Granted Patent B2

US 12,731,694 · App. 18/941,342 · Granted Sep 8, 2026

Inventors: Jeremy John Kasmann (Evergreen, CO); Christopher Stanley Funk (Highlands Ranch, CO)

Assignee: Health Language, Inc.

G16H50/30

Abstract

Claims (42)

1 . A system, comprising:

a non-transitory memory;

a processor communicatively coupled to the non-transitory memory, wherein the processor is configured to read a set of instructions to:

receive an anthology document associated with a domain;

implement a trained multimodal extraction model to identify one or more individual encounter records within the anthology document, wherein the trained multimodal extraction model is configured to receive at least two input types generated from the anthology document;

for each of the one or more individual encounter records:

generate one or more resource data structures representative of an entity in a selected one of the one or more individual encounter records, wherein the entity is identified by a trained natural language processing model;

generate one or more value set data structures comprising at least one of the one or more resource data structures, at least one code element comprising a code selected from a code library associated with the domain, and at least one linkage element representative of a linkage between the entity of the at least one of the one or more resource data structures and the code element, wherein:

each of the one or more value set data structures comprise a primary resource data structure and at least one secondary resource data structure, wherein the at least one linkage element is representative of a linkage between the entity of the primary resource data structure and the code element, and

each of the one or more value set data structures and each of the one or more resource data structures include one of a plurality of types, and wherein the type of the value set data structure is determined by a type of the primary resource data structure; and

generate instructions configured to display a user interface that includes recommended codes associated with the individual encounter records, the user interface comprising the anthology document, the code selected from the code library, and a visual element representative of the at least one linkage element, wherein the visual element representative of the at least one linkage element is user-selectable so as to associate the code selected from the code library with the individual encounter record.

2 . The system of claim 1 , wherein the multimodal extraction model is configured to receive image input data and text input data.

3 . The system of claim 1 , wherein the trained multimodal extraction model is configured to generate a page prediction for each page in the anthology document, and wherein the one or more individual encounter records are generated by a rules-based module based on the page prediction for each page in the anthology document.

4 . The system of claim 1 , comprising generating encounter metadata for each of the one or more individual encounter records.

5 . The system of claim 1 , wherein the one or more resource data structures each comprise one of a condition resource data structure, a medication resource data structure, an observation resource data structure, or a procedure resource data structure.

6 . The system of claim 1 , wherein each of the one or more resource data structures comprise an annotation data element, at least one context data element, and at least one metadata element.

7 . The system of claim 6 , wherein the entity of each of the one or more resource data structures is defined by the annotation data element.

8 . The system of claim 6 , wherein each of the one or more resource data structures are generated by a rules-based resource construction module based on the annotation data element, the at least one context data element, and the at least one metadata element.

9 . A computer-implemented method, comprising:

receiving an anthology document associated with a domain;

implementing a trained multimodal extraction model to identify one or more individual encounter records within the anthology document;

for each of the one or more individual encounter records:

generating one or more resource data structures comprising an annotation data element representative of an entity in a selected one of the one or more individual encounter records, wherein the entity is identified by a trained natural language processing model;

generating one or more value set data structures comprising at least one of the one or more resource data structures, at least one code element comprising a code selected from a code library associated with the domain, and at least one linkage element representative of a linkage between the entity of the at least one of the one or more resource data structures and the code element, wherein:

each of the one or more value set data structures comprise a primary resource data structure and at least one secondary resource data structure, wherein the at least one linkage element is representative of a linkage between the entity of the primary resource data structure and the code element, and

each of the one or more value set data structures and each of the one or more resource data structures include one of a plurality of types, and wherein the type of the value set data structure is determined by a type of the primary resource data structure; and

generating instructions configured to display a user interface that includes recommended codes associated with the individual encounter records, the user interface comprising the anthology document, the code selected from the code library, and a visual element representative of the at least one linkage element, wherein the visual element representative of the at least one linkage element is user-selectable so as to associate the code selected from the code library with the individual encounter record.

10 . The computer-implemented method of claim 9 , wherein the multimodal extraction model is configured to receive image input data and text input data.

11 . The computer-implemented method of claim 9 , wherein the trained multimodal extraction model is configured to generate a page prediction for each page in the anthology document, and wherein the one or more individual encounter records are generated by a rules-based module based on the page prediction for each page in the anthology document.

12 . The computer-implemented method of claim 9 , wherein each of the one or more resource data structures comprise at least one context data element, at least one metadata element, or a combination thereof.

13 . The computer-implemented method of claim 12 , wherein each of the one or more resource data structures are generated by a rules-based resource construction module based on the annotation data element and the at least one context data element, at least one metadata element, or the combination thereof.

14 . A non-transitory computer readable medium having instructions stored thereon, wherein the instructions, when executed by at least one processor, cause at least one device to perform operations comprising:

receiving an anthology document associated with a domain;

implementing a trained multimodal extraction model to identify one or more individual encounter records within the anthology document, wherein the trained multimodal extraction model is configured to receive a text input and an image input, wherein each of the text input and the image input are generated from the anthology document;

for each of the one or more individual encounter records:

generating one or more resource data structures comprising an annotation data element representative of an entity in a selected one of the one or more individual encounter records, wherein the entity is identified by a trained natural language processing model;

generating one or more value set data structures comprising at least one of the one or more resource data structures, at least one code element comprising a code selected from a code library associated with the domain, and at least one linkage element representative of a linkage between the entity of the at least one of the one or more resource data structures and the code element, wherein:

each of the one or more value set data structures comprise a primary resource data structure and at least one secondary resource data structure, wherein the at least one linkage element is representative of a linkage between the entity of the primary resource data structure and the code element, and

each of the one or more value set data structures and each of the one or more resource data structures include one of a plurality of types, and wherein the type of the value set data structure is determined by a type of the primary resource data structure; and

generating instructions configured to display a user interface that includes recommended codes associated with the individual encounter records, the user interface comprising the anthology document, the code selected from the code library, and a visual element representative of the at least one linkage element, wherein the visual element representative of the at least one linkage element is user-selectable so as to associate the code selected from the code library with the individual encounter record.

15 . The non-transitory computer readable medium of claim 14 , wherein the trained multimodal extraction model is configured to generate a page prediction for each page in the anthology document, and wherein the one or more individual encounter records are generated by a rules-based module based on the page prediction for each page in the anthology document.

16 . The non-transitory computer readable medium of claim 14 , wherein each of the one or more resource data structures comprise at least one context data element, at least one metadata element, or a combination thereof, and wherein each of the one or more resource data structures are generated by a rules-based resource construction module based on the annotation data element and the at least one context data element, at least one metadata element, or the combination thereof.

Assignments (1)

Continuity (2)

References Cited (29)