Congress Abstracts 2005

20

HANGING CHILDHOOD CANCER SURVIVORS HEALTH BEHAVIORS TO MODIFY RISKS OF LATE EFFECTS. Cheryl Cox, RN, PhD, Rosemary McLaughlin, MSN, Shesh Rai, PhD, Brenda Steen, RN, CCRA, and Melissa Hudson, MD, St. Jude Children’s Research Hospital, Memphis, TN.

The late effects of therapy increase childhood cancer survivors’ risk of chronic health problems. Young survivors fail to engage in important health-promoting behaviors, and they practice risk behaviors at alarming rates. Interventions to target behavior change must be developed if there is to be a significant impact on reducing late effects.

A recent (2002) clinical trial failed to demonstrate a clear impact of an intervention on changing health behaviors in adolescent survivors. The study outcome was a single summative measure: risk behaviors comprised the lower end and health-protective behaviors comprised the higher end of a unidimensional scale. We re-examined these same data by separately evaluating the impact of the intervention on 5 health risk and 9 health protective behaviors.

The Health Belief Model guided the selection of study variables and the development of the intervention in the parent study. The Interaction Model of Client Health Behavior guided the secondary analysis through the re-configuration of study variables and their relationships.

The randomized trial compared 132 adolescent survivors in the intervention arm with 135 in the standard care arm at baseline and at 1-year follow-up relying on self-report and medical record data. Disease and treatment knowledge were compared against the medical record; late effects risk perceptions and health/risk behaviors were assessed on Likert scales (Cronbach’s alpha = 0.75 to 0.92).

Each of the 14 behaviors were evaluated separately in contrast to the parent study’s single summative measure; age, gender, and the wide variation in baseline behaviors were addressed within an ANCOVA model. Breast (p = 0.0001)/testicular (p = 0.004) self-examination increased as did perceptions about needing to change behavior (p = 0.004), effort needed to stay healthy (p = 0.0001), and knowledge (p = 0.014). Significant interactions between gender and treatment group were demonstrated.

The intervention’s impact was demonstrated by treating health-risk and health-protective behaviors separately rather than as a single summative measure. Gender and age influenced the intervention’s effects, suggesting that a “one-size fits all” approach to changing behavior is ineffective in adolescent survivors. The development of interventions to modify health behaviors in adolescent survivors must extend beyond cognitive processing theories and include important concepts of motivation and tailoring of intervention approaches.