In the 2006 article referenced in the post, E. Jane Costello, Alaattin Erkanli, and Adrian Angold conducted a meta-analysis of more than twenty academic studies published of children born from the mid-1960s to the mid-1990s. The criteria for the epidemiological studies they analyzed included the use of structured diagnostic interviews in making a formal diagnosis of depression.
In the period reviewed, the authors concluded, the evidence did not support the epidemic categorization. Their findings did not show any significant increase in depressive disorders, although the extent of such illnesses gave reason for concern. However, those factors did not add up to an “epidemic” categorization. They found it likely that there had been “heightened awareness of a disorder that was long under-diagnosed by clinicians,” which influenced popular belief in an increase.
Their research is useful in assessing the current state of depression among adolescents and younger children, up to a point. The first relevant issue is the time period, as their article appeared fourteen years ago. A 2019 publication by Susan D. Swick and Michael S. Jellinek addresses the issue of possible increase in child mental health disorders since the study in this question.
One of the difficulties they note is the relative lack and unevenness of data, with many estimates based on surveys done with parents rather than the children’s direct reporting or rigorously collected scientific evidence. They also note that the documented increase in treatment for psychiatric disorders may be understood as a positive development, as those affected are seeking and getting help, and does not necessarily correspond to an increase in the number of youth affected.