Systems and methods for analyzing, interpreting, and acting on continuous glucose monitoring data
Inventors
Liu, Shiping • SHOMALI, Mansur • KUMBARA, Abhimanyu • Iyer, Anand • Peeples, Malinda • DUGAS, Michelle • CROWLEY, Kenyon • GAO, Guodong
Assignees
Interested in licensing this patent?
MTEC can help explore whether this patent might be available for licensing for your application.
Publication Number
US-11357427-B2
Publication Date
2022-06-14
Expiration Date
Abstract
Methods and devices include automated coaching for management of glucose states by receiving a user's glucose levels using a continuous glucose monitoring (CGM) device, determining a time in range (TIR) value, determining a TIR state, receiving a glucose variability (GV) value, determining a GV state, determining a starting state based on the TIR state and the GV state, determining that the starting state corresponds to a non-ideal state, generating an optimized pathway to reach an ideal state based on one or more account vectors such as addressing self-management behavior including food, activity, and medication use. The optimized pathway may further be based on computer detection and classification of significant events of interest over time.
Core Innovation
The invention provides a computer-implemented method and system for managing glucose states of a user by moving from a non-ideal state to an ideal state. The ideal state corresponds to a good time in range (TIR) state and a good glucose variability (GV) state, while the non-ideal state comprises at least one of a bad TIR state or a bad GV state.
A current TIR state is determined from a TIR value over a first period of time using a threshold band over a base time period. A current GV state is determined based on a GV value associated with the user’s glucose level, where the GV value indicates a standard deviation (SD) of glucose levels or a coefficient of variance (CV).
A plurality of optimization profiles is received for reaching the ideal state from the non-ideal state, and one optimization profile is identified based on one or more user vectors and one or more user attributes. An optimized pathway is then identified based on the identified optimization profile, the current TIR state, and the current GV state, where the optimized pathway comprises one or more adjustments to the one or more user vectors.
The optimized pathway is provided as a machine learning model output comprising a user vector change based on machine learning inputs. The machine learning model inputs further comprise user attributes including a medical attribute, a user preference, a metabolic attribute, and a user demographic, and the optimized pathway is further based on a habit index score determined from a cohort of users with one or more user attributes in common with the user.
Claims Coverage
Three independent claims define a shared approach for managing glucose states by computing a current good/bad TIR state and a current GV state, selecting an optimization profile using user vectors and user attributes, and generating an optimized pathway as a machine learning model output that includes medication, food consumption, and exercise adjustments conditioned on a habit index score.
Good/bad TIR state determination with threshold TIR and base time period
Determining a current TIR state based on a TIR value of the user’s glucose level over a first period of time, where the TIR value is based on an amount of time the user’s glucose level is within a threshold band over a base time period, and the current TIR state is one of a good current TIR state or a bad current TIR state.
Glucose variability state from SD or CV over base time period
Determining a current GV state based on a GV value associated with the user’s glucose level, where the GV value indicates a standard deviation (SD) of glucose levels or a coefficient of variance (CV), and the CV is variability of the user’s glucose level in view of a standard deviation of the glucose level over the base time period.
Optimization profile selection from user vectors and user attributes
Identifying one of the optimization profiles based on one or more user vectors and one or more user attributes, where the optimization profiles are for reaching an ideal state from a non-ideal state and are associated with the ideal state corresponding to a good TIR state and a good glucose variability (GV) state and the non-ideal state comprising at least one of a bad TIR state or a bad GV state.
Machine learning optimized pathway with user vector adjustments and habit index conditioning
Identifying an optimized pathway based on the identified optimization profile, the TIR state, and the GV state, where the optimized pathway comprises one or more adjustments to the one or more user vectors including a medication adjustment, a food consumption adjustment, and an exercise value, and where the optimized pathway is provided as a machine learning model output comprising a user vector change based on machine learning inputs including the TIR state and the GV state and further comprising user attributes including a medical attribute, a user preference, a metabolic attribute, and a user demographic, and is further based on a habit index score determined from a cohort of users with one or more user attributes in common with the user.
Across independent claims, the inventive coverage is centered on computing good/bad TIR and SD/CV-based GV states, selecting an optimization profile using user vectors and user attributes, and producing a machine learning model output optimized pathway that adjusts user vectors and is conditioned on a habit index score derived from a user cohort.
Stated Advantages
Not explicitly described in patent.
Documented Applications
Not explicitly described in patent.
Interested in licensing this patent?