In this article, Serretti et al. assessed the performance of a machine learning algorithm called XGBoost in predicting treatment resistant depression (TRD) using large, real-world clinical data. XGBoost is an implementation of gradient-boosted decision trees that is especially suited to this task due to its ability to handle missing data. The data sample used was collected within the “Group for the Study of Resistant Depression” research project and included three previously reported samples as well as a new one collected using the same protocol. XGBoost was trained on clinical and sociodemographic features from the data, with 80% of the dataset being used for training and 20% for testing. Secondary analysis was performed using different samples for training and testing to confirm generalizability of results. The XGBoost model had a ROC AUC of 0.8009, demonstrating that the model correctly differentiated between outcomes in about 80.09% of cases. Overall classification accuracy was 61%, and the model performed best for resistant patients, though performance was relatively balanced across all cases. Longer duration of current episode, longer duration of disease, higher age, and onset and number of hospitalizations were the most influential predictors. This study’s results support the idea that longer illness duration and more frequent depressive episodes contribute to TRD. The study also highlights the importance of other factors, such as BMI and depression severity.