Systems and methods for identifying errors and/or critical results in medical reports

Patent metadata

Field	Value
Publication number	US-11024406-B2
Application number	US-201313795741-A
Country	US
Kind code	B2
Filing date	Mar 12, 2013
Priority date	Mar 12, 2013
Publication date	Jun 1, 2021
Grant date	Jun 1, 2021

Abstract

Systems and methods for analyzing a medical report to determine whether the medical report includes at least one instance of at least one category selected from a group consisting of: gender error, laterality error, and critical finding. In some embodiments, one or more portions of text are identified from the medical report. Contextual information associated with the medical report is used to determine whether the identified one or more portions of text comprise at least one instance of at least one category selected from the group.

First claim

What is claimed is:

1. A system comprising:
at least one processor; and
at least one storage medium storing executable instructions that, when executed by the at least one processor, cause the at least one processor to carry out a method comprising:
analyzing a medical report to determine whether the medical report includes at least one critical finding, wherein a critical finding indicates that content of the medical report indicates a patient is experiencing a medical condition to which an urgent response is warranted, wherein analyzing the medical report comprises acts of:
applying at least one statistical model, trained to recognize in medical reports text indicative of a critical finding, to at least a first portion of text of the medical report and/or to contextual information associated with the medical report, wherein applying the at least one statistical model comprises:
tokenizing text of at least the first portion of text of the medical report;
determining, with the at least one statistical model, a concept related to one or more words and/or phrases of at least the first portion of text of the medical report;
identifying, with the at least one statistical model, a candidate critical finding from at least the first portion of text of the medical report;
determining, with the at least one statistical model, a confidence value representing a likelihood that the candidate critical finding is a critical finding; and
evaluating the confidence value determined for the candidate critical finding using the at least one statistical model to determine whether the candidate critical finding is a critical finding; and

in response to determining that the medical report includes the at least one critical finding, triggering a notification for at least one medical professional, wherein the notification indicates that the medical report indicates that the patient is experiencing a medical condition to which an urgent response is warranted.

2. The system of claim 1 , wherein:
the contextual information comprises metadata associated with the medical report; and
applying the at least one statistical model to at least the first portion and/or the contextual information comprises applying the at least one statistical model to at least the first portion and to the metadata associated with the medical report.

3. The system of claim 1 , wherein applying the at least one statistical model to at least the first portion and/or the contextual information comprises applying the at least one statistical model to the first portion and to one or more second portions of text from the medical report that are different from the first portion.

4. The system of claim 1 , wherein the at least one statistical model is at least one first statistical model and the likelihood is a first likelihood and wherein the method further comprises:
analyzing the medical report to determine whether the medical report includes at least one instance of at least one type of fact selected from a group consisting of: gender error and laterality error, comprising applying at least one second statistical model, trained to recognize in medical reports text indicative of a gender error and/or a laterality error to the at least the first portion and/or to contextual information to determine a second likelihood that the first portion includes at least one instance of at least one type of fact selected from the group;
evaluating the second likelihood determined for the first portion using the at least one second statistical model to determine whether the first portion includes at least one instance of at least one type of fact selected from the group; and
in response to determining that the first portion of the medical report includes at least one instance of a gender error and/or a laterality error, triggering presentation to a user of at least one indication that the first portion of the medical report has been determined to contain the at least one instance of gender error and/or laterality error.

5. The system of claim 4 , wherein the method further comprises:
presenting the at least one indication to the user, wherein presenting the at least one indication to the user comprises:
displaying to the user at least a part of the medical report comprising the first portion that was determined to contain the at least one instance of a gender error and/or a laterality error; and
differentiating the first portion of the text from at least one other portion of the text of the medical report to identify the first portion as including the at least one instance of a gender error and/or a laterality error.

6. The system of claim 5 , wherein displaying to the user the first portion of the medical report comprises displaying the first portion in response to activation by the user of at least one user interface feature that, when activated by the user, causes at least the first portion of the text of the medical report to be displayed to the user.

7. The system of claim 4 , wherein the at least one type of fact is laterality error.

8. The system of claim 4 , wherein the at least one type of fact is gender error.

9. The system of claim 4 , wherein the method further comprises:
training the at least one first statistical model to recognize text in medical reports indicative of a critical finding and training at least one second statistical model to recognize text in medical reports indicative of a gender error and/or a laterality error, wherein training the at least one first statistical model comprises training the at least one first statistical model with a corpus of medical reports that are each associated with information indicating whether the medical report includes a critical finding and wherein training the at least one second statistical model comprises training the at least one second statistical model with a corpus of medical reports that are each associated with information indicating whether the medical report includes a gender error and/or a laterality error.

10. The system of claim 4 , wherein evaluating the first confidence value determined for the candidate critical finding using the at least one first statistical model to determine whether the candidate critical finding is a critical finding and evaluating the second likelihood determined for the first portion using the at least one second statistical model to determine whether the first portion includes at least one instance of at least one type of fact selected from the group comprises determining whether at least one of the confidence value or the second likelihood is above a threshold level.

11. The system of claim 10 , wherein determining whether at least one of the first likelihood or the second likelihood is above a threshold level comprises one or more of:
determining whether the first portion includes a gender error by determining whether the second likelihood is above a first threshold;
determining whether the first portion includes a laterality error by determining whether the second likelihood is above a second threshold; or
determining whether the first includes a critical finding by determining whether the confidence value is above a third threshold,
wherein the first threshold, second threshold, and third thresholds have different threshold values.

12. The system of claim 1 , wherein evaluating the confidence value determined for the candidate critical finding using the at least one statistical model to determine whether the candidate critical finding is a critical finding comprises evaluating the confidence value using one or more thresholds.

13. The system of claim 12 , wherein evaluating the confidence value using

Assignees

• Nuance Communications Inc

Inventors

• Sadeghi Sepehr
• Mistry Atul T

Classifications

• G06Q10/10

  Office automation; Time management · CPC title

• G16H15/00 Primary

  ICT specially adapted for medical reports, e.g. generation or transmission thereof · CPC title

• G16H40/67

  for remote operation · CPC title

• G06F40/237

  Lexical tools · CPC title