There are over 40 million diabetics in the United States, with a concomitant annual health care cost of over $100 billion, with 10-20% of this representing costs related to vascular complications of the disease. Current therapeutic methods, such as insulin treatment and glucose monitoring for insulin treatment, are expensive and invasive, with the cost of one blood glucose monitor in the range of $50-300, and typically requiring at least 5 finger pricks per day for testing. Frequent monitoring is required because glucose levels can vary substantially throughout the day, and may often be quite low when the patient is sleeping or at work. Early detection of dangerously low glucose levels is critical, so that intervention measures may prevent more severe negative consequences, and so that the patient's health can be better managed. These issues are especially important to diabetic patients having impaired glucose regulation, defined as the inability to maintain glucose levels within certain normal ranges. These patients are at risk for impaired consciousness, hypoglycemia (low blood glucose) and hyperglycemia (high blood glucose). The ability to sense when a patient is approaching dangerous levels of hypoglycemia or hyperglycemia, or has already arrived at dangerous levels, and to then take preventive action can literally mean the difference between life and death for these patients. There is also increasing evidence that fluctuations in blood glucose levels (including hypoglycemia and hyperglycemia) during the day result in short-term fluctuation of blood pressure. It is known that diabetes itself is associated with abnormal hemodynamics (including increased peripheral vascular resistance) and this relationship is exacerbated in subjects with impaired glucose regulation. However, there is a surprising lack of knowledge about the physiology of blood pressure and its relationship to hypoglycemia or hyperglycemia. There are almost no known data about the effect of glucose level on blood pressure over the day (day to day variability) or within the day (24 hour variability). Most of what is known is based on studies that measured very large population groups (over 100 subjects) and assessed blood pressure by one single blood sample obtained at the time that the glucose level was measured. In addition, glucose levels themselves may be responsible for some blood pressure variability. Some studies show a link between hypoglycemia and “crashes” (increased blood pressure variability) whereas others do not find such a link. Thus, while it is clear that both hypoglycemia and hyperglycemia can contribute to variability in blood pressure, what is not clear is the contribution of short-term glucose variability to short-term blood pressure variability. Despite advances in the treatment of diabetes, many diabetic patients are frequently subjected to the debilitating effects of unregulated glucose levels. This patent application describes a non-invasive device and method for measuring blood glucose levels and/or predicting blood glucose levels based on analyte concentration measurements of subcutaneous body fluids. The information derived from the described methods can be used to determine blood glucose levels and blood glucose levels trends in patients, for purposes of implementing, modifying or terminating a treatment regimen for these patients, as well as for research purposes. For example, the described information may be used by the patient, the physician, the healthcare provider, or a combination thereof, to monitor or assess treatment efficacy and safety. In some embodiments, analyte concentration values can be collected continuously, if desired. In some embodiments, analyte concentration values can be collected periodically, such as once a day. In some embodiments, the information may also be used to predict the future behavior of the subject and/or the subject's response to various medical interventions. Such predictions can be used to optimize a treatment regimen for the patient. In some embodiments, prediction information can be used to control the patient's treatment by prompting the subject to take medications or by modifying the subject's life style. In some embodiments, the non-invasive analyte concentration measurement and prediction method can be used to determine or predict the future level of analyte in a diabetic patient. In some embodiments, information regarding analyte concentration measurements may be used to determine a diabetes treatment strategy. In some embodiments, information regarding blood glucose levels may be used to determine or determine a diabetes treatment strategy. In some embodiments, information regarding blood glucose levels may be used to determine, improve or modify a diabetes treatment regimen. In some embodiments, information regarding blood glucose levels may be used to provide feedback to a patient for purposes of monitoring, tracking, or modifying a diabetes treatment regimen. In some embodiments, information regarding blood glucose levels may be used to provide feedback to a patient for purposes of predicting and monitoring insulin delivery and/or blood glucose levels in response to a calculated diabetes treatment regimen. In some embodiments, the method of this patent application can be used to determine the proper diabetes treatment regimen for a diabetic patient by determining the blood glucose levels, insulin response or other analyte levels or characteristics that can be obtained by the method. In some embodiments, the method may provide information about the glucose levels and/or glucose response to any treatment regimen and whether the subject will be adequately controlled. In some embodiments, the method may determine a need to switch or supplement the treatment regimen, for example, by adjusting insulin dose or type or changing the subject's diet. In some embodiments, the method may make automatic recommendations regarding the subject's treatment regimen. In some embodiments, the method may provide information to a healthcare provider to aid in making treatment decisions and treatment protocol decisions, including the determination of the proper dosage, type and timing of treatment, for a particular subject.