Joint state estimation prediction that evaluates differences in predicted vs. corresponding received data

Abstract

Systems and methods are provided for reconciling untrusted data of a subject using trusted data pertaining to the subject. Systems and methods are directed to evaluating differences in predicted data with respect to corresponding received data. Systems and methods estimate metabolic states from a combination of trusted and untrusted metabolic inputs, along with optionally using a personalized mathematical model with parameter optimization. Systems and methods provide for reconciled untrusted inputs with their measured impact of the glycemic signals that is consistent with a metabolic model. Estimation of future metabolic states for decision support and automated insulin dosing is enabled. Replay of scenarios with estimated or reconciled data is also provided.

What is claimed is: 
     
         1 . A method comprising:
 receiving, at an input estimator, untrusted data pertaining to a subject;   receiving, at the input estimator, trusted data pertaining to the subject;   reconciling, using an input reconciler, the untrusted data using the trusted data; and   outputting the reconciled untrusted data.   
     
     
         2 . The method of  claim 1 , wherein the untrusted data comprises at least one of timing of insulin, amount of insulin, meal data, activity data, a stream of data inputs, reported carbs wherein the reported carbs are unreliable or unavailable, or diabetes management data wherein the diabetes management data is estimated diabetes management data, wherein the untrusted data is untrusted because of behavioral anomalies or human error in at least one of timing, amount, estimation, or entry. 
     
     
         3 . The method of  claim 2 , wherein the trusted data comprises diabetes management data corresponding to the estimated diabetes management data, wherein the diabetes management data is received from a connected device or user entry. 
     
     
         4 . The method of  claim 1 , wherein the trusted data comprises at least one of CGM data, insulin pump data, computer generated data, computer generated models, or an individualized model that describes the glucose and insulin dynamics of the subject. 
     
     
         5 . The method of  claim 1 , further comprising predicting a future glucose state of the subject based on the reconciled untrusted data. 
     
     
         6 . The method of  claim 1 , further comprising receiving at least one of additional untrusted data or additional trusted data, reconciling the additional untrusted data using the trusted data, and reconciling the untrusted data using the additional trusted data. 
     
     
         7 . The method of  claim 1 , further comprising tuning an AP using the reconciled untrusted data. 
     
     
         8 . The method of  claim 1 , further comprising updating a behavior model of the subject using the reconciled untrusted data. 
     
     
         9 . The method of  claim 1 , further comprising determining that the untrusted data is unreliable or unknown, wherein determining that the untrusted data is unreliable or unknown comprises at least one of: (1) computing, using modeling, local variance of the untrusted data and comparing the local variance to the overall variance using the untrusted data to determine a comparison amount, wherein when the comparison amount is above a threshold, the untrusted data is determined to be unreliable or unknown, or (2) determining differences between the untrusted data and a model of trusted data. 
     
     
         10 . The method of  claim 1 , further comprising determining a credibility score for the untrusted data relative to trusted data. 
     
     
         11 . The method of  claim 1 , further comprising generating alerts pertaining to the subject based on the reconciled untrusted data. 
     
     
         12 . The method of  claim 1 , further comprising determining behavior patterns of the subject using the reconciled untrusted data. 
     
     
         13 . The method of  claim 12 , further comprising generating smart alerts pertaining to the subject based on the behavior patterns. 
     
     
         14 . The method of  claim 1 , further comprising comparing the untrusted data to the trusted data to identify a behavioral root cause of glycemic dysfunction. 
     
     
         15 . The method of  claim 1 , further comprising identifying a behavioral root cause of glycemic dysfunction using the reconciled untrusted data. 
     
     
         16 . The method of  claim 1 , wherein the reconciling comprises:
 receiving the untrusted data at the input reconciler, wherein the untrusted data comprises untrusted metabolic inputs;   receiving the trusted data at the input reconciler, wherein the trusted data comprises estimated untrusted metabolic inputs; and   combining the untrusted data and the trusted data using a weighting function to generate reconciled untrusted metabolic inputs.   
     
     
         17 . The method of  claim 16 , wherein the untrusted data and the trusted data received at the input reconciler are in the form of vectors, and wherein the reconciled untrusted metabolic inputs are in the form of vectors. 
     
     
         18 . The method of  claim 16 , wherein the weighting function is based on at least one of time-relevance or relative confidence of the untrusted data and the trusted data. 
     
     
         19 . The method of  claim 16 , wherein the untrusted data comprises reported untrusted metabolic inputs, and wherein the combining comprises reconciling differences between the reported untrusted metabolic inputs and the estimated untrusted metabolic inputs. 
     
     
         20 . The method of  claim 19 , wherein the reconciling comprises at least one of: (1) making the reported untrusted metabolic inputs and the estimated untrusted metabolic inputs consistent with a behavior model, or (2) reconciling differences between the amount and timing of the untrusted metabolic inputs and the estimated untrusted metabolic inputs with measured data. 
     
     
         21 . A method comprising:
 receiving, at an input estimator, untrusted data pertaining to a subject, wherein the untrusted data comprises user entered data comprising at least one of insulin data, meal data, or activity data; and   reconciling, using an input reconciler, the untrusted data using trusted data pertaining to the subject, wherein the trusted data comprises computer generated data.   
     
     
         22 . The method of  claim 21 , wherein the untrusted data comprises at least one of timing of insulin, amount of insulin, meal data, activity data, a stream of data inputs, diabetes management data wherein the diabetes management data is estimated diabetes management data, or reported carbs wherein the reported carbs are unreliable or unavailable, wherein the untrusted data is untrusted because of behavioral anomalies or human error in at least one of timing, amount, estimation, or entry, and wherein the trusted data comprises at least one of CGM data, insulin pump data, computer generated models, an individualized model that describes the glucose and insulin dynamics of the subject, or diabetes management data corresponding to the estimated diabetes management data, wherein the diabetes management data is received from a connected device or user entry. 
     
     
         23 . The method of  claim 21 , further comprising predicting a future glucose state of the subject based on the reconciled untrusted data. 
     
     
         24 . The method of  claim 21 , further comprising at least one of: (1) receiving additional untrusted data and reconciling the additional untrusted data using the trusted data, or (2) receiving additional trusted data and reconciling the untrusted data using the additional trusted data. 
     
     
         25 . The method of  claim 21 , further comprising at least one of (1) tuning an AP using the reconciled untrusted data, or (2) updating a behavior model of the subject using the reconciled untrusted data. 
     
     
         26 . The method of  claim 21 , further comprising determining that the untrusted data is unreliable or unknown, wherein determining that the untrusted data is unreliable or unknown comprises at least one of: (1) computing, using modeling, local variance of the untrusted data and comparing the local variance to the overall variance using the untrusted data to determine a comparison amount, wherein when the comparison amount is above a threshold, the untrusted data is determined to be unreliable or unknown, or (2) determining differences between the untrusted data and a model of trusted data. 
     
     
         27 . The method of  claim 21 , further comprising determining a credibility score for the untrusted data relative to trusted data. 
     
     
         28 . The method of  claim 21 , further comprising at least one of: (1) generating alerts pertaining to the subject based on the reconciled untrusted data, or (2) determining behavior patterns of the subject using the reconciled untrusted data. 
     
     
         29 . The method of  claim 28 , further comprising generating smart alerts pertaining to the subject based on the behavior patterns. 
     
     
         30 . The method of  claim 21 , further comprising comparing the untrusted data to the trusted data to identify a behavioral root cause of glycemic dysfunction. 
     
     
         31 . The method of  claim 21 , further comprising identifying a behavioral root cause of glycemic dysfunction using the reconciled untrusted data. 
     
     
         32 . The method of  claim 21 , wherein the reconciling comprises:
 receiving the untrusted data at the input reconciler, wherein the untrusted data comprises untrusted metabolic inputs;   receiving the trusted data at the input reconciler, wherein the trusted data comprises estimated untrusted metabolic inputs; and   combining the untrusted data and the trusted data using a weighting function to generate reconciled untrusted metabolic inputs,   wherein the untrusted data and the trusted data received at the input reconciler are in the form of vectors, and wherein the reconciled untrusted metabolic inputs are in the form of vectors, and wherein the weighting function is based on at least one of time-relevance or relative confidence of the untrusted data and the trusted data.   
     
     
         33 . The method of  claim 32 , wherein the untrusted data comprises reported untrusted metabolic inputs, and wherein the combining comprises reconciling differences between the reported untrusted metabolic inputs and the estimated untrusted metabolic inputs. 
     
     
         34 . The method of  claim 33 , wherein the reconciling comprises at least one of: (1) making the reported untrusted metabolic inputs and the estimated untrusted metabolic inputs consistent with a behavior model, and (2) reconciling differences between the amount and timing of the untrusted metabolic inputs and the estimated untrusted metabolic inputs with measured data. 
     
     
         35 . The method of  claim 21 , further comprising at least one of: (1) performing replay prediction using the reconciled untrusted data and the trusted data and outputting simulated metabolic states based on the replay prediction, or (2) performing real time prediction using the reconciled untrusted data and the trusted data and outputting simulated metabolic states based on the real time prediction. 
     
     
         36 . A system comprising:
 an input estimator configured to receive untrusted data pertaining to a subject and to receive trusted data pertaining to the subject; and   an input reconciler configured to reconcile the untrusted data using the trusted data, and to output the reconciled untrusted data.   
     
     
         37 . The system of  claim 36 , wherein the untrusted data comprises at least one of timing of insulin, amount of insulin, meal data, activity data, a stream of data inputs, reported carbs wherein the reported carbs are unreliable or unavailable, or diabetes management data wherein the diabetes management data is estimated diabetes management data, wherein the untrusted data is untrusted because of behavioral anomalies or human error in at least one of timing, amount, estimation, or entry. 
     
     
         38 . The system of  claim 36 , wherein the trusted data comprises diabetes management data corresponding to the estimated diabetes management data, wherein the diabetes management data is received from a connected device or user entry. 
     
     
         39 . A system comprising:
 an input estimator configured to receive untrusted data pertaining to a subject, wherein the untrusted data comprises user entered data comprising at least one of insulin data, meal data, or activity data; and   an input reconciler configured to reconcile the untrusted data using trusted data pertaining to the subject, wherein the trusted data comprises computer generated data.   
     
     
         40 . The system of  claim 39 , wherein the input reconciler is further configured to:
 receive the untrusted data, wherein the untrusted data comprises untrusted metabolic inputs;   receive the trusted data, wherein the trusted data comprises estimated untrusted metabolic inputs; and   combine the untrusted data and the trusted data using a weighting function to generate reconciled untrusted metabolic inputs,   wherein the untrusted data and the trusted data received at the input reconciler are in the form of vectors, and wherein the reconciled untrusted metabolic inputs are in the form of vectors, and wherein the weighting function is based on at least one of time-relevance or relative confidence of the untrusted data and the trusted data.