Although associations with outcome have been inconsistent, therapist adherence and competence continues to garner attention, particularly within the context of increasing interest in the dissemination, implementation, and sustainability of evidence-based treatments. To date, research on therapist adherence and competence has focused on average levels across therapists. With a few exceptions, research has failed to address multiple sources of variability in adherence and competence, identify important factors that might account for variability, or take these sources of variability into account when examining associations with symptom change. Objective: (a) statistically demonstrate between- and within-therapist variability in adherence and competence ratings and examine patient characteristics as predictors of this variability and (b) examine the relationship between adherence/competence and symptom change. Method: Randomly selected audiotaped sessions from a randomized controlled trial of cognitive-behavioral therapy for panic disorder were rated for therapist adherence and competence. Patients completed a self-report measure of panic symptom severity prior to each session and the Inventory of Interpersonal Problems-Personality Disorder Scale prior to the start of treatment. Results: Significant between- and within-therapist variability in adherence and competence were observed. Adherence and competence deteriorated significantly over the course of treatment. Higher patient interpersonal aggression was associated with decrements in both adherence and competence. Neither adherence nor competence predicted subsequent panic severity. Conclusions: Variability and "drift" in adherence and competence can be observed in controlled trials. Training and implementation efforts should involve continued consultation over multiple cases in order to account for relevant patient factors and promote sustainability across sessions and patients. (Contains 3 tables and 2 footnotes.)