It is important to take into account the possible selection bias in
some aspects in our study. We have already mentioned that none of
the participating subjects had been adopted or accepted into foster
care from the hospital. Nonetheless, this may be explained in that
the addresses or telephone numbers of those who had been adopted
or taken into foster care were not available and thus, these subjects
would not be located. Other of the limitations of the present study is
the socioeconomic assessment. We used a test designed to establish the
socioeconomic status in children and adolescents who live with their
parents, however once finalized the study we observed that about a half
of the subjects lived independently of their families, so the FAS test
could have estimated their socioeconomic situation incorrectly.