The purpose of this study was to examine the association between characteristics of the symptomatology change curve (i.e., initial symptomatology, rate of change, and curvature) and final treatment outcome using a multilevel approach to account for therapist effects. The sample consisted of community clients (N = 492) seen by 204 student therapists at a training clinic. Linear, quadratic, and cubic trajectories of anxiety and depression symptomatology, as assessed by the Shorter Psychotherapy and Counselling Evaluation (sPaCE; Halstead, Leach, & Rust, 2007) were estimated. The multilevel quadratic trajectory best fit the data and depicted a descending curve (partial U-shaped). The quadratic growth parameters (intercept, slope, quadratic) were used as predictors of final treatment outcome as assessed by the Outcome Questionnaire-45.2 (OQ-45.2; Lambert et al., 1996). Patterns for two different successful outcomes, as conceptualised by (a) change in general symptomatology and (b) reliable improvement, were identified. For symptom change, successful outcomes followed a pattern of low initial levels of depression and anxiety, high initial rates of change, and high (flattening after initial drop) curvature, and the pattern applied to both within- and between-therapist levels. For reliable improvement at within-therapist level, successful outcomes followed a pattern of high initial rate of change and high curvature. For reliable improvement at between-therapist level, successful outcomes were associated with a pattern of low initial levels of depression and anxiety. Implications for clinical practice are discussed. (PsycINFO Database Record (c) 2018 APA, all rights reserved)