after therapy design. One hundred and ten medicated and stabilized patients with schizophrenia (26 females and 86 males, mean age=38.4, SD=9.6) achieved the 3-months program (i.e. ten therapy sessions), and before/after theory-of-mind assessments. These skills were assessed with the V-SIR (Bazin et al., 2009), a video-based attribution of intentions test that is structured the same way than the videos of ToMRemed. In addition, 38 patients were assessed with the Positive and Negative Syndrome Scale (PANSS), 18 patients with the Peter Delusional Inventory 21 (PDI), and 21 patients with the Personal and Social Performance Scale (PSPS).

As expected, the main criteria (V-SIR score) decreased significantly from 19.7 to 17.4 (t=5.4; df=109; P<0.000001). Mean improvement of the scores increased with the initial impairment level of the V-SIR score. When divided into three groups, of equal sizes, based on their initial V-SIR scores, i.e. high, medium and low performers, a favorable change of this score was observed in respectively 54, 81 and 84 % of the patients. Concerning the secondary variables, significant improvements in the scores before and after the remediation program were found for PANSS positive, negative and general subscales, PSPS scores, but not PDI subscales. No significant correlation was found between V-SIR changes and sociodemographic, and aforementioned clinical variables.

The present study brings some evidence about the acceptability (for the patients and the clinicians) and the usefulness of ToMRemed as a way to improve social cognition and, potentially, real-life functioning. Even though the data presented here comes from a study of standard care with few methodological constraints, it is consistent with accumulated research evidence of the utility of social cognitive training in clinical practice. In addition, improvement of the attribution of intentions scores was also observed in the patients with low or medium initial performances, indicating that such therapy may be proposed to patients with most impaired social cognitive profiles. Last, the absence of significant correlation between V-SIR changes and clinical/functioning assessment does not provide evidence that ToMRemed would be more efficient with a subgroup of patients. Reports of field data are still rare and should be considered for their interest to promote and guide future investigation meeting high methodological standards. However, the methodological limitations of the present study are the absence of a control group and the selection bias (the subjects were selected once they completed all the sessions and V-SIR assessments).