By applying hierarchical linear modeling (HLM) techniques, patient clinical
characteristics at the beginning of treatment were used to predict individ
ual patient responses (N = 160) to psychotherapy. Four diagnostic groups (m
ood, anxiety, other, and no diagnosis) were formed among the patients based
on intake-administered Structured Diagnostic Interview for the Diagnosis o
f DSM-III-R axis I Disorders. Patients with mood and anxiety disorders had
predicted courses of response to psychotherapy that were similar but differ
ent from patients with other disorders and no diagnosis. Predicted and obse
rved courses of response to psychotherapy in a subsample (N = 75) who had p
rovided enough data to model the actual course of treatment showed high lev
els of congruence, thus supporting the validity of predicting course of res
ponse. HLM predictive profiling offers a new approach for assessing treatme
nt effectiveness of psychotherapy with patients having axis I diagnostic co
nditions by considering an individual patient's clinical characteristics.