Using dissertations to examine potential bias in child and adolescent clinical trials