STUDY OBJECTIVE: We examine access to care for acute depression by insurance status compared to access for acute medical conditions in 9 metropolitan areas in the United States. METHODS: Using an audit study design, trained research assistants posing as patients referred from a local emergency department (ED) for treatment of depression called each clinic twice, with differing insurance status. The main outcome measure was the ability to schedule a mental health appointment within 2 weeks of the ED visit. RESULTS: In 45% of 322 calls to mental health clinics, the research assistant reached an answering machine compared with 8% of calls to medical clinics. As a result, only 31% of callers with depression vignettes were able to determine whether they could get an appointment versus 78% of callers with medical complaints. When they reached appointment personnel by telephone, 57% of depression callers successfully arranged an appointment (39% within 14 days). Among depression callers who reached appointment personnel, 67% of privately insured and 33% of Medicaid callers were able to make an appointment, for overall appointment rates of 22% and 12%, respectively. Appointment success for the uninsured was comparable to that of Medicaid patients. The high percentage of callers who encountered answering machines prevented us from completing the designed analysis of paired calls to individual clinics. CONCLUSION: Our findings indicate that the process for obtaining urgent follow-up appointments is systematically different for patients seeking behavioral health care than for those with physical complaints. The use of voicemail, in lieu of having a person answer the telephone, is much more prevalent in behavioral than physical health settings. More work is needed to determine the effect of this practice on depressed individuals and vulnerable populations.