The authors declare they have no competing interests.
ADV was the project leader and managed all aspects of the study and together with NPS contributed to the write up of the study. NPS, ADV, CD and EVL conceptualized the study and contributed to all other aspects, including write up. CL and NG carried out the data analyses and contributed to the write-up. ZA, LD, MDW and JH participated in the intervention mapping process, in questionnaire development, and in fieldwork. All authors read and approved the final manuscript.
Availability of data and materials
Numerous studies in schools in the Western Cape Province, South Africa have shown that children have an unhealthy diet with poor diversity and which is high in sugar and fat. HealthKick (HK) was a three-year randomised controlled trial aimed at promoting healthy eating habits.
Sixteen schools were selected from two low-income school districts and randomly allocated to intervention (n = 8) or control school (n = 8) status. The HK intervention comprised numerous activities to improve the school nutrition environment such as making healthier food choices available and providing nutrition education support. Dietary intake was measured by using a 24-h recall in 2009 in 500 grade 4 learners at intervention schools and 498 at control schools, and repeated in 2010 and 2011. A dietary diversity score (DDS) was calculated from nine food groups and frequency of snack food consumption was determined. A school level analysis was performed.
The mean baseline (2009) DDS was low in both arms 4.55 (SD = 1.29) and 4.54 (1.22) in the intervention and control arms respectively, and 49 % of learners in HK intervention schools had a DDS ≤4 (=low diversity). A small increase in DDS was observed in both arms by 2011: mean score 4.91 (1.17) and 4.83 (1.29) in the intervention and control arms respectively. The estimated DSS intervention effect over the two years was not significant [0 .04 (95 % CI: −0.37 to 0.46)]. Food groups least consumed were eggs, fruit and vegetables. The most commonly eaten snacking items in 2009 were table sugar in beverages and/or cereals (80.5 %); followed by potato crisps (53.1 %); non-carbonated beverages (42.9 %); sweets (26.7 %) and sugar-sweetened carbonated beverages (16 %). Unhealthy snack consumption in terms of frequency of snack items consumed did not improve significantly in intervention or control schools.
The results of the HK intervention were disappointing in terms of improvement in DDS and a decrease in unhealthy snacking. We attribute this to the finding that the intervention model used by the researchers may not have been the ideal one to use in a setting where many children came from low-income homes and educators have to deal with daily problems associated with poverty.
The HK intervention did not significantly improve quality of diet of children.