EnglishTo the Editor In their meta-analysis,1 Dr Mata and colleagues combined point and period estimates from 54 studies and reported a 29% prevalence of “depression or depressive symptoms” among resident physicians. This is more than 4 times the 12-month general population period prevalence for major depressive disorder (MDD) in North America.2,3 EnglishMs Levis and colleagues comment on our finding of high levels of depression and depressive symptoms among residents—based primarily on data from self-report inventory studies—and question whether these levels are truly elevated compared with criterion-based MDD in the general population. We included multiple subanalyses to assess whether the depression found in residents was due to inaccurate self-report inventories rather than a true increase in pathology. In a subanalysis examining 7 prospective studies on more than 4200 residents using the same inventory to assess depressive symptoms before and during residency, we found a median absolute increase in symptoms of 15.8% (relative risk, 4.5) with the onset of training. This analysis, which controlled for inventory and cutoff score, showed that the prevalence of preresidency depressive symptoms was essentially the same as that of the general population and suggests that the depression levels identified in the primary studies were an effect of residency. If Levis and colleagues were correct that elevated resident depression was an artifact of self-report inventories, then spuriously elevated depression levels would likely have appeared in the preresidency assessments as well. However, this was not the case.