IP Library › Granted Patent US 12,646,620
Granted Patent B2
US 12,646,620 · App. 17/096,763 · Granted Jun 2, 2026

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

Inventor: Stephen D. Patek (Charlottesville, VA)
Assignee: Dexcom, Inc.
G16H50/20A61B5/14532A61B5/486A61B5/4866A61B5/7221A61B5/7275A61B5/742A61M5/172G06F16/2365G16H10/20G16H10/60G16H20/13G16H20/17G16H20/30G16H20/60G16H40/67G16H50/30G16H50/50G16H50/70G16H70/20A61B5/1118
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 12,646,620
App. No.
17/096,763
Granted
Jun 2, 2026
Kind
B2
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.

Claims (42)

1 . 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 untrusted insulin data, untrusted meal data, or untrusted activity data, the untrusted insulin data, the untrusted meal data, and the untrusted activity data each being associated with timing data;

determining, at a credibility assessor, a credibility of the untrusted data;

receiving trusted data at the credibility assessor; and updating, at the credibility assessor, the credibility of the untrusted data using the trusted data;

reconciling differences between the trusted data and the untrusted data to generate reconciled data that includes reconciled insulin data, reconciled meal data, and/or reconciled activity data wherein the reconciled insulin data, reconciled meal data, and/or reconciled activity data are generated at least in part by (i) using a weighted average of the untrusted data and estimated untrusted data, the estimated untrusted data being estimated using the untrusted data, the trusted data and a mathematical model; or (ii) choosing the untrusted data to maximize a classification objective;

performing replay prediction or real time prediction using the trusted data and the reconciled insulin data, the reconciled meal data, and/or the reconciled activity data;

adjusting a therapeutic parameter employed for controlling glucose in a diabetic patient based on the updated credibility of the untrusted data;

optimizing an amount of basal insulin to be delivered to the subject using the adjusted therapeutic parameter; and

delivering the optimal amount of basal insulin to the subject using an insulin delivery device.

2 . The method of claim 1 , wherein determining the credibility of the untrusted data comprises: determining a first credibility based on at least one of a lack of completeness of the untrusted data or a lack of continuity of the untrusted data; determining a second credibility based on expected behaviors indicative of at least one of the lack of completeness of the untrusted data or the lack of continuity of the untrusted data; determining a third credibility based on artifacts in estimated inputs indicative of systemic unknown factors; and aggregating the first credibility, the second credibility, and the third credibility.

3 . The method of claim 2 , wherein determining the first credibility comprises using measured signals as an input, determining the second credibility comprises using the untrusted data and trusted data as inputs, and determining the third credibility comprises using the untrusted data, estimated untrusted data, and reconciled data as inputs.

4 . The method of claim 1 wherein reconciling differences between the trusted data and the untrusted data to generate reconciled data includes estimating values for the untrusted data using an individual mathematical model of metabolic processes to produce or consume glucose.

5 . A method comprising:

receiving, at an input estimator, first untrusted data pertaining to a subject, wherein the first untrusted data comprises user entered data comprising at least one of untrusted insulin data, untrusted meal data, or untrusted activity data, the untrusted insulin data, the untrusted meal data, and the untrusted activity data each being associated with timing data;

determining, at a credibility assessor, a credibility of the first untrusted data;

receiving, at the input estimator, second untrusted data pertaining to the subject;

determining, at a credibility assessor, a credibility of the second untrusted data;

updating, at the credibility assessor, the credibility of the first untrusted data using the second untrusted data or the credibility of the second untrusted data;

receiving trusted data at the credibility assessor; and

updating, at the credibility assessor, the credibility of the first untrusted data and the second untrusted data using the trusted data and reconciling differences between the trusted data and the untrusted data to generate reconciled data that includes reconciled insulin data, reconciled meal data and/or reconciled activity data and reconciled second data, wherein the reconciled insulin data, reconciled meal data, and/or reconciled activity data are generated at least in part by (i) using a weighted average of the untrusted data and estimated untrusted data, the estimated untrusted data being estimated using the untrusted data, the trusted data and a mathematical model; or (ii) choosing the untrusted data to maximize a classification objective;

performing replay prediction or real time prediction using the trusted data and the reconciled insulin data, the reconciled meal data, and/or the reconciled activity data and the reconciled second data; adjusting a therapeutic parameter employed for controlling glucose in a diabetic patient based on the updated credibility of the first and/or the second untrusted data;

optimizing an amount of basal insulin to be delivered to the subject using the adjusted therapeutic parameter; and

delivering the optimal amount of basal insulin to the subject using an insulin delivery device.

6 . The method of claim 5 , wherein determining the credibility of the first and/or second untrusted data comprises: determining a first credibility based on at least one of a lack of completeness of the first and/or second untrusted data or a lack of continuity of the first and/or second untrusted data; determining a second credibility based on expected behaviors indicative of at least one of the lack of completeness of the first and/or second untrusted data or the lack of continuity of the first and/or second untrusted data; determining a third credibility based on artifacts in estimated inputs indicative of systemic unknown factors; and aggregating the first credibility, the second credibility, and the third credibility.

7 . The method of claim 6 , wherein determining the first credibility comprises using measured signals as an input, determining the second credibility comprises using the first and/or second untrusted data and trusted data as inputs, and determining the third credibility comprises using the first and/or second untrusted data, estimated untrusted data, and reconciled data as inputs.

8 . 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 untrusted insulin data, untrusted meal data, or untrusted activity data, the untrusted insulin data, the untrusted meal data, and the untrusted activity data each being associated with timing data; and

a credibility assessor configured to: determine a credibility of the untrusted data, receive trusted data, and update the credibility of the untrusted data using the trusted data;

an input reconciler configured to reconcile differences between the trusted data and the untrusted data to generate reconciled data that includes reconciled insulin data, reconciled meal data, and/or reconciled activity data, wherein the reconciled insulin data, reconciled meal data, and/or reconciled activity data are generated at least in part by (i) using a weighted average of the untrusted data and estimated untrusted data, the estimated untrusted data being estimated using the untrusted data, the trusted data and a mathematical model; or (ii) choosing the untrusted data to maximize a classification objective;

a prediction engine configured to perform replay prediction or real time prediction using the trusted data and the reconciled insulin data, the reconciled meal data, and/or the reconciled activity data; a parameter estimator configured to adjust a therapeutic parameter employed for controlling glucose in a diabetic patient based on the updated credibility of the untrusted data; and

an insulin delivery device delivering an optimal amount of basal insulin to the subject, wherein the optimal amount of basal insulin is optimized using the adjusted therapeutic parameter.

9 . The system of claim 8 , wherein determining the credibility of the untrusted data comprises: determining a first credibility based on at least one of a lack of completeness of the untrusted data or a lack of continuity of the untrusted data; determining a second credibility based on expected behaviors indicative of at least one of the lack of completeness of the untrusted data or the lack of continuity of the untrusted data; determining a third credibility based on artifacts in estimated inputs indicative of systemic unknown factors; and aggregating the first credibility, the second credibility, and the third credibility.

10 . The system of claim 9 , wherein determining the first credibility comprises using measured signals as an input, determining the second credibility comprises using the untrusted data and trusted data as inputs, and determining the third credibility comprises using the untrusted data, estimated untrusted data, and reconciled data as inputs.

11 . A system comprising:

an input estimator configured to: receive first untrusted data pertaining to a subject, wherein the first untrusted data comprises user entered data comprising at least one of untrusted insulin data, untrusted meal data, or untrusted activity data, and receive second untrusted data pertaining to the subject; and

a credibility assessor configured to: determine a credibility of the first untrusted data, determine a credibility of the second untrusted data, update the credibility of the first untrusted data using the second untrusted data or the credibility of the second untrusted data, receive trusted data at the credibility assessor, and update the credibility of the first untrusted data and the second untrusted data using the trusted data;

an input reconciler configured to reconcile differences between the trusted data and the untrusted second data, untrusted insulin data, untrusted meal data and/or untrusted activity data to generate reconciled that includes reconciled second data, reconciled insulin data, reconciled meal data, and/or reconciled activity data, wherein the reconciled insulin data, reconciled meal data, and/or reconciled activity data are generated at least in part by (i) using a weighted average of the untrusted data and estimated untrusted data, the estimated untrusted data being estimated using the untrusted data, the trusted data and a mathematical model; or (ii) choosing the untrusted data to maximize a classification objective;

a prediction engine configured to perform replay prediction or real time prediction using the reconciled data and the trusted data, the reconciled data including the reconciled second data, the reconciled insulin data, the reconciled meal data, and/or the reconciled activity data;

a parameter estimator configured to adjust a therapeutic parameter employed for controlling glucose in a diabetic patient based on the updated credibility of the first and/or second untrusted data; and

an insulin delivery device delivering an optimal amount of basal insulin to the subject, wherein the optimal amount of basal insulin is optimized using the adjusted therapeutic parameter.

12 . The system of claim 11 , wherein determining the credibility of the first and/or second untrusted data comprises: determining a first credibility based on at least one of a lack of completeness of the first and/or second untrusted data or a lack of continuity of the first and/or second untrusted data; determining a second credibility based on expected behaviors indicative of at least one of the lack of completeness of the first and/or second untrusted data or the lack of continuity of the first and/or second untrusted data; determining a third credibility based on artifacts in estimated inputs indicative of systemic unknown factors; and aggregating the first credibility, the second credibility, and the third credibility.

13 . The system of claim 12 , wherein determining the first credibility comprises using measured signals as an input, determining the second credibility comprises using the first and/or second untrusted data and trusted data as inputs, and determining the third credibility comprises using the untrusted data, estimated untrusted data, and reconciled data as inputs.

Assignments (1)
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Feb 4, 2021
From: PATEK, STEPHEN D.
To: DEXCOM, INC.
Reel/Frame 055153/0764 →
Continuity (2)
Provisional Application 62935920 · Nov 15, 2019
Related Publication 20210151141A1 · May 20, 2021
References Cited (58)
US 8185412B1 · Harpale · 2012 [cited by examiner]
US 10341866B1 · Spencer et al. · 2019 [cited by applicant]
US 11024429B2 · Patek et al. · 2021 [cited by applicant]
US 20050192843A1 · Pratt et al. · 2005 [cited by applicant]
US 20070016449A1 · Cohen et al. · 2007 [cited by applicant]
US 20070033074A1 · Nitzan et al. · 2007 [cited by applicant]
US 20070083335A1 · Moerman · 2007 [cited by applicant]
US 20090006129A1 · Thukral · 2009 [cited by examiner]
US 20090069636A1 · Zivitz et al. · 2009 [cited by applicant]
US 20100249561A1 · Patek et al. · 2010 [cited by applicant]
US 20110098548A1 · Budiman · 2011 [cited by examiner]
US 20110257893A1 · Taylor et al. · 2011 [cited by applicant]
US 20130184547A1 · Taub et al. · 2013 [cited by applicant]
US 20130317316A1 · Kandeel · 2013 [cited by applicant]
US 20140031959A1 · Glode · 2014 [cited by examiner]
US 20140095201A1 · Farooq et al. · 2014 [cited by applicant]
US 20140214450A1 · Bechtold et al. · 2014 [cited by applicant]
US 20150190098A1 · Patek et al. · 2015 [cited by applicant]
US 20150269337A1 · Schulte et al. · 2015 [cited by applicant]
US 20160342754A1 · Vettoretti et al. · 2016 [cited by applicant]
US 20160354039A1 · Soto et al. · 2016 [cited by applicant]
US 20170091259A1 · Curran · 2017 [cited by examiner]
US 20170182248A1 · Rosinko · 2017 [cited by applicant]
US 20170262604A1 · Francois · 2017 [cited by applicant]
US 20170323064A1 · Bates · 2017 [cited by applicant]
US 20180174675A1 · Roy · 2018 [cited by examiner]
US 20180310822A1 · Indorf · 2018 [cited by examiner]
US 20180336286A1 · Shah · 2018 [cited by examiner]
US 20190099555A1 · Patek et al. · 2019 [cited by applicant]
US 20190173885A1 · Kamath et al. · 2019 [cited by applicant]
US 20190221307A1 · Mazlish et al. · 2019 [cited by applicant]
US 20190252079A1 · Constantin et al. · 2019 [cited by applicant]
US 20190374137A1 · Kovatchev et al. · 2019 [cited by applicant]
US 20200065681A1 · Wolf · 2020 [cited by examiner]
US 20210151189A1 · Patek · 2021 [cited by applicant]
US 20210151190A1 · Patek · 2021 [cited by applicant]
US 20210151196A1 · Patek · 2021 [cited by applicant]
CN 101801262A · 2010 [cited by applicant]
CN 102596307A · 2012 [cited by applicant]
CN 103907116A · 2014 [cited by applicant]
CN 104837517A · 2015 [cited by applicant]
CN 107851224A · 2018 [cited by applicant]
EP 4057892A1 · 2022 [cited by applicant]
JP 2014534483A · 2014 [cited by applicant]
WO 2016201120A1 · 2016 [cited by applicant]
WO 2021097092A1 · 2021 [cited by applicant]
Young et al., DAMON: A Data Authenticity Monitoring System for Diabetes Management, 2018, 2018 IEEE/ACM Third International Conference on Internet-of-Things Design and Implementation (IoTDI) (Year: 2018). [cited by examiner]
Li et al., Resolving Conflicts in Heterogeneous Data by Truth Discovery and Source Reliability Estimation, 2014, SIGMOD'14 (Year: 2014). [cited by examiner]
Dogan, ProTru: a provenance-based trust architecture for wireless sensor networks, Feb. 24, 2016, International Journal of Network Management (Year: 2016). [cited by examiner]
Bleiholder et al., Conflict Handling Strategies in an Integrated Information System, 2006, Institut fur Informatik, Humboldt-Universitat zu Berlin (Year: 2006). [cited by examiner]
Bleiholder et al., Data fusion, Jan. 2009, ACM Computing Surveys, vol. 41, Issue 1 (Year: 2009). [cited by examiner]
International Search Report and Written Opinion dated Mar. 12, 2021 for Application No. PCT/US2000/60241 in 13 pages. [cited by applicant]
Arnolds S., et al., “Common Standards of Basal Insulin Titration in T2DM,” Journal of Diabetes Science and Technology, May 2013, vol. 7, No. 3, pp. 771-788. [cited by applicant]
El Youssef J., et al., “A Review of Closed-Loop Algorithms for Glycemic Control in the Treatment of Type 1 Diabetes,” Algorithms, 2009, vol. 2, pp. 518-532. [cited by applicant]
Fritzen K., et al., “Modeling of Diabetes and Its Clinical Impact,” Journal of Diabetes Science and Technology, 2018, vol. 12, No. 5, pp. 976-984. [cited by applicant]
Triplitt C., “How to Initiate, Titrate, and Intensify Insulin Treatment in Type 2 Diabetes,” US Pharmacist, Oct. 18, 2007, vol. 32, No. 10, pp. 10-16. [cited by applicant]
Roppenheimer: “Understanding the AndroidAPS Screens,” Sep. 17, 2018, [Retrieved on May 9, 2023], XP093045426, 10 pages, Retrieved from the Internet: URL: https://github.com/roppenheimer/AndroidAPS-alt-Guide/blob/9685cf7… [cited by applicant]
Hajizadeh I., et al., “Adaptive Personalized Multivariable Artificial Pancreas Using Plasma Insulin Estimates,” Journal of Process Control, Oxford, GB, vol. 80, May 27, 2019, pp. 26-40. [cited by applicant]