Longitudinal data from the World Health Organization Psychological Problems in General Health Care study were used to examine the relationship between recognition and outcomes among depressed primary care patients. A representative sample of primary care patients at 15 sites completed a baseline assessment including the Composite International Diagnostic Interview (CIDI), the 28-item General Health Questionnaire (GHQ), and the Brief Disability Questionnaire (BDQ). The GHQ and BDQ were readministered after 3 months, and the GHQ, BDQ, and CIDI were readministered after 12 months. Of 948 patients with major depression at the baseline assessment, 42% were recognized by the primary care physician and given an appropriate diagnosis. Recognized patients were more severely ill (mean GHQ score 16.2 vs. 12.9, t = 5.44, p < 0.001) and more disabled (mean BDQ score 9.8 vs 8.2, t = 3.22, p < 0.001) at baseline. Recognized patients showed a significantly greater decrease in GHQ score at the 3-month assessment (6.1 vs 4.1, F = 5.33, df = 1, p = 0.02). At 12 months, recognized and unrecognized groups did not differ in either change in GHQ score or change in diagnostic status from baseline. Results were consistent across study sites. Our data suggest that recognition and appropriate diagnosis of depression in primary care is associated with significantly greater short-term improvement. The absence of a relationship between recognition and long-term outcomes may reflect limitations of this observational study. When considered along with other recent studies, these findings suggest that increasing recognition of depression in primary care is only a first step toward more appropriate treatment.