BACKGROUND: Improved cure rates for childhood cancer have produced a growing population of survivors at risk for late toxicities of chemotherapy and radiation therapy. Healthy behaviors can modify these risks. We initiated a controlled prospective trial to determine if a multi-component behavioral intervention could induce change in childhood cancer survivors' health knowledge, health perceptions, and practice of health-protective behaviors. PROCEDURE: Adolescent cancer survivors attending a long-term follow-up clinic were randomized to receive standard follow-up care or standard care plus the educational intervention. Baseline measures were obtained at randomization (T(0)) and repeated 1 year (T(1)) later during the survivors' annual check-up. RESULTS: Of 272 patients enrolled and randomized, 251 are evaluable at both time points. Treatment and control groups were similar in regards to diagnosis, gender, race, and age. The change in outcome measures over the year (T(1)-T(0)) was not significantly different between the two groups as assessed by a two-sample pooled t test. However, additional exploratory analyses indicated a significant gender difference in knowledge with female survivors in the intervention group having higher scores. In addition, patients who choose certain individual health goals, such as breast/testicular self-examination, showed improved practice of the health behavior. In addition, in a very exploratory analysis, a gender difference in response to the intervention was noted, with females exhibiting a greater improvement in knowledge scores than did males. CONCLUSIONS: Although the multi-behavioral educational intervention did not induce change in health knowledge, perceptions, and behaviors of childhood cancer survivors for the treatment group as a whole, gender differences and specific health goal differences were found. These findings suggest that future interventions should be tailored to reflect gender differences and the nature of the health goal under assessment.