Eating disorders in children and adolescents might be the only psychiatric condition with some positive connotations in the Westernized ‘cult of thinness’. Nevertheless, these conditions are not free from stigmatization (Russell, 2004). Furthermore, the victims and their families suffer a lot, much developmental potential is not fulfilled, a huge burden of care is placed on both families and healthcare delivery systems (Nielsen & Bará-Carrill, 2003) and excess mortality in anorexia nervosa (AN) is well documented (Nielsen et al., 1998; Nielsen, 2001). Findings from a few centres of excellence deviate from the general pattern in that excess mortality is not reported for AN (Korndörfer et al., 2003; Crisp et al., 2006).

I will focus on publications reporting findings for children and adolescents, but I cannot avoid touching upon themes of a more general epidemiological nature. As always when you are synthesizing evidence, you are standing on the shoulders of giants. My major inspirations are Szmukler (1985), Hodes (1993), Hoek (1993), van't Hoff (1994), Fombonne (1995), Hsu (1996) to whom I shall refer the interested reader.

As the findings differ dependent on source population, screening instruments and diagnostic system, I shall be clear about the type of population behind the data presented here. Hoek (1993) defined the following sources of data – level 1 ‘community’; level 2 ‘total in primary care’; level 3 ‘conspicuous in primary care’; level 4 ‘total psychiatric patients’; level 5 ‘psychiatric inpatients only’.