Using modern risk engines and machine learning/artificial intelligence to predict diabetes complications: A focus on the BRAVO model.

Management of diabetes requires a multifaceted approach of risk factor reduction; through management of risk factors such as glucose, blood pressure and cholesterol. Goals for these risk factors often vary and guidelines suggest that this is based on patient characteristics and need to be individualized. Evaluating risk is therefore critically important to determine goals and choose appropriate treatments. A risk engine is an analytic tool that collects a large amount of population data allowing the simulation of the progression of diabetes with set equations over a period of time. Recently, a number of data cohorts have become available, leading to the development of newer risk engines that are more dynamic and generalizable. An example is the Building, Relating, Assessing, and Validating Outcomes in (BRAVO) diabetes model which was built on the ACCORD trial database. It is capable of accurately predicting diabetes comorbidities in an international population based on calibration with international clinical trial data. It has potential uses in risk stratification of patients, evaluation of interventions and calculation of their long term cost effectiveness. Recently, it has been used to simulate long term outcomes based on short term data, using difference modelling scenarios.