Patterns of age of onset of anxiety disorders (EDSP; N = 3021). (Note: In phobias impairment was required among subjects aged 18 years or older; *Separation Anxiety Disorder was only assessed in a subsample at T1.)

The threshold problem in diagnosing children and adolescent. Diagnostic criteria define thresholds for diagnoses by specifying type and number of symptoms, the duration and persistence that the symptoms need to be present, and the clinical significance (A). Due to the current categorical classification system, being short of just one criterion (eg, only 2 instead of 3 symptoms, only 5 months’ instead of 6 months’ duration, only 3 instead of 4 days a week, all symptom criteria met but no distress or impairment reported) leads to non-diagnosis (or nonspecific classification as “Anxiety Disorder Not Otherwise Specified” or “Other Conditions That May Be a Focus of Clinical Attention”). The variation of symptoms over time and difficulties in retrospective assessment may negatively affect correct diagnosis. Therefore it is also crucial to take a lifetime approach to diagnosis (B). Mere cross-sectional assessment may lead to erroneous nondiagnosis based on transient alleviation of symptoms.

Assessing onset of anxiety disorders. The age of onset of anxiety disorders can be directly assessed by asking “When was the first time you experienced….” The retrospectively reported ages often reflect the syndrome, rather than disorder onset, and can be subject to recall bias. This fact is indirectly reflected by observations from longitudinal studies whereby different ages of onset are reported for the same condition at various assessment waves. Other sources of information on age of onset are prevalence estimates for disorders in aggregated age groups (mostly reported in cross-sectional studies). More reliably but also more rare, incidence reports from longitudinal studies (ie, the proportion of new cases in a defined time interval) provide insights into the disorder onset.

Cumulative Incidence of anxiety disorders (EDSP; N=3021). (Note: Percentages in the legends refer to the estimated cumulative incidence rate at age 33, *age 19 for Separation Anxiety Disorder which was only assessed in a subsample at T1; in phobias impairment was required among subjects aged 18 years or older.)

Assessing the course of anxiety disorders. Several approaches exist to study the course of anxiety disorders. Cross-sectional studies most frequently use retrospective age of onset and age of recency reports to calculate the duration of a condition in years. This approach assumes a continuous disorder course, and may thus overestimate the duration and chronicity because symptom-free intervals are not taken into account. Another indirect measure of disorder chronicity is the proportion of point to lifetime prevalence. The higher the proportion, the higher the chronicity. Because only categorical diagnoses are considered here (no symptomatic improvements below the diagnostic threshold), this may lead to underestimation of chronicity. Overall, cross-sectional studies allow for only crude estimations of course and chronicity of anxiety disorders. Longitudinal studies, in contrast, allow for a more realistic description of the course of a disorder. Taking a prospective approach, the proportion of individuals meeting or not meeting the criteria again at follow-up is frequently used to describe stability and remission. Considering only the full DSM-IV diagnostic level, higher remission rates are possible because improvements below the diagnostic threshold are not taken into account. Thus, the most valid way to describe the course of anxiety disorders is to consider also subthreshold or subsyndromal conditions.