As minor depression is perhaps the most common form of mood disorder seen in primary care, we sought to explore the effects of both pharmacologic and psychosocial interventions for primary care patients with this condition. Three hundred and eighteen primary care patients meeting criteria for minor depression (defined as endorsing 3 or 4 of the nine DSM-IV symptoms of major depression, at least one of which was either depressed mood or anhedonia, for a period of at least four weeks, and scoring > or = 10 on the Hamilton Rating Scale for Depression) from 4 diverse geographic sites were enrolled in a randomized controlled 11 week trial of paroxetine, problem-solving therapy or placebo. Patients who attended at least 4 treatment sessions and who received a Hamilton score by an independent rater at either 6 or 11 weeks were used in the analysis (77% of enrolled patients). A score of < or = 6 on the Hamilton was defined as a positive response to treatment. Fifty four percent of patients met our criteria for remission (HRS-D < or = 6) by week 11, with no difference among treatments. Patients who were women, younger, of European descent, homemakers or retired persons (as opposed to unemployed) and who had lower baseline severity of depression were more likely to remit across all treatment conditions. Although explicitly addressed in the data analysis, differences in outcomes across the four sites of the investigations limit our confidence in the generalizabilty of our findings. In addition, patients with lower levels of educational attainment had a higher dropout rate, suggesting further caution about the generalizability of the findings. Defining remission in this categorical way, we found no differences among the interventions studied, but did find that outcome was related to demographic and clinical characteristics of the patients. While it is difficult to know why female patients were more likely to remit, this may be a function of the association in our subject populations between male gender and the likelihood (approximately .46) of being a patient in the VA system. The remaining variables associated with higher probability of remission appear to reflect social advantage and lower severity or complexity of illness.