Translation modeling methods and systems for simulating sensor measurements

Patent metadata

Field	Value
Publication number	US-11670425-B2
Application number	US-202016848687-A
Country	US
Kind code	B2
Filing date	Apr 14, 2020
Priority date	Dec 9, 2019
Publication date	Jun 6, 2023
Grant date	Jun 6, 2023

Abstract

Medical devices and related systems and methods are provided. A method of estimating a physiological condition involves determining a translation model based at least in part on relationships between first measurement data corresponding to instances of a first sensing arrangement and second measurement data corresponding to instances of a second sensing arrangement, obtaining third measurement data associated with the second sensing arrangement, determining simulated measurement data for the first sensing arrangement by applying the translation model to the third measurement data, and determining an estimation model for a physiological condition using the simulated measurement data, wherein the estimation model is applied to subsequent measurement output provided by an instance of the first sensing arrangement to obtain an estimated value for the physiological condition.

First claim

What is claimed is:

1. A method of estimating a physiological condition using a first sensing arrangement influenced by the physiological condition, the method comprising:
determining, by a computing device, a translation model based at least in part on relationships between first measurement data corresponding to instances of the first sensing arrangement and second measurement data corresponding to instances of a second sensing arrangement, wherein the second sensing arrangement is different from the first sensing arrangement;
obtaining, by the computing device, third measurement data associated with the second sensing arrangement;
determining, by the computing device, simulated measurement data for the first sensing arrangement by applying the translation model to the third measurement data;
determining, by the computing device, an estimation model for the physiological condition using the simulated measurement data, wherein the estimation model is different from the translation model; and
causing, by the computing device, calibration of an instance of the first sensing arrangement by providing the estimation model to the instance of the first sensing arrangement or to an electronic device used in connection with the instance of the first sensing arrangement, wherein the estimation model is applied to subsequent measurement output provided by the instance of the first sensing arrangement to obtain an estimated value for the physiological condition.

2. The method of claim 1 , further comprising obtaining reference measurement data corresponding to the first measurement data and the second measurement data, wherein determining the translation model comprises optimizing the translation model using the reference measurement data.

3. The method of claim 2 , wherein determining the estimation model comprises determining the estimation model based at least in part on relationships between the simulated measurement data and the reference measurement data.

4. The method of claim 3 , wherein determining the estimation model comprises training the estimation model using the simulated measurement data as an input to the estimation model and using at least a portion of the reference measurement data as an output of the estimation model.

5. The method of claim 1 , wherein determining the estimation model comprises determining the estimation model based at least in part on the simulated measurement data and the first measurement data.

6. The method of claim 5 , wherein determining the translation model comprises training the translation model using the first measurement data and the second measurement data as an input to the translation model.

7. The method of claim 5 , further comprising obtaining reference measurement data concurrent to the first measurement data and the second measurement data, wherein determining the estimation model comprises determining the estimation model based at least in part on relationships between the simulated measurement data, the first measurement data, and the reference measurement data.

8. The method of claim 7 , wherein determining the estimation model comprises training the estimation model using the simulated measurement data and the first measurement data as an input to the estimation model and at least a portion of the reference measurement data as an output of the estimation model.

9. The method of claim 1 , wherein determining the translation model comprises training the translation model using the first measurement data as an output of the translation model and the second measurement data as an input to the translation model.

10. The method of claim 1 , the first sensing arrangement comprising a target sensing device influenced by the physiological condition and the second sensing arrangement comprising a template sensing device influenced by the physiological condition, wherein the method further comprises:
obtaining the first measurement data from instances of the target sensing device; and
obtaining the second measurement data from instances of the template sensing device concurrent to the first measurement data, wherein respective sets of the first measurement data and the second measurement data are concurrently obtained from a respective common patient using respective instances of the target sensing device and the template sensing device.

11. The method of claim 10 , further comprising obtaining reference measurement data concurrent to the third measurement data, wherein:
obtaining the third measurement data comprises obtaining the third measurement data from second instances of the template sensing device and creating the simulated measurement data for the target sensing device by applying the translation model to the third measurement data; and
determining the estimation model comprises training the estimation model using the simulated measurement data as an input to the estimation model and using the reference measurement data as an output of the estimation model.

12. The method of claim 11 , wherein determining the translation model comprises training the translation model using the first measurement data and the second measurement data as an input to the translation model.

13. The method of claim 11 , wherein determining the translation model comprises determining the translation model based on relationships between the first measurement data and the second measurement data using a shifting modeling technique comprising a series of invertible, memory-less transforms to shift sensing signals between a first domain associated with the first sensing arrangement and a second domain associated with the second sensing arrangement.

14. The method of claim 10 , wherein the target sensing device comprises a new glucose sensor and the template sensing device comprises a legacy glucose sensor.

15. The method of claim 1 , further comprising providing the estimation model associated with the first sensing arrangement to the instance of the first sensing arrangement via a network.

16. The method of claim 1 , further comprising determining a delivery command for operating an actuation arrangement of an infusion device based at least in part on the estimated value for the physiological condition.

17. A system comprising:
a database to store first sensor measurement data corresponding to instances of a first sensing arrangement influenced by a physiological condition and second sensor measurement data corresponding to instances of a second sensing arrangement influenced by the physiological condition, wherein a type or configuration of the second sensing arrangement is different from the first sensing arrangement; and
a server coupled to the database and a network to:
determine a translation model associated with the first sensing arrangement based at least in part on relationships between the first sensor measurement data and a first subset of the second sensor measurement data,
determine simulated measurement data for the first sensing arrangement by applying the translation model to a second subset of the second sensor measurement data,
determine an estimation model for the physiological condition using the simulated measurement data, wherein the estimation model is different from the translation model, and
cause calibration of an instance of the first sensing arrangement by providing, via the network, the estimation model associated with the first sensing arrangement to the instance of the first sensing arrangement or to a computing device used in connection with the instance of the first sensing arrangement.

18. The metho

Assignees

• Medtronic Minimed Inc

Inventors

• Gee Elaine
• Ajemba Peter
• Engheta Bahman
• Nishida Jeffrey
• Varsavsky Andrea
• Nogueira Keith

Classifications

• A61B5/0022

  Monitoring a patient using a global network, e.g. telephone networks, internet · CPC title

• G16H70/60

  relating to pathologies · CPC title

• G06N3/088

  Non-supervised learning, e.g. competitive learning · CPC title

• G06F18/214

  Generating training patterns; Bootstrap methods, e.g. bagging or boosting · CPC title

• A61M2205/50

  with microprocessors or computers · CPC title