EVALUATION DE L'IMPACT DES MODALITES DE PRISES EN CHARGE HOSPITALIERES DE L'ANOREXIE MENTALE : EFFICACITE ET FACTEURS PREDICTIFS – EVALHOSPITAM

Anorexia Nervosa (AN) occurring in adolescents is associated with an extremely bad prognostic (mortality is 12 times higher than in same age normal individuals). Hospitalisation treatment programs vary a lot and they have never been compared to one another. Although these hospitalisations last long and are expensive, their respective efficiency remain questionable. Objectives: The main aim is to compare the efficacy of the treatments (psychological outcome , eating symptoms, somatic outcome, social adjustment).on 5 dimensions inpatient treatment programs for AN in adolescents (for such severe cases that hospitalisation is absolutely necessary) with a therapeutic study « here/there » (5 dimensions are : discharge BMI, separation or no separation,, eating rehabilitation present or not; level of family implication,, phase or not).. Moreover, in a largest sample (300 subjects) including also AN adults, we aim at determining outcome predictive factors in terms of the presence/absence of short term efficiency (discharge), in terms of clinical improvement duration (i.e., 12 months after hospitalisation), as well as predictors of treatment program expenses (i.e., hospitalisation duration). We will collect subject's and family opinion about treatment. Expected Results: Clinical improvement of AN due to hospitalisation will be quantified. At the end of hospitalisation and 12 months after hospitalisation, Poor and Good responders to hospitalisation program will be identified. The results may help to focus on specific therapeutic targets and to improve the patients prognostic, for whom there is a seriously increased risk of mortality and morbidity. Methodology : We will include in a prospective naturalistic study 300 AN recruited in one year in 11 different Psychiatric Department located in France. Patients are hospitalisation for maximal duration of one year and followed one year (study of 4 years). For practical and ethical reasons, patients can not be randomly assigned to one of the different treatment programs (Halmi 2005, Stober 2005). Moreover, given that the study focuses on severe AN cases (for which hospitalisation is necessary), there's no rationale for including a control group of AN outpatients. The first outcome criteria is the Morgan and Russell global outcome score. The secondary criteria are :BMI, psychological state, cognitive state , eating symptoms, somatic symptoms, social adjustment. Instead, the multivariate analysis analyses will include a propensity score adjustment procedure (D'Agostino, 1998), on all the various prognostic factors and complex variability factors . Using previously validated instruments, the patients and their relatives (parents) will be repetitively investigated, during and after hospitalisation, in order to estimate improvements of : eating and weight symptoms, somatic state, , mental state, cognitive processes, familial relationships, social adaptation As a whole, the recruitment procedure should allow to collect initial and 12 month follow-up data from 300 patients and their over a period of 3 years. The study, of such a cohort of 300 AN patients, also offers the opportunity to gather data for a future project on the evaluation of outcome predictive factors after several years of follow-up. We hope to find what are the better ways to treat theses patients and predictive factors will be new target for treatment in order to diminish the mortality.