Systems and methods for enhancing natural language processing

Patent metadata

Field	Value
Publication number	US-12164874-B2
Application number	US-202318490662-A
Country	US
Kind code	B2
Filing date	Oct 19, 2023
Priority date	Feb 23, 2018
Publication date	Dec 10, 2024
Grant date	Dec 10, 2024

Abstract

Methods and systems for enhanced natural language processing of clinical documentation are provided. Using natural language processing, a clinical condition is extracted from unstructured data within a current electronic document. A clinical ontology identifying itemsets associated with the clinical condition is retrieved, and indicators of relevant clinical concepts, as identified from the ontology, are searched from within the patient's longitudinal record, which comprises documentation from at least a prior encounter. Based on the whether the clinical concepts are present in the patent's record, a confidence is assigned to the NLP-extracted clinical condition, and one or more actions may be performed.

First claim

What is claimed is:

1. A computer-program product tangibly embodied in a non-transitory machine-readable storage medium, including instructions configured to cause one or more data processors to perform a set of operations comprising:
receiving a current electronic document with unstructured health-related data associated with an individual;
receiving an identifier of the individual associated with the unstructured health-related data;
parsing and extracting at least one clinical condition within the unstructured health-related data using one or more natural language processing techniques;
identifying one or more clinical concepts related to the clinical condition using one or more clinical ontologies for the at least one clinical condition, each clinical ontology providing contextual relationships between the clinical condition and the one or more clinical concepts;
retrieving one or more portions of a longitudinal electronic health record (EHR) associated with the individual, the longitudinal EHR comprising documentation of at least one prior encounter and having some structured data;
determining a confidence value for the clinical condition based on a statistical likelihood of whether at least one parameter of the clinical condition matches with the current electronic document and the longitudinal EHR;
based on whether the clinical condition has a sufficient confidence value, providing a notification, in real time, to a user of an application for entering the current electronic document, the notification either stating the clinical condition has a sufficient confidence value, the clinical condition has an insufficient confidence value, or requesting supplemental information to support a diagnosis for the clinical condition;
triggering a coding document quality process to check whether current documentation supports a given coding level assigned to the clinical condition, wherein determining whether the current documentation supports a given coding level comprises determining whether at least a threshold confidence metric has been met with regard to a presence of a clinical code corresponding to the clinical condition; and
providing the notification when the current documentation supports a coding level.

2. The computer-program product of claim 1 , wherein the confidence value comprises one of a positive, neutral, or negative state indicating a statistical likelihood the clinical condition matches the clinical concept is above, within, or below a threshold value.

3. The computer-program product of claim 2 , wherein the set of operations further comprises determining the current documentation is insufficient to provide the notification when the confidence value is negative or neutral.

4. The computer-program product of claim 2 , wherein the set of operations further comprises determining the current documentation is sufficient to provide the notification when the confidence value is positive.

5. The computer-program product of claim 1 , wherein the set of operations further comprises electronically modifying a relational database by adding an entry for the clinical condition, the entry indicating a likelihood that a diagnosis of the clinical condition is accurate.

6. The computer-program product of claim 5 , wherein determining the current documentation is sufficient to support a diagnosis includes measuring a presence or level of the one or more clinical concepts in the current documentation.

7. The computer-program product of claim 6 , wherein the likelihood that the diagnosis of the clinical condition is based on the confidence value and the presence or level of the related clinical concepts in the current documentation.

8. The computer-program product of claim 1 , wherein based on whether the clinical condition is verified, additionally performing triggering a coding document quality process to check whether the current documentation supports a given coding level assigned to the clinical condition, wherein determining whether the documentation supports a given coding level comprises determining a high confidence value to the presence of a clinical code corresponding to the clinical condition.

9. The computer-program product of claim 8 , wherein the set of operations further comprises providing a notification when the current documentation supports a coding level.

10. The computer-program product of claim 1 , wherein the parsing and extracting at least one clinical condition further comprises:
generating a tag for a clinical condition where it is determined there is an ambiguity in the unstructured health-related data; and
providing a notification including an indication of the tag, in real time, to the user of the application for entering the current electronic document.

11. A computer-implemented method comprising:
receiving a current electronic document with unstructured health-related data associated with an individual;
receiving an identifier of the individual associated with the unstructured health-related data;
parsing and extracting at least one clinical condition within the unstructured health-related data using one or more natural language processing techniques;
identifying one or more clinical concepts related to the clinical condition using one or more clinical ontologies for the at least one clinical condition, each clinical ontology providing contextual relationships between the clinical condition and the one or more clinical concepts;
retrieving one or more portions of a longitudinal electronic health record (EHR) associated with the individual, the longitudinal EHR comprising documentation of at least one prior encounter and having some structured data;
determining a confidence value for the clinical condition based on a statistical likelihood of whether at least one parameter of the clinical condition matches with the current electronic document and the longitudinal EHR;
based on whether the clinical condition has a sufficient confidence value, providing a notification, in real time, to a user of an application for entering the current electronic document, the notification either stating the clinical condition has a sufficient confidence value, the clinical condition has an insufficient confidence value, or requesting supplemental information to support a diagnosis for the clinical condition;
triggering a coding document quality process to check whether current documentation supports a given coding level assigned to the clinical condition, wherein determining whether the current documentation supports a given coding level comprises determining whether at least a threshold confidence metric has been met with regard to a presence of a clinical code corresponding to the clinical condition; and
providing the notification when the current documentation supports a coding level.

12. The computer-implemented method of claim 11 , wherein the confidence value comprises one of a positive, neutral, or negative state indicating a statistical likelihood the clinical condition matches the clinical concept is above, within, or below a threshold value.

13. The computer-implemented method of claim 12 , wherein the set of operations further comprises determining the current documentation is insufficient to provide the notification when the confidence value is negative or neutral.

14. The computer-implemented method of claim 12 , wherein the set of operations further comprises determining the current documentation is sufficient to provide the notification when the confidence value is positive.

15. The computer-implemented method of claim 11 , wherein the set of operations further comprises electronically

Assignees

• Cerner Innovation Inc

Inventors

• Kailasam Kanakasabha
• Mcnair Douglas S
• Murrish John Christopher
• Cohen David
• Smith Josh
• Vickers Alex
• Gupta Tanuj

Classifications

• G06F40/205

  Parsing · CPC title

• G06F40/237

  Lexical tools · CPC title

• G06F40/30 Primary

  Semantic analysis · CPC title