Understanding how different groups of patients change at different rates is important for treatment selection, planning and evaluation. This study aimed to assess whether an approach to classifying patients on the basis of initial symptom distress profiles (ISDPs) derived from a self-rated questionnaire measuring psychological distress may be useful in predicting treatment response. The Clinical Outcome in Routine Evaluation-Outcome Measure were collected from 1,095 first line mental health service patients (M [SD] age = 37.3 [14.3] years; 74% female) prior to every session. Latent profile analysis was performed on the questionnaires from the first session to classify participants into subtypes, which were then used to predict change rates. The best-fitting model identified 4 ISDP subtypes with significantly different treatment responses. Profile 1 predicted very slow change rate and indicated low initial distress coupled with low deviations among problem areas. Profile 2 predicted slow change rate with average initial distress and low emphasis on questions relating to risk of harming oneself and/or others. Profile 3 predicted fast improvement rate and showed high initial distress combined with low emphasis on the risk area. Profile 4 predicted moderate change rate and displayed very high initial distress accompanied with more emphasis on the risk area. Findings support the potential utility of ISDP subtypes to predict treatment response, suggesting that intake data that is easily collected by the clinician contain reliable information about treatment prognosis. The study is exploratory and needs to be replicated before stable conclusions can be drawn. (PsycINFO Database Record (c) 2019 APA, all rights reserved).