OBJECTIVES: The proportion of depressed individuals in usual practice who develop treatment resistant depression (TRD) is not known, in part because of varied definitions. One definition of TRD that has gained some consensus is the failure of two antidepressant trials of adequate dose and duration for the same depressive episode. The objective of this study was to operationalize this definition of TRD in administrative claims data and estimate the prevalence of TRD among individuals with major depression. METHODS: Retrospective analysis of PharMetrics Integrated Outcomes Database of adjudicated medical and pharmaceutical claims from multiple commercial U.S. health plans. We identified 168,533 adults (age 18-64) with a diagnosis of major depression (ICD-9-CM 296.20-296.39), treatment with an antidepressant, and with continuous enrollment for one year following the initiation of the antidepressant. TRD was identified based on the initiation of a third antidepressant after treatment with two antidepressants of adequate dose and duration. In sensitivity analyses, we varied the definition of “adequate duration” and the length of gaps between antidepressants. RESULTS: There were 12,125 individuals (7.2%) who started a third antidepressant after taking 2 different antidepressants for at least 42 days each, with at least one script of adequate dose, and a gap of less than 30 days between different antidepressant treatments. Increasing the gap between the ending of one antidepressant treatment and the initiation of the next to 60 days or 90 days resulted in identification of 16,494 (9.8%) and 19,273 (11.4%) cases of TRD, respectively. CONCLUSION: In usual medical care, 7.2% of individuals who are diagnosed with major depression appear to develop treatment resistant depression. For this small, but meaningful proportion of individuals with major depression, three or more antidepressant trials appear to be needed to sufficiently control symptoms.