Therapeutic decisions systems and method using stochastic techniques

• US 20030060690A1
• Filed: 06/19/2002
• Published: 03/27/2003
• Est. Priority Date: 06/19/2001
• Status: Active Grant

• Current Assignee: University of Southern California
• Original Assignee: University of Southern California
• Inventors: Jelliffe, Roger, Botnen, Andreas, Shoemaker, William, Bayard, David
• Granted Patent: US 6,896,660 B2
• Time in Patent Office: Days
• US Class Current: 600/300
• CPC Class Codes: A61B 5/0205 Simultaneously evaluating both cardiovascular conditions and different types of body conditions, e.g. heart and respiratory condition
• CPC Class Codes: A61B 5/021 Measuring pressure in heart or blood vessels A61B5/0205 takes precedence
• CPC Class Codes: A61B 5/0537 Measuring body composition by impedance, e.g. tissue hydration or fat content
• CPC Class Codes: A61B 5/1455 using optical sensors, e.g. spectral photometrical oximeters
• CPC Class Codes: A61B 5/1468 using chemical or electrochemical methods, e.g. by polarographic means A61B5/1486 takes precedence
• CPC Class Codes: A61B 5/318 Heart-related electrical modalities, e.g. electrocardiography [ECG]
• CPC Class Codes: A61B 5/412 Detecting or monitoring sepsis
• CPC Class Codes: A61B 5/7275 Determining trends in physiological measurement data; Predicting development of a medical condition based on physiological measurements, e.g. determining a risk factor
• CPC Class Codes: G16H 10/60 for patient-specific data, e.g. for electronic patient records
• CPC Class Codes: G16H 20/30 relating to physical therapies or activities, e.g. physiotherapy, acupressure or exercising
• CPC Class Codes: G16H 40/67 for remote operation
• CPC Class Codes: G16H 50/70 for mining of medical data, e.g. analysing previous cases of other patients
• CPC Class Codes: Y10S 128/923 by comparison of patient data to other data
• CPC Class Codes: Y10S 128/924 using artificial intelligence

Abstract

Patients are evaluated in terms of various physiological parameters such as hemodynamic parameters. Covariate information indicative of the patient'"'"'s individual characteristics may also be determined. A state variable indicative of the patient'"'"'s characteristics is determined. This state variable is used to compare the patient to other patients in the database. The other patients in the database are evaluated in terms of similar parameters and their probability of survival as well as different treatments and the probability of survival based on those different treatments. The patient may be simply evaluated based on the database or a proper course of treatments may be determined based on the database.

Claims

1. A system comprising:

a monitoring device monitoring a plurality of physiological signals of a current patient;

a database, which stores records indicative of a plurality of patients including at least some of said physiological signals for said plurality of patients and prognosis information for said plurality of patients;

a processor, comparing said physiological signals to said stored records in a database, to determine mathematically patients in the database which are most similar to a current patient; and

to make available information about treatment and results for said patients who are most similar to the current patient.

2. A system as in claim 1, wherein said physiological signals include hemodynamic information.

3. A system as in claim 2, wherein said hemodynamic information includes bio impedance information.

4. A system as in claim 2, wherein said hemodynamic information includes temporal patterns within the hemodynamic information.

5. A system as in claim 2, further comprising obtaining covariate information indicative of a patient'"'"'s individual characteristics.

6. A system as in claim 5, wherein said processor mathematically determines said similar patients by determining a state variable for a patient based on at least said hemodynamic information and said covariate information.

7. A system as in claim 6, wherein said state variable is defined as a vector over time, where each point in time includes measurements at that point in time.

8. A system as in claim 7, wherein said state variable also includes derivatives of the hemodynamic measurements and integrals of the hemodynamic measurements.

9. A system as in claim 2, wherein said hemodynamic measurements include at least cardiac index, blood pressure, pulse oximetry information, and transcutaneous gas tension.

10. A system as in claim 2, wherein said processor also carries out interpolation to place a plurality of said values on a common time grid.

11. A system as in claim 7, wherein said state variable is also a function of non covariate parameters, and inconsistencies between said state variable are explained in terms of said covariate parameters.

12. A system as in claim 7, wherein said processor models said patient as a control system, with said state parameter representing a patient'"'"'s state, a control input representing a therapy, and the patient'"'"'s state following a nonlinear dynamic system with process noise being explained in terms of covariate parameters.

13. A system as in claim 12, wherein said processor models said patient with a survival probability based on nearest neighbor states to other patients in the database.

14. A system as in claim 7 wherein said processor uses said state variable to find nearest neighbors in the database which represent patients that are most similar to the current patient.

15. A system as in claim 14, wherein said finding nearest neighbors comprises determining a quadratic distance between the current patient and other patients in the database.

16. A system as in claim 14, wherein said processor uses a weighting matrix to determine said closest neighbors.

17. A system as in claim 16, wherein said weighting matrix is based on relationships between hemodynamic patterns and survival versus non survival.

18. A method, comprising:

monitoring a plurality of physiological signals of a current patient;

comparing said physiological signals of the current patient to stored physiological signals for previous patients stored in a database;

determining patients in said database who are most similar to the current patient; and

determining information about treatments and results for said patients who are most similar to the current patient.

19. A method as in claim 18, wherein said determining information comprises determining a likelihood that the patient will survive.

20. A method as in claim 18, wherein said determining information comprises determining a therapy for the current patient which has proved most statistically successful for said patients who are most similar to the current patient.

21. A system as in claim 18, wherein said physiological signals include signals indicative of hemodynamic information.

22. A system as in claim 21, wherein said signals indicative of hemodynamic information include at least cardiac index, blood pressure, pulse oximetry information and transcutaneous gas tension.

23. A system as in claim 18, further comprising obtaining covariate information indicative of a patients individual characteristics, and said determining most similar patients and said determining information is also based on said covariate information.

24. A system as in claim 21, wherein said hemodynamic information includes bio impedance information.

25. A system as in claim 23, wherein said comparing and determining most similar patients includes comparing temporal patterns within the physiological information.

26. A system as in claim 23, wherein said determining most similar patients comprises mathematically determining a state variable indicative of the current patient based on at least said hemodynamic information and said covariate information.

27. A method as in claim 23, wherein said determining a state variable comprises investigating derivatives of the hemodynamic information and integrals of the hemodynamic information.

28. A method as in claim 18, wherein said comparing comprises interpolating among said physiological signals to evaluate said physiological signals on a common time scale.

29. A method as in claim 18, wherein said determining information comprises modeling said patient as a control system with a state parameter representing a patient'"'"'s state, a control input representing a therapy, and a nonlinear dynamic system, with process noise being explained in terms of covariate information indicative of a patients individual characteristics.

30. A method as in claim 18, wherein said determining patients comprises determining quadratic distances between a state variable representing a current patient and state variables representing other patients in said database.

31. A method as in claim 30, further comprising using a weighting matrix to determine said closest patients.

32. A method as in claim 31, further comprising determining said weighting matrix based on hemodynamic parameters and survival versus non survival.

33. A method, comprising:

determining hemodynamic parameters of an individual patient;

determining covariate parameters of the individual patient representing the patients individual characteristics;

using said covariate characteristics and said hemodynamic parameters to develop a state variable that represents information that is continuously variable in time;

comparing said state variable with information indicative of a plurality of previously treated patients in the database; and

obtaining survival information about said patients in said database.