Intelligent Decision Support

The 4 Diabetes Support System Project

There are over 422 million people in the world who have diabetes. From 5 to 10% of these people have type 1 diabetes (T1D), the most severe kind. In T1D, the pancreas fails to produce insulin, an essential hormone needed to convert food into energy. Therefore, T1D patients depend upon external supplies of insulin to live. T1D can not, at present, be prevented or cured; however, it can be treated and effectively managed.

The key to diabetes management is blood glucose control. Good blood glucose control can help delay or prevent serious long-term complications of diabetes, including blindness, amputations, kidney failure, strokes, and heart attacks. Achieving and maintaining good blood glucose control is a difficult task for patients, who must continuously monitor their blood glucose levels and daily activities. It is a difficult task for physicians, who must review large quantities of blood glucose and life event data, looking for problems and making therapeutic adjustments to correct them.

The goal of the 4 Diabetes Support System (4DSS) project is to: (1) automatically detect problems in blood glucose control; (2) propose solutions to detected problems; and (3) remember which solutions are effective or ineffective for individual patients. A case-based reasoning approach is used to capitalize on the extensive experiences of clinicians and T1D patients working together to manage the disease. An overview of the 4DSS system architecture is shown at left. The 4DSS could assist busy clinicians managing multiple T1D patients, and, once proven safe and effective, could eventually be embedded in medical devices or smart phones to provide low-risk advice to patients in real time.

Two other SmartHealth Lab projects, Machine Learning Models for Blood Glucose Prediction in Diabetes Management and A Consensus Perceived Glycemic Variability Metric, grew out of work on the 4DSS. In addition to their standalone capabilities, the blood glucose prediction models and glycemic variability metric are integral components of the situation assessment module of the 4DSS.