Healthcare information technology system for predicting or preventing readmissions

Patent metadata

Field	Value
Publication number	US-10943676-B2
Application number	US-201414225549-A
Country	US
Kind code	B2
Filing date	Mar 26, 2014
Priority date	Jun 8, 2010
Publication date	Mar 9, 2021
Grant date	Mar 9, 2021

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 for predicting or preventing hospital readmission, the method comprising:
receiving an indication of an event for a patient from a hospital at which the patient is being treated;
training at least one machine-trained classifier, the training comprising:
acquiring training data of various patients from a plurality of sources;
determining one or more variables from the training data, the one or more variables linked to a likelihood of readmission; and
generating the at least one machine-trained classifier;

automatically triggering application of a predictor of readmission specific to the hospital in response to the receiving of the indication, the predictor comprising:
the at least one machine-trained classifier;
a combination matrix that weights and combines the one or more variables; and
at least one feature vector;

populating the at least one feature vector by assigning values, using the combination matrix, to the one or more variables in a separate data storage device or location;
applying, by a processor, the predictor of readmission to an electronic medical record of the patient by inputting the populated at least one feature vector to the combination matrix;
predicting, automatically by the processor, a probability of readmission of the patient based on the applying of the predictor to the electronic medical record of the patient; and
outputting a readmission indicator for the patient as a function of the probability.

2. The method of claim 1 , wherein the values for the one or more variables are obtained by mining the electronic medical record, the mining comprising:
mining from a first data source of the electronic medical record and mining from a second data source of the electronic medical record, the first data source comprising structured data and the second data source comprising unstructured data; and
mining output values for the at least one feature vector in a structured format from the first data source and the second data source.

3. The method of claim 1 , wherein the values are obtained by mining, wherein each of the values is inferred by a probabilistic combination of probabilities associated with different possible values from different sources, and wherein the inferred values comprise the at least one feature vector.

4. The method of claim 3 , wherein the mining comprises mining as a function of existing knowledge, guidelines, best practices, or about specific institutions regarding readmissions.

5. The method of claim 1 , wherein outputting the readmission indicator comprises generating a cell phone alert, a bedside monitor alert, an alert associated with prevention of data entry, or combinations thereof.

6. The method of claim 1 further comprising:
automatically scheduling a job entry in a workflow of a case manager, the job entry being for examination to avoid readmission; and
outputting at least one of the one or more variables having one of the values for the patient associated with a strongest link to the probability indicating a risk of readmission, the strongest link being relative to links for other values of other variables to the risk.

7. The method of claim 1 , wherein the event is a discharge, a readmission, or an entry of new data, and wherein applying the at least one machine-trained classifier comprises a statistical model.

8. The method of claim 1 , wherein outputting the readmission indicator comprises generating a mitigation plan associated with the predicting, wherein the mitigation plan reduces the patient's likelihood of readmission.

9. The method of claim 1 , wherein outputting the readmission indicator comprises outputting based on a criteria set for the hospital, and wherein the plurality of sources is various hospitals with similar readmission concerns, sizes, or patient populations.

10. The method of claim 9 , wherein outputting the readmission indicator further comprises outputting instructions based on the probability, and wherein the criteria set is readmission within a time period.

11. The method of claim 10 further comprising:
applying a predictor of compliance by the patient with the instructions to the electronic medical record of the patient; and
predicting a probability of compliance of the patient based on the applying of the predictor of compliance.

12. A system for predicting or preventing hospital readmission, the system comprising:
at least one memory operable to store data for a plurality of readmitted patients of a first hospital; and
a processor configured to:
identify at least one variable contributing to patient readmission specific to the first hospital based on the data for the plurality of the readmitted patients of the first hospital, wherein the at least one variable is linked to a likelihood of readmission;
train at least one machine-trained classifier using (1) the at least one variable and (2) interrelationships of the at least one variable to an outcome of readmission;
generate, based on the at least one variable, a predictor of readmission for a second patient of the first hospital, wherein the predictor comprises:
the at least one machine-trained classifier; and
a combination matrix, wherein the combination matrix weights and combines one or more variables of at least one feature vector;

populate the at least one feature vector by assigning values, using the combination matrix, to the at least one variable;
in response to receiving an indication of an event for the second patient, apply the generated predictor to an electronic medical record of the second patient by inputting the populated at least one feature vector to the combination matrix; and
based on the applying, output a readmission probability indicator for the second patient.

13. The system of claim 12 , wherein the processor is configured to identify and generate by machine learning a statistical model from the data, the predictor comprising the combination matrix corresponding to the statistical model.

14. The system of claim 12 , wherein the processor is configured to mine the data for the plurality of the readmitted patients including mining unstructured information, the mining providing the assigned values by inference from different possible values in the data and probabilities assigned to the possible values.

15. The system of claim 12 , wherein the processor is configured to associate different workflows with different possible predictions of the predictor.

16. The system of claim 12 , wherein the processor is configured to identify another variable contributing to patient readmission specific to the first hospital based on data from a plurality of readmitted patients of a second hospital.

17. The system of claim 16 , wherein the processor is configured to automatically predict the probability of readmission of the patient based on the data from the plurality of readmitted patients of the second hospital.

18. A non-transitory computer readable storage medium storing computer-useable instructions, that when used by one or more computing devices, cause the one or more computing devices to perform operations comprising:
training at least one machine-trained classifier using training variables comprising (1) readmission data sets from a plurality of sources and (2) a set of interrelationships among the training variables and an outcome of readmission, wherein the training variables are based on patient healthcare data and are linked to a likelihood of readmission specific to a hospital;
predicting a probability of readmission of a patient, the

Assignees

• Cerner Innovation Inc

Inventors

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

Classifications

• G06Q10/10

  Office automation; Time management · CPC title

• G16H40/20

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

• G16H50/20

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