Objective: To investigate side effects, medication compliance, and assumpti
on of medication assignment in adolescents taking imipramine versus placebo
in a clinical trial. Method: Sixty-three anxious-depressed adolescents in
an 8-week double-blind study of imipramine versus placebo, each in combinat
ion with cognitive-behavioral therapy for school refusal, were evaluated. M
easures of side effects, global improvement, family functioning, medication
compliance based on pill counts, and guesses of drug assignment (imipramin
e versus placebo) were analyzed. Results: Mean side effects ratings were si
gnificantly higher for the imipramine group compared with the placebo group
(p = .001), Side effects were not associated with noncompliance or with dr
opping out. Oppositional defiant disorder (ODD) in the adolescents was sign
ificantly associated with medication noncompliance (p = .036). On the Famil
y Adaptability and Cohesion Evaluation Scale II (FACES II), low family adap
tability (i.e., rigidity), low family cohesion (i.e., disengagement), and e
xtreme family type were significantly associated with greater noncompliance
with medications. Accuracy rates for guessing medication assignment (imipr
amine versus placebo) were 66% for subjects, 62.5% for mothers, and 79.5% f
or the psychiatrist. Logistic regression demonstrated that side effects (p
= .005) and global improvement scores (p = .06) predicted the psychiatrist'
s guesses of drug assignment. Conclusions: Side effects were not associated
with noncompliance. Nonadherence with taking medications was associated wi
th ODD in the adolescents and problematic family functioning on FACES II,Th
e psychiatrist, who was blind to treatment condition, guessed the subjects'
medication assignments with high accuracy. Thus, because of expectancy bia
s, the data support the use of blind independent evaluators for rating chan
ges in medication trials.