Contrasting mediating variables in two 5-a-day nutrition intervention programs

Background. Researchers have advocated mediational analysis for behavioral intervention studies to link the supporting theory used in an intervention with the mediating variables and with its ultimate success or failure. Few mediational analyses have been reported for school-based nutrition studies. The conduct of mediational analyses within multi-site studies may provide advantages for the standardization of methods and for the replication and generalizability of findings. Methods. This study identified mediators of two school-based nutrition interventions for 4th graders. Three variables were tested on the four criteria necessary to establish mediation of intervention effects on changes in fruit and vegetable consumption (FVC) in 4th graders (Alabama, N = 1584; Minnesota, N = 522). FVC was measured in children using 24-h dietary recalls. Mediators were assessed using questionnaires completed by children and parents. Results. All criteria were met in Alabama for a single-item measure of knowledge of the 5-a-day daily consumption guideline. Knowledge and parent consumption satisfied one criterion in Minnesota. Knowledge accounted for 9.78% of the total intervention effect in Alabama. Conclusions. Knowledge of the 5-a-day guideline for fruit and vegetable consumption may mediate intervention effects. Future work should include tests of mediational models in multi-site studies. (C) 2004 The Institute For Cancer Prevention and Elsevier Inc. All rights reserved.