Effect of Children's Diabetes Education on HbA1c and Wellbeing in the United Kingdom: A Randomised Controlled Trial and Selective Meta-Analysis

Abstract

Background: Children's self-management of diabetes in the United Kingdom (UK) is poor. We aimed to estimate the effectiveness and cost-effectiveness of standardised, quality-assured self-management kits for children with type 1 diabetes, and to synthesise HbA1c outcomes with five other contemporaneous UK trials. Methods: The EPIC trial was a pragmatic trial with randomisation ratio of 2 intervention: 1 control in 11 paediatric diabetes clinics in England and Wales, with a qualitative process evaluation. We designed kits to empower children to achieve glycaemic control, notably by recording blood glucose and titrating insulin. Clinic staff offered these kits to children in routine consultations. The comparator was usual treatment. Trial outcome measures at 3 and 6 months: Primary - Diabetes PedsQL. Secondary - HbA1c; General PedsQL; EQ-5D; and healthcare resource use. Trial registration: ISRCTN17551624 We undertook a selective meta-analysis of HbA1c using Cochrane methods of 6 UK children's diabetes education interventions compared with treatment as usual. Findings Trial: Between February 2010 and August 2011 we validly randomised 308 children aged six to 18 years; 201 received the intervention. The process evaluation was conducted up to December 2013. Meta-analysis: 6 trials including 1083 intervention and 935 control. Trial: Of the five dimensions of the Diabetes PedsQL, Worry showed adjusted scores significantly favouring self-management kits at three months (mean child-reported difference = +5.87; standard error [SE] = 2.19; 95% confidence interval [CI] from +1.57 to +10.18; p = 0.008); but Treatment Adherence significantly favoured controls at six months (mean child-reported difference = -4.68; SE = 1.74; 95% CI from -8.10 to -1.25; p = 0.008). Furthermore, intervention children reported significantly worse changes between three and six months on four of the five Diabetes PedsQL dimensions and thus on the total score (mean difference = -3.20; SE = 1.33; 95% CI from -5.73 to -0.67; p = 0.020). Moreover, there was no evidence of change in HbA1c; indeed only 18% of participants in each group achieved recommended levels at 6 months. Clinic staff reported no serious adverse reactions attributable to the intervention or its absence. Children's use of kits was poor. Few children or parents associated blood glucose readings with better glycaemic control. The kits, which cost £185 each, alienated many children and parents. Meta-analysis: When combined in a meta-analysis, five other recent contemporaneous UK trials of various educational interventions also reported no benefit, and little rapport between children, their parents, clinic staff and the interventions. Interpretation: Standardised kits showed no evidence of benefit, inhibited self-management and increased worry. Similar results from five contemporaneous UK trials indicate a serious problem with how children's diabetes education and support is conceptualised and delivered. Future research should study relationships between children and professionals, and seek new methods of helping children and parents to manage diabetes Trial Registration Number: ISRCTN17551624. Funding: National Institute for Health Research. Declaration of Interest: Professor John Gregory has received payments from Pfizer, Bayer and Ipsen for lectures, development of educational presentations and travel/accommodation to attend scientific meetings and advisory board meetings. His employer (Cardiff University) has also received funding from Novo Nordisk to support the development of patient-support materials used in the Development and Evaluation of a Psychosocial Intervention for Children and Teenagers Experiencing Diabetes (DEPICTED) research study. The remaining authors declare no competing interests. Ethical Approval: The Wales Multi-Centre Research Ethics Committee gave approval (08/MRE09/57).