Grant: Catalyst Gr: Disease Prev./Risk Factor Modification - Non-communicable diseases

Childhood obesity tracks through adolescence and adulthood and has been forecasted to become the leading cause of cardiovascular disease in the next century. One in four children in Canada is currently overweight or obese, with children from ethnic minority and low-income households more likely to be impacted. Successful interventions to prevent childhood obesity exist (e.g. family-based interventions), but can be time and resource intensive. A nuanced approach is needed, where obesity prevention interventions can be appropriately targeted according to a child's risk of obesity. 'Clinical prediction tools' are statistical models that help doctors predict future events or poor outcomes. An accurate obesity risk prediction too which considers ethnicity and income will enable preventive interventions to be delivered to children who are most likely to benefit, while avoiding unnecessary interventions and cost for low risk children. The study will develop, test and validate a new clinical prediction tool designed to be used at the doctor's office through existing electronic medical record systems which will identify young children who are at high risk for childhood obesity in the future. We will create the Early Childhood Obesity clinical prediction tool (EChO) which will be used at 2 years of age to predict obesity at 6 years of age using two methods: traditional statistical techniques (EChO-T) and cutting-edge machine learning methods (EChO-M). We will assess how well EChO-T and EChO-M can predict childhood obesity in children across Ontario, and then specifically in children from ethnic minority and low-income households. EChO will help healthcare providers to determine which young children are most likely to benefit from early obesity prevention programs. EChO holds the potential to prevent childhood obesity, and its long term health consequences, minimizing future health care costs, and reducing obesity-related health inequality.