Health information transformation system

• US 10,628,553 B1
• Filed: 10/08/2012
• Issued: 04/21/2020
• Est. Priority Date: 12/30/2010
• Status: Active Grant

• Current Assignee: Cerner Innovation, Inc. (Oracle Corporation)
• Original Assignee: Cerner Innovation, Inc. (Oracle Corporation)
• Inventors: Murrish, John Christopher, Kailasam, Kanakasabha, McNair, Douglas S., Ash, Michael Alan, Fangman, Thomas Anthony
• Primary Examiner(s): Adams, Charles D
• Application Number:
  US13/647,187
• Time in Patent Office: 2,752 Days
• Field of Search: None
• CPC Class Codes: G06F 16/2228 Indexing structures
• CPC Class Codes: G06F 16/2365 Ensuring data consistency and integrity
• CPC Class Codes: G06F 16/951 Indexing; Web crawling techniques
• CPC Class Codes: G16H 10/60 for patient-specific data, e.g. for electronic patient records
• CPC Class Codes: G16H 50/70 for mining of medical data, e.g. analysing previous cases of other patients
• CPC Class Codes: G16H 70/00 ICT specially adapted for the handling or processing of medical references

Abstract

Methods, computer systems, and computer readable media for transforming raw healthcare data into relevant healthcare data are provided. Transformation of raw healthcare data into relevant data is accomplished by receiving raw data from a plurality of disparate data sources, indexing the data, and mapping the data. Embodiments transform raw data into relevant data through the use of natural language processing, synonymy and ontology mapping, and adaptive knowledge processing. The relevant data is stored in a one record that includes information necessary to understand and represent that patient.

Claims

1. One or more non-transitory computer hardware devices having computer-executable instructions embodied thereon, that, when executed, implement a method for transforming raw healthcare data into usable healthcare data, the method comprising:

receiving raw health information from at least one data collector, the data collector accessing the raw health information from one or more raw data sources;

indexing a portion of the raw health information to create a searchable index;

transforming the indexed portion of the raw health information into usable data corresponding to a patient, using one or more content-mapping services, the transforming comprising;

(1) sorting the portion of the raw health information into structured data of standard nomenclature, structured data of non-standard nomenclature, and unstructured raw data;

(2) when the portion of the raw health information is unstructured raw data, identifying a first discrete clinical concept having an associated codified meaning and a second discrete clinical concept without a codified meaning via natural language processing, the first and second discrete clinical concepts identified based on a level of detail in the unstructured raw data;

(3) transforming the first discrete clinical concept into a first health concept that is structured data of standard nomenclature by applying the associated codified meanings, and transforming the second discrete clinical concept into a second health concept that is structured data of standard nomenclature by associating the second discrete clinical concept with a broader health concept having the codified meaning, the codified meaning and the associated codified meaning based on one or more standardized clinical codes; and

(4) mapping the first and second health concepts to a relevant piece of information based on nomenclature and ontology mapping, the ontology mapping providing discovery of nomenclatural linkages between pairs of terms or combinations of terms that are extant on multiple different health information systems, at least two of the different health information systems not sharing at least one of a set of unified codesets, nomenclatures, or ontologies to tag or code the documents and records;

storing the usable data into a data repository; and

utilizing the usable data by a computer application to perform a content-mapping, the content-mapping service comprising a synonymic-discovery service or an ontology mapping service.

2. The non-transitory computer hardware devices of claim 1, further comprising storing a sacred copy of the received raw health information.

3. The non-transitory computer hardware devices of claim 1, wherein mapping the first and second concepts to the relevant piece of information further comprises performing record linkage on the portion of the raw health information.

4. The non-transitory computer hardware devices of claim 3, wherein mapping the first and second health concepts to the relevant piece of information further comprises performing reconciliation on the portion of the raw health information, and wherein performing reconciliation comprises eliminating duplicative information, removing outliers, filling in holes, or correcting inconsistent information.

5. The non-transitory computer hardware devices of claim 1, wherein at least one of the ontology-mapping service and the synonymic-discovery service are carried out by a software agent.

6. The non-transitory computer hardware devices of claim 1, wherein indexing the portion of the raw health information includes natural-language processing of the portion of the raw health information.

7. The non-transitory computer hardware devices of claim 1, wherein the transforming further comprises labeling each discrete clinical concept as positive, suspected, or negative based on context of text in the portion of raw health information identified via natural language processing.

8. The non-transitory computer hardware devices of claim 7, wherein labeling each discrete clinical concept comprises:

identifying discrete clinical concepts with codified meanings, and based on a degree of confidence in the identification, assigning a label to each discrete clinical concept, the label indicating each of the discrete clinical concepts as one of positive, suspected, or negative.

9. The non-transitory computer hardware devices of claim 1, wherein the relevant piece of information comprises data represented by a standardized structure, nomenclature, format, or ontology.

10. The non-transitory computer hardware devices of claim 1, wherein the one or more raw data sources comprise at least two disparate sources.

11. The non-transitory computer hardware devices of claim 1, wherein the at least one data collector comprises at least one of an electronic health record crawler, a clinical and administrative data collector, a claims data collector, a document feed collector, or a medical device feed collector.

12. The non-transitory computer hardware devices of claim 1, wherein the raw health information comprises historical health information.

13. The non-transitory computer hardware devices of claim 1, wherein the computer application includes a virtual clinical trial.

14. A system for transforming raw healthcare data into relevant healthcare data, the system comprising:

a staging area repository for receiving raw health information from at least a first source of a plurality of sources of raw health information;

one or more processors;

memory, the memory comprising;

a crawler agent for accessing the raw health information from the first source and providing a copy of the raw health information to the staging area;

an observer agent for identifying a plurality of discrete clinical concepts with corresponding clinical codes from the copy of raw health information in the staging area; and

a concept mapping agent for transforming a portion of the copy of raw health information into relevant data associated with a patient based on the identified concepts using one or more content mapping services, the transforming comprising;

(1) sorting the portion of the raw health information into structured data of standard nomenclature, structured data of non-standard nomenclature, and unstructured raw data;

(2) when the portion of the raw health information is unstructured raw data, identifying a plurality of discrete clinical concepts represented in non-standard nomenclature using a natural language processing component;

(3) labeling each of the plurality of discrete clinical concepts as one of positive, suspected, or negative based on context of text in the portion of raw health information identified using the natural language processing component;

(4) transforming the plurality of discrete concepts into a plurality of health concepts by associating the plurality of discrete clinical concepts with broader health concepts having codified meanings and applying the codified meanings to the plurality of discrete clinical concepts, the codified meanings based on one or more standardized clinical codes; and

(5) mapping the plurality of health concepts to a relevant piece of information based in existing relationships within a nomenclature and ontology mapping component, the nomenclature and ontology mapping component associating the plurality of health concepts to one or more broader health concepts, the one or more broader health concepts being structured data of standard nomenclature suitable for processing by a distributed adaptive knowledge engine component; and

a data repository for storing the relevant data, at least one content-mapping service comprising a synonymic-discovery or an ontology mapping service.

15. The system of claim 14, wherein the data repository is accessible to a software application or service component.

16. The system of claim 14, wherein the concept mapping agent is configured for performing a reconciliation service on the portion of the copy of raw health information, including at least one of eliminating duplicative information, removing outliers, filling in holes, or correcting inconsistent information.

17. The system claim 15, wherein the software application comprises an application for facilitating performing virtual clinical trials on the relevant data.

18. One or more non-transitory computer hardware devices having computer-executable instructions embodied thereon, that, when executed, implement a method for generating a universal patient health record, the method comprising:

receiving raw health information from one or more data sources;

from the received raw health information, identifying a plurality of health concepts, the identifying comprising;

(1) sorting the portion of the raw health information into structured data of standard nomenclature, structured data of non-standard nomenclature, and unstructured raw data;

(2) when the portion of the raw health information is unstructured raw data, identifying a plurality of discrete clinical concepts represented in non-standard nomenclature via natural language processing;

(3) labeling each of the plurality of discrete clinical concepts as one of positive, suspected, or negative based on context of text in the portion of raw health information identified via natural language processing; and

(4) transforming the plurality of discrete clinical concepts into the plurality of health concepts by associating the plurality of discrete clinical concepts with broader health concepts having codified meanings and applying the codified meanings to the plurality of discrete clinical concepts, the codified meanings based on one or more standardized clinical codes;

based on the identified plurality of health concepts, transforming at least a portion of the raw health information into relevant data associated with a patient, the transforming comprising;

(1) performing content mapping on the portion of the raw health information using one or more content-mapping services including at least one of a synonymic-discovery service or an ontology mapping service, the ontology mapping service providing discovery of nomenclatural linkages between pairs of terms or combinations of terms that are extant on multiple different health information systems, at least two of the different health information systems not sharing at least one of a set of unified codesets, nomenclatures, or ontologies to tag or code the documents and records;

or

(2) performing patient record linkage on the at least the portion of the raw health information; and

storing the relevant data into a data repository, thereby creating the universal patient health record;

the transformed relevant data comprising data represented by a standardized structure, nomenclature, format, or ontology.

19. The non-transitory computer hardware devices of claim 18, wherein performing content mapping further comprises performing reconciliation on the portion of the raw health information, and wherein performing reconciliation comprises eliminating duplicative information, removing outliers, filling in holes, or correcting inconsistent information.

20. The non-transitory computer hardware devices of claim 18, wherein at least one of the ontology-mapping service and synonymic-discovery service are carried out by a software agent.