[ipaddr]

Dutch study (Steensma 2013) followed 127 children who were referred to the Gender Identity clinic in Amsterdam under the age of 12. The study looked to see if these children were still gender dysphoric by the time they reached adolescence at age fifteen. 37% of these children had persisted. However 64% of children had either desisted or were no longer traceable. Since there is only one Gender Identity clinic for children in Holland it can be assumed that the latter no longer required support and so had also desisted.

Most change their mind.

[a_shovel]

The "desistance" measured in that study is between going to the clinic as a child (before any treatment would be offered) and the time they would go to receive medical treatment from the clinic. Most sources I've seen suggest that change in gender identity is more common in younger children and drops off around the time puberty would start, which is the earliest that medical treatments are offered. The meaning of "desistance" being discussed here is regret after receiving medical treatment - basicly, the Type I error (false positive) rate of the treatment decision. These aren't really comparable.