Communication of randomization in childhood leukemia trials

Abstract

Context: Most children diagnosed as having leukemia become research subjects in randomized clinical trials (RCTs), but little is known about how randomization is explained to or understood by parents.

Objective: To investigate physicians' explanation and parental understanding of randomization in childhood leukemia RCTs.

Design and setting: A multisite study of the informed consent communication process for RCTs of childhood leukemia. Consecutive cases were recruited from pediatric oncology inpatient wards at 6 US children's hospitals associated with major academic medical centers from July 1, 1999, until December 31, 2001. The informed consent conferences were observed and audiotaped, and the information obtained was coded and analyzed. Parents were interviewed shortly after the conference to ascertain their understanding.

Participants: Parents and members of the health care team who participated in 137 informed consent conferences for children with newly diagnosed acute leukemia.

Main outcome measures: Observed explanations of randomization and parental understanding of randomization after the consent conference.

Results: Randomization was explained by physicians in 83% of cases and a consent document was presented during the conference in 95% of cases. Interviews after the conference demonstrated that 68 (50%) of 137 parents did not understand randomization. Parents of racial minority and lower socioeconomic status were less likely to understand randomization (P<.001 for each). Discussion of specific clinical trial details and the presence of a nurse during the conference were associated with understanding. Eighty-four percent of children were enrolled in a leukemia trial.

Conclusions: Despite oral and written explanation, half of the parents in this study did not understand randomization for childhood leukemia trials. To make informed consent more effective, future research must seek to improve communication during this critical interchange.