To facilitate the improvement of future interventions, it's important to know the determinants of healthy behaviors. Our aim was to determine the predictors of healthy habits in a school based intervention study to promote healthy diet and physical activity among schoolchildren in the region of Sousse, Tunisia.

It was a quasi-experimental intervention study with two groups: control and intervention group with pre–post evaluation of nutrition and physical activity intention and behavior in each group. The target population was composed with students aged 12–16 years schooled in colleges of Sousse in Tunisia. To evaluate the intervention, a sampling was used to include 2200 students who participated to the questionnaire. All the students of intervention group received a standardized program with information about healthy nutrition and physical activity. An Arabic pre-tested and auto-administered questionnaire was used to assess nutrition and physical activity intention and behavior before and after the intervention.

The intervention group's posttest knowledge and behavioral intention were significantly higher than the control group's posttest. No significant differences occurred in posttest attitudes between the control and intervention groups. A multivariate logistic regression models were used to identify the predictors of students’ “healthy behavior”.

This school based intervention improved eating and physical activity intentions and behaviors among schoolchildren. “Healthy behaviors” were determined by age, father's profession and home characteristics. This finding could direct future interventions to disadvantaged categories.