Rensselaer researchers have developed a simple algorithm designed to safely modulate insulin delivery and prevent hypoglycemia in outpatients using a continuous glucose monitor (CGM) and insulin pump. This technology applies a Kalman filter to the readings from a CGM to estimate the value and rate of change of the glucose level. This estimate is then used to predict current or impending low blood glucose levels. Then it suspends baseline insulin delivery if the algorithm identifies a risk of hyperglycemia.