Studies have found that up to two-thirds of patients with major depressive disorder (MDD) do not fully respond to the first antidepressant. While switching antidepressants is a common strategy for antidepressant non-responders, there is still a lack of consensus about the optimal timing of a switch. Many clinicians wait for 6-12 weeks before considering a switch. The objectives of this paper are to (1) review the evidence for positive and negative predictive value (NPV) of early improvement at 2-4 weeks to predict final antidepressant response; (2) review randomized controlled trials (RCTs) that examine early switching strategies; and (3) provide future research directions and clinical recommendations for timing of antidepressant switching. We conducted a literature search for English-language studies via PubMed and Google Scholar, from 1984 to May 2013, with the following terms: 'antidepressants', 'MDD', 'time course', 'trajectory', 'early response', 'onset', 'delayed response', 'early improvement', 'predictors', 'switch', 'combination therapy', and 'augmentation'. Replicated evidence indicates that lack of early improvement (e.g. <20% reduction in a depression scale score) at 2-4 weeks can be an accurate predictor to identify eventual non-responders. The NPVs suggest that only about one in five patients with lack of improvement at 4 weeks will have a response by 8 weeks. Three RCTs examined early switch strategies, but results are inconsistent and comparisons limited by methodological differences. Future studies should incorporate a standard consensus definition of early improvement, discern whether the effect of early switching is specific to certain types of antidepressants, and determine whether early switch is superior to other strategies such as augmentation or combination. Notwithstanding these limitations, there is reasonable evidence to recommend earlier assessment for improvement. If there is no indication of early improvement at 2-4 weeks after starting an antidepressant, and taking into account other patient and clinical factors, a change in management can be considered.