Tracking of fruit, vegetables and unhealthy snacks consumption from childhood to adulthood (15 year period): does exposure to a free school fruit programme modify the observed tracking?

Background: The rationale for promoting increased consumption of fruit and vegetables (FV) at an early age is based on results from previous tracking-studies, indicating that dietary habits learned in childhood sustain into adulthood. Previous tracking studies have several limitations (e.g. low study sample, few repeated measurements and/or short a follow-up period). In addition, to our knowledge, no study has shown that a dietary intervention initiated in childhood affects tracking of dietary behaviour. The main objectives in this study were therefore to assess tracking of FV and unhealthy snacks in a large sample with multiple follow-up surveys over 15-years, and whether exposure to free school fruit for one school year modified tracking.
Method: The longitudinal cohort-study, Fruit and Vegetables Make the Marks, included 38 randomly drawn schools in Norway; nine intervention schools received free fruit (or vegetable) in the school year 2001/2002 and 29 schools severed as control. The baseline sample included 1950 subjects, and 16-92% participated at five follow-up surveys (2002-2016). FV consumption and unhealthy snacks were measured by FFQ. Mixed models were applied to estimate overall tracking coefficients, and to assess whether the intervention modified tracking ((from baseline, from follow-up one (while intervention was running) and from follow-up two (after end of intervention)).
Results: Overall tracking coefficients were 0.33 for fruit, 0.36 for vegetables and differed by sex for unhealthy snacks: 0.46 males and 0.39 for females (interaction p = 0.065). Most analyses showed no significant difference in tracking between the intervention group and control group. However, from follow-up one, tracking coefficients were different for unhealthy snacks, 0.46 vs. 0.38 (interaction p = 0.036), and from follow-up two for vegetables, 0.35 vs 0.48 (p = 0.036), in the intervention group and control group, respectively.
Conclusion: Our results indicate low to moderate tracking of FV and unhealthy snacks from childhood to adulthood. We found little evidence that the free fruit intervention modified tracking of fruit, vegetables or unhealthy snacks. More research is needed on if or how we can influence the tracking of fruit, vegetables and unhealthy snacks consumption to improve public health.