Method of continuous prediction of patient severity of illness, mortality, and length of stay

• US 9,569,723 B2
• Filed: 11/03/2011
• Issued: 02/14/2017
• Est. Priority Date: 11/08/2010
• Status: Active Grant

• Current Assignee: Koninklijke Philips Electronics N.V. (Koninklijke Philips N.V.)
• Original Assignee: Koninklijke Philips N.V.
• Inventors: Saeed, Mohammed
• Primary Examiner(s): Rifkin, Ben
• Assistant Examiner(s): BROWN JR, NATHAN H
• Application Number:
  US13/883,766
• Publication Number:
  US 20130232103A1
• Time in Patent Office: 1,930 Days
• Field of Search: None
• US Class Current: 1/1
• CPC Class Codes: G01N 2800/52 Predicting or monitoring the response to treatment, e.g. for selection of therapy based on assay results in personalised medicine; Prognosis
• CPC Class Codes: G06F 18/00 Pattern recognition
• CPC Class Codes: G06N 5/01 Dynamic search techniques; Heuristics; Dynamic trees; Branch-and-bound
• CPC Class Codes: G16H 10/20 for electronic clinical trials or questionnaires
• CPC Class Codes: G16H 10/60 for patient-specific data, e.g. for electronic patient records
• CPC Class Codes: G16H 50/20 for computer-aided diagnosis, e.g. based on medical expert systems
• CPC Class Codes: G16H 50/30 for calculating health indices; for individual health risk assessment
• CPC Class Codes: G16H 50/50 for simulation or modelling of medical disorders
• CPC Class Codes: G16H 50/70 for mining of medical data, e.g. analysing previous cases of other patients
• CPC Class Codes: G16Z 99/00 Subject matter not provided for in other main groups of this subclass

Abstract

A method for predicting a patient'"'"'s outcome variable, such as a probability of mortality/recovery, includes accessing at least one of a plurality past patients'"'"' data fields including physiological and/or laboratory data, and a time of stay indicating how long each patient had been under care at the time. An outcome variable estimation algorithm is generated by data mining from a plurality of past patients'"'"' physiological and/or laboratory data, corresponding time of stay, and associated outcome variables. A current patient'"'"'s outcome variable is determined from the current patient'"'"'s physiological and/or laboratory data, the current time of stay, and the outcome variable estimation algorithm.

Claims

1. A patient monitoring system comprising:

a patient monitor device which collects physiological data from a patient;

a control processor configured to;

receive the collected physiological data from the patient monitor device,

receive laboratory data about the patient from a laboratory,

determine a time of stay indicative of a time the patient has been in a healthcare facility,

predict an outcome variable continuously in real time by applying an outcome variable estimation algorithm to the patient'"'"'s physiological and laboratory data using the determined time of stay.

2. The system according to claim 1, wherein the patient is in an intensive care unit (ICU) of the healthcare facility and the time of stay is indicative of a time the patient has been in the ICU and wherein the outcome variables include a predicted length of stay in the ICU and a probability of mortality.

3. The system according to claim 1, wherein the outcome variable estimation algorithm is generated by:

accessing at least one of a plurality past patients'"'"' data fields including physiological and laboratory data, and a time of stay indicating how long each patient had been in the healthcare facility at the time the physiological and laboratory data were generated and associated outcome variables from a clinical database;

generating the outcome variable estimation algorithm from the plurality past patients physiological and laboratory data, corresponding time of stay, and associated outcome variables.

4. The system according to claim 1, wherein the outcome variable estimation algorithm is generated using at least one of:

linear regression analysis, nonlinear regression analysis, neural networks, support vector machines, radial basis functions, expert rule-based, and classifier decision trees.

5. The system according to claim 1, wherein the physiological data and the laboratory data include values indicative of the patient'"'"'s health, wherein predicting the outcome variable by applying the outcome variable algorithm includes:

generating patient health scores for the values;

weighting each patient health score based on the determined time of stay; and

combining the weighted scores, the combined weighted scores being indicative of the outcome variable.

6. The system according to claim 1, wherein the outcome variable algorithm is configured to calculate a Simplified Acute Physiology (SAP) score weighted based on the time of stay.

7. A patient monitoring system comprising:

a patient monitor device which collects physiological data from a patient;

a laboratory server configured to transmit laboratory data of the patient;

a control processor configured to;

receive the collected physiological data from the patient monitor device,

receive the laboratory data of the patient from the laboratory,

determine a time of stay indicative of a time the patient has been in a healthcare facility,

predict an outcome variable by applying an outcome variable estimation algorithm to the patient'"'"'s physiological and laboratory data using the determined time of stay.

8. A patient monitoring system comprising:

a patient monitor device configured to collect physiological data from a patient;

a control processor configured to;

receive the collected physiological data from the patient monitor device,

receive the laboratory data of the patient from the laboratory,

determine a time of stay indicative of a time the patient has been in a healthcare facility,

predict an outcome variable by applying an outcome variable estimation algorithm to the patient'"'"'s physiological and laboratory data using the determined time of stay; and

a display configured to display the predicted outcome variable.

9. A method for predicting a patient'"'"'s outcome variable, the method comprising, with one or more processors:

accessing at least one of a plurality past patients'"'"' data fields including physiological and laboratory data, and a time of stay indicating how long each patient had been in a healthcare facility at the time the physiological and laboratory data were generated and associated outcome variables from a clinical database;

generating a outcome variable estimation algorithm from the plurality past patients physiological and laboratory data, corresponding time of stay, and their associated outcome variables;

receiving a current patient'"'"'s physiological and laboratory data;

determining the current patient'"'"'s current time of stay in a healthcare facility; and

applying the outcome variable estimation algorithm to the current patient'"'"'s physiological and laboratory data and the current time of stay to estimate the current patient'"'"'s probability of outcome continuously in real time; and

controlling a display monitor to display a current probability of outcome.

10. The method according to claim 9, wherein the outcome variables include at least a probability of mortality.

11. The method according to claim 9, wherein in the past patient data fields include one or more vital signs, automated or user-input problem lists including chronic problems, acute admitting problems, clinical laboratory data, medications, subjective nursing assessment, patient demographics, and current length of stay in a hospital under intensive care.

12. The method according to claim 9, wherein the past patient associated outcome variables include time of stay, recovery/mortality, and intervention.

13. The method according to claim 9, wherein the outcome variable estimation algorithm is identified using at least one of:

linear regression analysis, nonlinear regression analysis, neural networks, support vector machines, radial basis functions, expert rule-based, and classifier decision trees.

14. A non-transitory computer readable medium containing software which when loaded into one or more processors programs the one or more processors to perform the method according to claim 9.

15. A method of predicting a probability of outcome, comprising:

measuring physiological and laboratory data for a current patient who is staying under intensive care in a medical facility;

determining a time of stay which the current patient has been staying under intensive care in the medical facility;

applying a time of stay dependent outcome prediction algorithm using the measured physiological and laboratory data and the determined time of stay to predict a probability of recovery/mortality; and

controlling a display device to display the probability of recovery/mortality.

16. A non-transitory computer readable medium containing software which when loaded into a processor programs the processor to perform the method according to claim 15.

17. A processor programmed to perform the method according to claim 15.

18. The method according to claim 15, wherein predicting the probability of recovery/mortality includes:

assigning scores to the measured physiological and laboratory data;

weighting the scores based on a current time of stay under intensive care; and

combining the weighted scores to generate an acuity score indicative of the probability of recovery/mortality.