Computer-Implemented System And Method For Facilitating Accurate Glycemic Control By Modeling Blood Glucose Using Circadian Profiles

摘要

A computer-implemented system and method for facilitating accurate glycemic control by modeling blood glucose using circadian profiles is provided. Anti-hyperglycemic medications are categorized based on similar glucose lowering effects. A circadian profile for a diabetic patient is built by assembling meal period data including at least two recent typical measurements of pre-meal and post-meal self-measured blood glucose, identifying a dose of an anti-hyperglycemic medication, and identifying the class of the anti-hyperglycemic medication. A model of glucose management through the circadian profile is created by estimating expected blood glucose values and their predicted errors at each of the meal periods, visualizing the expected blood glucose values and their predicted errors over time for each meal period in a log-normal distribution, and selecting one of the meal periods and, for each anti-hyperglycemic medication in the identified class, and modeling a change in the dose of the anti-hyperglycemic medication.

主权项

1. A computer-implemented method for facilitating accurate glycemic control by modeling blood glucose using circadian profiles, comprising the steps of:
defining a plurality of meal periods that each occur each day; defining a plurality of classes that each class comprises a set of anti-hyperglycemic medications that has similar glucose lowering effects; building a circadian profile for a diabetic patient, comprising the steps of:
choosing an observational time frame for the circadian profile comprising a plurality of days that have occurred recently;assembling meal period data comprising at least two typical measurements of pre-meal and post-meal self-measured blood glucose that were recorded at each of the meal periods that occurred each day in the observational time frame;identifying a dose of an anti-hyperglycemic medication that was taken during each of the meal periods for which the blood glucose measurements were recorded; andidentifying the class to which the anti-hyperglycemic medication is comprised; and creating a model of glucose management through the circadian profile for the diabetic patient for each of the anti-hyperglycemic medications in the identified class, comprising the steps of:
defining a modeling period comprising a plurality of days, which each comprise the same plurality of the meal periods that occurred each day in the observational time frame;estimating expected blood glucose values and their predicted errors at each of the meal periods from the blood glucose measurements based on the meal periods in the circadian profile;visualizing the expected blood glucose values and their predicted errors over time for each meal period in a log-normal distribution; andselecting one of the meal periods and, for each anti-hyperglycemic medication in the identified class, modeling a change in the dose of the anti-hyperglycemic medication for the selected meal period, comprising the steps of:
obtaining glucose lowering effect of the modeled change in the dose of the anti-hyperglycemic medication and glucose lowering effects of other anti-hyperglycemic medications in the determined class;normalizing the glucose lowering effect of the modeled change in the dose of the anti-hyperglycemic medication with the glucose lowering effects of the other anti-hyperglycemic medications in the determined class;propagating the normalized glucose lowering effect over time for the modeled change in the dose of the anti-hyperglycemic medication to the expected blood glucose values, beginning with the selected meal period and continuing with each of the meal periods occurring subsequently in the modeling period, the normalized blood glucose lowering effect being adjusted in proportion to the set time of each subsequent meal period until the normalized blood glucose lowering effect is exhausted; andvisualizing the expected blood glucose values as propagated and their predicted errors in the log-normal distribution, wherein the steps are performed on a suitably-programmed computer.