Demographic, personality, and disease-related factors all contribute when patients disagree with physicians on the severity of subjective symptoms. This study aims to create a model, on the basis of patient factors at treatment initiation, for longitudinal prediction of disagreement on treatment response and remission in depressed patients. Four hundred patients with major depressive disorder were studied during a clinical drug trial. Repeated assessments with the Montgomery–Åsberg Depression Rating Scale (MADRS) and the self-rating version (MADRS-S) were used to indicate response or remission. Factors at baseline and week 2 were tested for inclusion in a model for the prediction of discordance on remission and response between patients and physicians at week 8. The models were then tested, in the same population, at weeks 12, 16, and 24. Model AUCs ranged from 0.71 to 0.74 for week 8. The models that were validated at weeks 12, 16, and 24 indicated stability in the predictive value of the models. The risk for longitudinal disagreement in the evaluation of depression treatment response and remission in clinical practice and drug trials can be predicted using factors at study initiation and at week 2.