Diabetes management therapy advisor
First Claim
1. A method for determining a treatment dose for a treated patient, the method comprising:
- obtaining, at data processing hardware, training data for a plurality of patients of a patient population from memory hardware in communication with the data processing hardware, the training data comprising training glucose history data and training patient-state information for each patient of the patient population, the training glucose history data comprising treatment meal boluses of insulin administered by each patient of the patient population during a scheduled time interval within a day and an outcome attribute associated with each of the treatment meal boluses of insulin administered by the patients of the patient population during the scheduled time interval, the outcome attribute of the training glucose history data for each patient of the patient population comprising a mean glucose percent error based on a function of a sum of next scheduled glucose measurements and a glucose target range, each next scheduled glucose measurement for the corresponding patient of the patient population corresponding to a glucose measurement occurring after administration of a corresponding treatment meal bolus of insulin during the scheduled time interval;
processing, by the data processing hardware, the training data obtained for each of the plurality of patients of the patient population to train a predictive model capable of predicting treatment doses for the treated patient, the trained predictive model identifying, for each patient of the patient population, an optimum meal bolus of insulin associated with the scheduled time interval that yields the outcome attribute associated with bringing and maintaining a glucose level of the corresponding patient of the patient population closest to a glucose target center of the glucose target range;
receiving, at the data processing hardware, patient-state information for the treated patient;
determining, by the data processing hardware, using the trained predictive model, a next recommended treatment meal bolus of insulin during the scheduled time interval for the treated patient based on one or more of the identified optimum treatment meal boluses of insulin associated with patients of the patient population having training patient-state information similar to the patient-state information for the treated patient; and
transmitting the next recommended treatment meal bolus of insulin to a portable device associated with the treated patient, the portable device configured to display the next recommended treatment meal bolus of insulin during the scheduled time interval.
5 Assignments
0 Petitions
Accused Products
Abstract
A method includes obtaining training data for a plurality of patients of a patient population. The training data includes training blood glucose history data including treatment doses of insulin administered by the patients of the patient population and one or more outcome attributes associated with each treatment dose. The method also includes identifying, for each patient of the patient population, one or more optimum treatment doses of insulin from the treatment doses yielding favorable outcome attributes. The method also includes receiving patient-state information for the treated patient, determining a next recommended treatment dose of insulin for the treated patient based on one or more of the identified optimum treatment doses associated with the patients of the patient population having training patient-state information similar to the patient-state information for the treated patient, and transmitting the next recommended treatment dose to a portable device associated with the treated patient.
-
Citations
24 Claims
-
1. A method for determining a treatment dose for a treated patient, the method comprising:
-
obtaining, at data processing hardware, training data for a plurality of patients of a patient population from memory hardware in communication with the data processing hardware, the training data comprising training glucose history data and training patient-state information for each patient of the patient population, the training glucose history data comprising treatment meal boluses of insulin administered by each patient of the patient population during a scheduled time interval within a day and an outcome attribute associated with each of the treatment meal boluses of insulin administered by the patients of the patient population during the scheduled time interval, the outcome attribute of the training glucose history data for each patient of the patient population comprising a mean glucose percent error based on a function of a sum of next scheduled glucose measurements and a glucose target range, each next scheduled glucose measurement for the corresponding patient of the patient population corresponding to a glucose measurement occurring after administration of a corresponding treatment meal bolus of insulin during the scheduled time interval; processing, by the data processing hardware, the training data obtained for each of the plurality of patients of the patient population to train a predictive model capable of predicting treatment doses for the treated patient, the trained predictive model identifying, for each patient of the patient population, an optimum meal bolus of insulin associated with the scheduled time interval that yields the outcome attribute associated with bringing and maintaining a glucose level of the corresponding patient of the patient population closest to a glucose target center of the glucose target range; receiving, at the data processing hardware, patient-state information for the treated patient; determining, by the data processing hardware, using the trained predictive model, a next recommended treatment meal bolus of insulin during the scheduled time interval for the treated patient based on one or more of the identified optimum treatment meal boluses of insulin associated with patients of the patient population having training patient-state information similar to the patient-state information for the treated patient; and transmitting the next recommended treatment meal bolus of insulin to a portable device associated with the treated patient, the portable device configured to display the next recommended treatment meal bolus of insulin during the scheduled time interval. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12)
-
-
13. A system comprising:
a dosing controller including data processing hardware and memory hardware in communication with the data processing hardware, the dosing controller; obtaining training data for a plurality of patients of a patient population from memory hardware in communication with the data processing hardware, the training data comprising training glucose history data and training patient-state information for each patient of the patient population, the training glucose history data comprising treatment meal boluses of insulin administered by each patient of the patient population during a scheduled time interval within a day and an outcome attribute associated with each of the treatment meal boluses of insulin administered by the patients of the patient population during the scheduled time interval, the outcome attribute of the training glucose history data for each patient of the patient population comprising a mean glucose percent error based on a function of a sum of next scheduled glucose measurements and a glucose target range, each next scheduled glucose measurement for the corresponding patient of the patient population corresponding to a glucose measurement occurring after administration of a corresponding treatment meal bolus of insulin during the scheduled time interval; processing the training data obtained for each of the plurality of patients of the patient population to train a predictive model capable of predicting treatment doses for the treated patient, the trained predictive model identifying, for each patient of the patient population, an optimum meal bolus of insulin associated with the scheduled time interval that yields the outcome attribute associated with bringing and maintaining a glucose level of the corresponding patient of the patient population closest to a glucose target center of the glucose target range; receiving patient-state information for a treated patient; determining, using the trained predictive model, a next recommended treatment meal bolus of insulin during the scheduled time interval for the treated patient based on one or more of the identified optimum treatment meal boluses of insulin associated with patients of the patient population having training patient-state information similar to the patient-state information for the treated patient; and transmitting the next recommended treatment meal bolus of insulin to a portable device associated with the treated patient, the portable device configured to display the next recommended treatment meal bolus of insulin during the scheduled time interval. - View Dependent Claims (14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24)
Specification