Predicting epilepsy surgery outcome in adult patients: May psychiatric diagnosis improve predictive models?

ABSTRACT

OBJECTIVE: With this study, we aimed to assess the importance of including psychiatric disorders in a comprehensive prediction model for epilepsy surgery. METHODS: Ambispective observational study with a sample of adults who underwent resective surgery. Participants were evaluated, before and one year after surgery, to collect data regarding their neurological and psychiatric history. The one-year post-surgical outcome was classified according to the Engel Outcome Scale. Previously identified predictors of post-surgical Engel Class were included in a logistic regression model. Then, the accuracy of alternative predictive models, including or excluding, past and current psychiatric diagnoses, were tried. RESULTS: One hundred and forty-six people participated in this study. The inclusion of psychiatric diagnosis resulted in a model with a higher AUC curve, however, the Delong method showed no significant statistical differences between the models. SIGNIFICANCE: Despite the fact that presurgical psychiatric diagnoses have shown to contribute to the prediction of epilepsy surgery outcome they do not contribute to a significant improvement of predictive models.