The data from blood glucose meters are discontinuous, providing only an estimate of the "amount" of glucose. Continuous sensors provide time series data, including "amount," but adding direction, rate of change, and rate of acceleration. Currently, continuous sensor labeling requires a fingerstick confirmation for treatment decisions and for periodic calibration. Because of the added information, sensors may be as efficacious and safe as meters when used for treatment decisions but have not yet been tested in stand-alone mode. Glucose sensors are currently in the early adopter's phase of technology development. To extend to the broader diabetes patient population, improvements are needed in reliability, comfort, ease of use, and integration with other technologies. In order to achieve a closed loop (artificial pancreas), the "Holy Grail" of the diabetes device industry, improvements in the accuracy of continuous sensors are also required. The main source of sensor inaccuracy is the calibration process. Calibration updates during the sensor wear period are required because of the gross changes in local metabolism (during the formation of the foreign body response) around the sensors after insertion and during the entire period of wear. Ambulatory sensors are currently approved and used only to treat insulin-using diabetes patients. In the future, the technology may have benefit for diagnostic use and behavior modification in obesity, metabolic syndrome, and prediabetes, as well as metabolic monitoring in elite athletes, soldiers, pilots, and the like. Hospitals currently use single-point information for tracking glucose in patients. Blood-based continuous sensing may provide the data needed to prevent in-hospital hypoglycemia and may lead to closed loop technology in the hospital.