Purpose This study examines the rate and prediction of referral for specialized psychosocial oncology care in 326 patients with metastatic GI or lung cancer. Patients and Methods Referral information was abstracted from medical records and hospital databases. Patients completed measures of psychosocial and physical distress and functioning. Results Routine referral occurred in 33% of patients, and in 42% and 44%, respectively, of those scoring high on measures of depression (Beck Depression Inventory [BDI]-II ≥ 15) and hopelessness (Beck Hopelessness Scale ≥ 8). Univariate analyses indicated that referral was associated with younger age, unmarried status, living alone, presence of more depressive symptoms, hopelessness, and attachment anxiety, and with less social support, self-esteem, and spiritual well-being (all P < .05). Among the significantly depressed (BDI-II ≥ 15), 100% of those less than 40 years of age, but only 22% of those age 70 years or older were referred. Multivariate analyses indicated that referral was associated with younger age, unmarried status, and presence of more depressive symptoms. Moreover, increasing age was associated with a progressively lower likelihood of referral independent of the level of distress. Conclusion Routine referral of patients with metastatic cancer for psychosocial oncology care was predicted by presence of more severe depressive symptoms, younger age, and unmarried status. The rate of referral progressively declined with each decade of age, even among those with significant distress. These findings are consistent with some aspects of Andersen's model of health care utilization. The extent to which referred patients represent those who are most likely to benefit deserves further investigation.