A web- and mobile phone-based obesity prevention intervention in 4-year-olds : a population-based randomized controlled trial

Abstract: Background: Childhood overweight and obesity has increased significantly over the past two decades. Many well-conducted obesity prevention trials have been conducted in pre-school aged children but the majority have not been able to show changes in obesity related markers. These trials have used traditional face-to-face methods to conduct the interventions, which are expensive and difficult to scale up. Therefore, new dissemination methods for intervention studies such as mobile health (mHealth) should be explored. Aims: The overall aim of this thesis was to determine whether a mHealth intervention targeted towards parents could improve obesity markers in pre-school aged children. Paper I: To outline the study design and methodologies utilized in the MINISTOP trial. Paper II: To evaluate the validity of reported energy and food intake assessed using the mobile based Tool for Energy Balance in Children (TECH) against total energy expenditure (TEE) and 24hr dietary recalls, respectively. Paper III: To evaluate the capacity of the wrist-worn ActiGraph wGT3x-BT accelerometer to capture variations in free-living activity energy expenditure (AEE) and to assess wear compliance of the ActiGraph using a seven day 24hr protocol. Paper IV: To assess the effectiveness of the MINISTOP intervention on body composition, intakes of fruits, vegetables, candy, and sweetened beverages, as well as the amount of time spent sedentary and in moderate-to-vigorous physical activity after the 6-month intervention. Paper V: To investigate if the MINISTOP intervention 12-months after baseline improved fat mass index (FMI) and had a maintained effect on a composite score (made up of FMI as well as dietary and physical activity variables). Methods Paper II: A nested validation study including 39 children aged 5.5 years. Energy and food intakes were measured using TECH and compared to TEE assessed using the doubly labelled water method and 24hr dietary recalls, respectively. Paper III: A nested validation study including 40 children aged 5.5 years. TEE was assessed using the doubly labelled water method and AEE was calculated as TEE minus a predicted basal metabolic rate. The ActiGraph was worn on the non-dominant wrist and the utilized outputs were mean of daily filtered vector magnitudes (mean VM total) and mean of awake filtered vector magnitudes (mean VM waking). Papers IV and V: A randomized controlled trial including 315 children aged 4.5 years. After baseline assessments, the children were randomly allocated into the intervention or control group for six months. The intervention group and control group received the MINISTOP app or a pamphlet on dietary and physical activity behaviors for pre-school children, respectively. The outcome measures were FMI (primary) and intakes of fruits, vegetables, candy, and sweetened beverages, as well as time spent sedentary and in moderate-to-vigorous physical activity (secondary). Two composite scores, a seven component (including all primary and secondary outcomes) and a six component (including only secondary outcomes) were computed. Results Paper II: No significant difference between mean energy intake and TEE was found (P = 0.064). For all eight food groups assessed no significant differences in the mean intakes were observed when using TECH and 24hr dietary recalls and all intakes were correlated when using both methods (range for rho: 0.665 to 0.896, all P < 0.001). Paper III: Mean VM total and mean VM waking alone were able to explain 14% (P = 0.009) and 24% (P = 0.001) of the variation in AEE. When adding fat and fat free mass to the models 58% and 62% (P < 0.001) of the variation in AEE was explained, using mean VM total and mean VM waking, respectively. Paper IV: No intervention effect for the primary outcome FMI was observed between the intervention and control group (P = 0.922). At the 6-month follow-up, for the seven component composite score the intervention group significantly increased their score compared to the control group (+0.36 ± 1.47 units vs. -0.06 ± 1.33 units, respectively, P = 0.021 between groups), with the difference being more evident in children with a higher FMI. For the six component composite score the children in the intervention group had a higher odds of increasing their score in comparison to the control group (odds ratio: 1.99; 95% confidence interval: 1.20, 3.30, P = 0.008). Paper V: For FMI there was no significant difference observed between the intervention and control group (P = 0.566) between the 12-month follow-up and baseline. Furthermore, there was no maintained effect observed in the change in the difference in the seven component composite score between the intervention and control group (P = 0.248). Conclusions: The results from this thesis suggest that both TECH and the wrist-worn ActiGraph have the potential to provide useful information in studies where diet and physical activity in young children are assessed. Furthermore, this thesis presents results from the first mHealth obesity prevention study in pre-school aged children. Although no difference between the intervention and control group for FMI was observed, the intervention group showed a significantly higher seven component composite score difference than the control group at the 6-month follow-up, especially in children with a higher FMI. Topics for future research include modifications of the MINISTOP app to more specifically target high risk children as well as further studies on to how maintain behavior changes in mHealth interventions.