Healthcare information technology system for predicting or preventing readmissions

Patent metadata

Field	Value
Publication number	US-11664097-B2
Application number	US-202016951515-A
Country	US
Kind code	B2
Filing date	Nov 18, 2020
Priority date	Jun 8, 2010
Publication date	May 30, 2023
Grant date	May 30, 2023

Abstract

Hospital readmissions may be prevented. Readmission is prevented by predicting the probability of a given patient to be readmitted. The probability alone may prevent readmission by educating the patient or medical professional. The probability may be predicted during a patient stay and used to generate a workflow action item to reduce the probability, to warn, to output appropriate instructions, and/or assist in avoiding readmission. The probability may be specific to a hospital, physician group, or other entity, allowing prevention to focus on past readmission causes for the given entity.

First claim

What is claimed is:

1. A method comprising:
receiving an indication of an event for a patient associated with a particular hospital;
determining one or more hospital-specific variables that meet a correlation with readmission at the particular hospital based on training data comprising hospital data identifying readmission for patients at the particular hospital;
generating, via one or more processors, a predictive model of hospital-specific readmission for the particular hospital that includes a plurality of machine-trained classifiers trained with the one or more hospital-specific variables, the plurality of machine-trained classifiers for predicting probability of readmission based on the one or more hospital-specific variables;
automatically triggering, via the one or more processors, application of the predictive model of hospital-specific readmission to an electronic health record that is specific to the patient in response to the indication of the event;
in response to application of the predictive model of hospital-specific readmission, automatically predicting, via the one or more processors, a probability of readmission of the patient for the particular hospital based on the plurality of machine-trained classifiers and based on one or more values in the electronic health record;
determining, via the one or more processors, that the probability of readmission meets a risk threshold; and
presenting a notification based on the probability of readmission meeting the risk threshold.

2. The method of claim 1 , wherein the event of the patient is an admission to the particular hospital, wherein the one or more hospital-specific variables are institution specific to the particular hospital, and wherein the training data comprises readmission information for other patients who were readmitted to the hospital, the readmission information comprising an amount of time prior to readmission.

3. The method of claim 1 , wherein the event is a discharge from the particular hospital, an admission to the particular hospital, or a scheduled discharge from the particular hospital, and wherein the training data comprises readmission information for other patients who were readmitted to the particular hospital and from at least one other hospital.

4. The method of claim 1 , wherein at least one hospital-specific variable of the particular hospital differs from a corresponding hospital-specific variable of a second hospital, and further comprising:
outputting a readmission indicator for the patient as a function of the probability of readmission, wherein outputting the readmission indicator comprises generating one or more of a cell phone alert, a bedside monitor alert, and an alert associated with prevention of data entry.

5. The method of claim 1 , wherein the predictive model of hospital-specific readmission differs from a predictive model of hospital-specific readmission of a second hospital.

6. The method of claim 1 , further comprising:
providing a list of one or more other variables that are linked to a reduced probability of readmission of the patient for the event; and
based at least in part on the list, providing a recommended clinical action corresponding to a mitigation plan for reducing the probability of readmission of the patient for the event.

7. The method of claim 1 , further comprising:
assigning, using at least one of the plurality of machine-trained classifiers, weights to the one or more hospital-specific variables indicating a strength of a correlation of the one or more hospital-specific variables to the increased probability of readmission, wherein the one or more hospital-specific variables are trained based on a plurality of hospital-specific patients, and wherein automatically predicting the probability of readmission is based at least in part on the weights assigned to the one or more hospital-specific variables.

8. The method of claim 1 , further comprising:
mining information from structured and unstructured data in an electronic medical record, the mining performed as a function of domain knowledge comprising factors that influence a risk of a disease and progression information of the disease; and
incorporating the information into the predictive model, wherein automatically predicting the probability of readmission is based at least in part on incorporating the information.

9. The method of claim 8 , wherein the mining is configured for different formats of data sources within electronic medical records of the particular hospital and at least one other hospital, and wherein the information comprises a value for the one or more hospital-specific variables and an inferred value for the one or more hospital-specific variables.

10. The method of claim 9 , wherein the domain knowledge provides an indication of reliability of the value and the inferred value based on a corresponding data source or a context.

11. A system comprising:
at least one memory; and
one or more processors configured to:
receive an indication of an event for a patient associated with a particular hospital;
determine one or more hospital-specific variables that meet a correlation with readmission at the particular hospital based on from training data comprising hospital data identifying readmission for patients at the particular hospital;
machine-train a plurality of classifiers using the one or more hospital-specific variables;
generate a predictive model of hospital-specific readmission for the particular hospital that includes the plurality of classifiers, the plurality of machine-trained classifiers for predicting probability of readmission based on the one or more hospital-specific variables;
automatically trigger application of the predictive model of hospital-specific readmission to an electronic health record that is specific to the patient in response to the indication;
in response to application of the predictive model of hospital-specific readmission, automatically predict a probability of readmission of the patient for the particular hospital based on the plurality of machine-trained classifiers and based on one or more values in the electronic health record;
determine that the probability of readmission meets a risk threshold; and
present a notification based on the probability of readmission meeting the risk threshold.

12. The system of claim 11 , wherein the training data comprises readmission data for patients readmitted to the particular hospital within a particular time period, and wherein at least one hospital-specific variable of the particular hospital differs from a corresponding hospital-specific variable of a second hospital.

13. The system of claim 11 , wherein the one or more hospital-specific variables used to train the plurality of classifiers are from data of patients readmitted to the particular hospital for admission problems, and wherein the one or more processers are configured to:
machine-train each of the plurality of classifiers using the one or more hospital-specific variables from the training data of patients who were readmitted to the hospital for a corresponding admission problem following a discharge;
wherein the probability of readmission of the patient is automatically predicted for the corresponding admission problem.

14. The system of claim 11 , wherein the one or more processors are configured to: receive a selection to alter a threshold associated with at least one of the plurality of classifiers in the predictive model, and wherein the probability of readmission of the patient is automatically predicted based at least in part on the altered threshold.

15. Th

Assignees

• Cerner Innovation Inc

Inventors

• Farooq Faisal
• Krishnapuram Balaji
• Rao Bharat R
• Rosales Romer E
• Yu Shipeng

Classifications

• G16H50/20

  for computer-aided diagnosis, e.g. based on medical expert systems · CPC title

• G16H50/70

  for mining of medical data, e.g. analysing previous cases of other patients · CPC title

• G16H50/30

  for calculating health indices; for individual health risk assessment · CPC title

• G16H10/60 Primary

  for patient-specific data, e.g. for electronic patient records · CPC title

• G16H40/20

  for the management or administration of healthcare resources or facilities, e.g. managing hospital staff or surgery rooms · CPC title