Reliability of cancer family history reported by parents in a case–control study of childhood leukemia
Menée auprès de 301 parents, cette étude cas-témoins analyse la fiabilité des données relatives à une histoire familiale de cancer rapportée par les parents d'un enfant atteint de leucémie lymphoblastique aiguë
Purpose To investigate the reliability of family history of cancer reported by parents of children with acute lymphoblastic leukemia (ALL) and parents of healthy control children. Methods A total of 301 parents were selected based on positive or negative family history of cancer at baseline, case–control status, and gender of the respondent (mother or father). Baseline responses were compared with responses at the second interview using the same questionnaire. Reliability was measured using proportion of overall agreement, Cohen’s kappa, and Cronbach’s alpha; a logistic regression model was also used to assess the role of the case–control status on overall agreement as the dependent variable. Results The overall agreement between interviews was high and similar for cases (85 %) and controls (86 %); there were no consistent effects of respondent gender, age at first interview, or time elapsed between interviews on agreement measures. Agreement measures did not materially vary according to whether respondents were reporting about their mothers, fathers, or siblings. Conclusions The study showed very good reliability of reporting family history among young parents of children affected with leukemia and parents of healthy control children.