E-triage: an electronic emergency triage system

Patent metadata

Field	Value
Publication number	US-12148529-B2
Application number	US-201916544016-A
Country	US
Kind code	B2
Filing date	Aug 19, 2019
Priority date	Feb 18, 2016
Publication date	Nov 19, 2024
Grant date	Nov 19, 2024

Abstract

An embodiment in accordance with the present invention provides a system that uses simple standardized patient information routinely collected at triage to distribute patients amongst triage levels based on critical and time-sensitive outcomes. The present invention estimates the probability of electronic medical record (EMR) recorded events for patients at triage. Predictions are made for patients based upon clinical information routinely collected at triage which include demographics (age and gender), vital signs (temperature, heart rate, systolic blood pressure, respiratory rate, and oxygen saturation), complaint(s), medical/surgical history, chronic conditions, and mode of arrival. Vital signs are categorized as normal or gradations of abnormal.

First claim

What is claimed is:

1. A method for performing a triage, the method comprising:
receiving, by a computing device having a trained machine learning prediction engine, data pertaining to an individual visiting an emergency department, the data pertaining to the individual including health-related information of the individual, wherein:
the trained machine learning prediction engine having been trained using supervised machine learning and training data related to a plurality of emergency department visits by individuals of an emergency department population, and
the trained machine learning prediction engine having used ensemble learning to create separate decision forest models, each of which being for prediction of a respective outcome of a plurality of predefined future acute clinical outcomes for the individuals of the emergency department population;

applying, in tandem by the computing device, the data pertaining to the individual to the separate decision forest models for prediction of the respective outcomes of the plurality of future acute clinical outcomes for the individual of the emergency department population to produce predicted probabilities of a risk for each of the respective outcomes for the individual of the emergency department population;
mapping, by the computing device, one of the produced predicted probabilities of the risk for the respective outcomes to one of a plurality of triage levels, the one of the plurality of triage levels being required to direct care for the individual in accordance with the one of the plurality of triage levels, the one of the produced predicted probabilities being mapped is based on values of the produced predicted probabilities;
making available, by the computing device, information regarding the one of the plurality of triage levels;
providing, by the computing device, an override feature to allow a health care provider to override the one of the plurality of triage levels;
collecting, by the computing device, a frequency of an override, a type of the override, and optionally reasons for the override; and
analyzing, by the computing device, the collected frequency of the override, the type of the override, and optionally the reasons for the override to iteratively improve an accuracy of the mapping of the one of the produced predicted probabilities of the risk for the respective outcomes to the one of the plurality of triage levels.

2. The method of claim 1 , wherein the health-related information of the individual comprises information of a group of one or more of demographic information, vital sign information, body temperature, heart rate, systolic blood pressure, respiratory rate, oxygen saturation, complaint information, medical history, information related to any chronic conditions, and mode of arrival.

3. The method of claim 1 , further comprising:
assigning, by the computing device, a triage level to the individual based on the predicted probabilities of the risk for the each future acute clinical outcome.

4. The method of claim 3 , further comprising:
displaying, by the computing device, a visual representation of the risk for future acute clinical outcomes associated with one or more of the individuals visiting the emergency department.

5. The method of claim 1 , wherein the plurality of predefined future acute clinical outcomes for the emergency department population include in-hospital mortality, intensive care unit admission, an emergent surgical procedure within a given time period of emergency department disposition, and inpatient hospitalization.

6. The method of claim 1 , further comprising:
prompting a health care provider for information.

7. The method of claim 1 , wherein the computing device communicates bidirectionally with an electronic medical records system.

8. At least one non-transitory computer readable medium having instructions stored thereon, which when executed by a computing device configure the computing device to:
receive data pertaining to an individual visiting an emergency department, the data pertaining to the individual including health-related information of the individual, wherein:
a trained machine learning prediction engine of the computing device having been trained using supervised machine learning and training data related to a plurality of emergency department visits by individuals of an emergency department population, and
the trained machine learning prediction engine having used ensemble learning to create separate decision forest models, each of which being for prediction of a respective outcome of a plurality of predefined future acute clinical outcomes for the individuals of the emergency department population;

apply, in tandem, the data pertaining to the individual to the separate decision forest models for prediction of the respective outcomes of the plurality of future acute clinical outcomes for the individual of the emergency department population to produce predicted probabilities of a risk for each of the respective outcomes for the individual of the emergency department population;
map one of the produced predicted probabilities of the risk for the respective outcomes to one of a plurality of triage levels, the one of the plurality of triage levels being required to direct care for the individual in accordance with the one of the plurality of triage levels, the one of the produced predicted probabilities being mapped is based on values of the produced predicted probabilities;
make available information regarding the one of the plurality of triage levels;
provide an override feature to allow a health care provider to override the one of the plurality of triage levels;
collect a frequency of an override, a type of the override, and optionally reasons for the override; and
analyze the collected frequency of the override, the type of the override, and optionally the reasons for the override to iteratively improve an accuracy of the mapping of the one of the produced predicted probabilities of the risk for the respective outcomes to the one of the plurality of triage levels.

9. The at least one non-transitory computer readable medium of claim 8 , wherein the health-related information of the individual comprises information of a group of one or more of demographic information, vital sign information, body temperature, heart rate, systolic blood pressure, respiratory rate, oxygen saturation, complaint information, medical history, information related to any chronic conditions, and mode of arrival.

10. The at least one non-transitory computer readable medium of claim 8 , wherein the instructions further configure the computing device to:
assign a triage level to the individual based on the predicted probabilities of the risk for the each future acute clinical outcome.

11. The at least one computer readable medium of claim 10 , wherein the instructions further configure the computing device to:
display a visual representation of the risk for future acute clinical outcomes associated with one or more individuals visiting the emergency department.

12. The at least one computer readable medium of claim 8 , wherein the plurality of predefined future acute clinical outcomes for the emergency department population include in-hospital mortality, intensive care unit admission, an emergent surgical procedure within a given time period of emergency department disposition, and inpatient hospitalization.

13. The at least one computer readable medium of claim 8 , wherein the instructions further configure the computing device to provide a graphical user interface that allows a health care provider to override the mapped one of the plurality

Assignees

• Univ Johns Hopkins

Inventors

• Levin Scott R
• Kirsch Thomas
• Toerper Matthew
• Kelen Gabor D
• Dugas Andrea

Classifications

• G06N5/01

  Dynamic search techniques; Heuristics; Dynamic trees; Branch-and-bound · CPC title

• G06N20/20

  Ensemble learning · CPC title

• G16H50/30

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

• G16H10/60

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

• G06N20/00

  Machine learning · CPC title