Exploring the clinical effectiveness of a mobile health intervention for adolescent weight management

BACKGROUND: There are few evidence‐based mobile health interventions for use in the treatment of clinical obesity in adolescence. In addition, studies often use narrow inclusion criteria to recruit participants and are not representative of the patient demographic presenting to secondary and tertiary care paediatric centres. AIM: The aim of this study is to assess the effects of an experimental remote telemedical system in reducing body mass index standard deviation score (BMI SDS) after 1 year, compared to standard care at an accredited paediatric obesity hospital‐based service. METHODS: The study was approved by the hospital ethics committee (11‐033). Adolescents (12 to 17 years) with a BMI at or above the 98th percentile were recruited from the W82GO Child and Adolescent Weight Management Service in Children's Health Ireland at Temple Street, Dublin, Ireland. W82GO delivers evidence‐based obesity intervention via either group care or individual outpatient care. Baseline demographic information and primary and secondary outcomes were collected. Morbid obesity was defined as per Cole. Participants were offered six sessions of standard care (SC) and then randomized for maintenance treatment to SC or care delivered via a telemedical system and smartphone application (App). We aimed to randomize 134 participants between the two study groups. An intention‐to‐treat analysis compared changes in BMI SDS across time using a continuous time model. We assumed that data were missing at random and attempted to fit a degree 2 polynomial time effect, but only the linear term was significant at 5%. We also included morbid obesity, gender, the interaction between assigned treatment and time effects in the model. As the study was randomized, group means were set equal at baseline. RESULTS: The trial was discontinued due to challenges in patient accrual and retention. A total of 109 participants (40 boys) were recruited and randomized (55 to App). Of these, 48% were morbidly obese and no differences between the randomized groups were observed at baseline except for body weight (94.53 ± 21.20 kg in App vs. 85.08 ± 15.25 kg in UC, p < 0.01) and proportions with learning difficulties (42% in App vs. 24% in SC, p < 0.05) or previous developmental delay (29% in App vs. 13% in SC, p < 0.05). Follow‐up data were available for 65 participants (60%) at 12 months. Baseline differences between dropouts and completers were observed for BMI SDS (2.85 ± 0.46 vs. 3.12 ± 0.46, respectively, p < 0.01). Over 12 months, BMI SDS reduced by ‐0.10, (95% CI: ‐0.15 to ‐0.05, p = 0.0003) in all participants combined, but there was no differences between treatment groups. Over 12 months, for those with morbid obesity at baseline, the model predicted a reduction of BMI SDS from 3.44 to 3.34, and for those who did not have morbid obesity at baseline, it predicted a reduction from 2.66 to 2.56. CONCLUSION: A modest reduction in BMI SDS was seen over time for the trial sample as a whole, but no differences were seen between the two treatments. Future research is warranted to explore the impact of treatments on secondary outcomes.