In the Sequenced Treatment Alternatives to Relieve Depression (STAR*D) trial of almost 3000 patients with depression in the United States, 50% responded to the initial trial of a selective serotonin reuptake inhibitor antidepressant, but only a third achieved remission (nil or minimal depressive symptoms). The final remission rate, even after four potential treatment steps, was only 70%. This finding reflects the reality of clinical practice and highlights the need to employ the best available evidence in the management of people with complex depression. Before adopting a pharmacological strategy for a patient with difficult-to-treat depression, general clinical issues (such as missed psychiatric diagnoses, unresolved psychological issues and treatment non-adherence) should be considered. While there is no strong evidence for the order of implementing evidence-based pharmacological strategies for difficult-to-treat depression, we recommend: i) increase antidepressant dose; ii) switch to different antidepressant; iii) augment with a non-antidepressant agent; and iv) combine antidepressants. Sometimes it may be more appropriate to consider augmentation before switching antidepressants. The use of psychological interventions or other physical treatments such as electroconvulsive therapy should be considered at each step in management.