Cocaine dependent outpatients (N = 252) were randomly assigned to one of th
ree individual treatments (cognitive, dynamic therapy, or drug counseling)
as part of the training phase of the National Institute of Drug Abuse Cocai
ne Collaborative Treatment Study. Patients and therapists completed the Hel
ping Alliance (HAq-II) and the California Alliance Scale (CALPAS), which do
cumented generally high alliance. Although patient reports of the alliance
did not predict outcome on drug related measures at the sixth month assessm
ent, it did at the one month assessment. Alliance, however, predicted impro
vement in depressive symptoms at six months. Overall, therapist ratings of
the alliance were less often predictive of outcome than patient ratings. Re
sults were similar across the completer and the intent-to-treat samples. Th
e CALPAS, but-not the HAq-II, predicted attrition. There was preliminary ev
idence that higher alliance predicted more change in depressive symptoms in
patients who remained for a longer time in treatment than for those who re
mained for a shorter time in treatment.