AUTOMATED CLINICAL INDICATOR RECOGNITION WITH NATURAL LANGUAGE PROCESSING

• US 20150066539A1
• Filed: 09/05/2014
• Published: 03/05/2015
• Est. Priority Date: 09/05/2013
• Status: Active Grant

• Current Assignee: Optum360, LLC (UnitedHealth Group Incorporated)
• Original Assignee: A-Life Medical, LLC
• Inventors: Morsch, Mark, Sheffer, Ronald, Wieczorek, Michelle
• Granted Patent: US 10,541,053 B2
• Time in Patent Office: Days
• US Class Current: 705/3
• CPC Class Codes: G16H 10/60 for patient-specific data, e.g. for electronic patient records
• CPC Class Codes: G16H 50/20 for computer-aided diagnosis, e.g. based on medical expert systems

Abstract

Computer-based, natural language processing systems and methods are provided for review of clinical documentation and other medical records, and for clinical documentation improvement. The systems and methods are configured to analyze received diagnoses and/or procedures in view of documents in the record using a natural language processor and a tiered information model to identify clinical indicators, and optionally markers. The identified information is compared with the received data for use in generating queries requesting evidence in support of the received diagnosis or procedure, or for use in validating the received information.

Claims

1. A computer-based, natural language processing system comprising:

a computer processor operatively coupled to an interactive display;

a natural language processor module executing on the computer processor; and

a database in communication with the computer processor, the database comprising memory for storing a medical record, the medical record comprising documents relating to patient care;

the computer processor configured to perform the steps of;

receiving a diagnosis for a patient associated with the medical record;

identifying clinical indicators and associated contextual information in the documents of the medical record at least using the natural language processor module;

analyzing the clinical indicators and associated contextual information in relation to rule-based scenarios in an information model;

generating markers according to the analysis, wherein the markers indicate a medical diagnosis;

identifying expected markers for the received diagnosis for the patient;

comparing the generated markers with the identified expected markers;

if the markers identified for the received diagnosis do not match one or more generated markers, then

generating a query requesting evidence missing from the medical record for supporting the received diagnosis, the missing evidence comprising clinical indicators required to generate scenarios and markers from the medical record for the received diagnosis, and

if the markers identified for the received diagnosis match the one or more generated markers, then

validating the received diagnosis.

2. The system of claim 1, wherein at least the natural language processor module performs the step of identifying clinical indicators by analyzing the clinical indicators according to a type of document in which the clinical indicator is located and according to one or more contextual factors related to the clinical indicator, the contextual factors including when the clinical indicator was mentioned in a hospital stay, when the clinical indicator was mentioned in a course of treatment, negation data associated with the clinical indicator, whether the clinical indicator was recorded as under consideration or ordered, or how many times the clinical indicator is repeated throughout the medical record.

3. The system of claim 1, wherein the step of identifying clinical indicators and associated contextual information further comprises synthesizing clinical indicators using information contained within a plurality of documents, and wherein the step of synthesizing clinical indicators comprises using the natural language processing module to interpret text into multiple sentences.

4. The system of claim 1, wherein if the markers identified for the received diagnosis do not match one or more generated markers, then the computer processor is further configured to perform the step of:

generating a query for a missing diagnosis, the query describing the one or more generated markers, in combination with evidence describing at least one of the clinical indicators and associated contextual information from which the marker was generated.

5. The system of claim 1, wherein the step of identifying clinical indicators and associated contextual information comprises analyzing the clinical indicators in relation to identified sections within the document containing the clinical indicators or in relation to a type of user creating the document containing the clinical indicators.

6. The system of claim 1, further comprising assigning the generated markers a confidence level and wherein the confidence level is used in the step of comparing the generated markers with the identified expected markers.

7. The system of claim 6, wherein a relatively high confidence level for one of the generated markers that matches one of the identified expected markers results in validating the received diagnosis.

8. The system of claim 1, wherein the query requesting evidence missing from the patient medical record for supporting the received diagnosis further comprises identification of one or more documents relating to patient care missing from the medical record required to be analyzed to identify clinical indicators and generate scenarios and markers for the received diagnosis.

9. A computer-based, natural language processing system comprising:

a computer processor operatively coupled to an interactive display;

a natural language processor module executing on the computer processor; and

a database in communication with the computer processor, the database comprising memory for storing a medical record, the medical record comprising documents relating to patient care;

the computer processor configured to perform the steps of;

receiving medical procedure data for a patient associated with the medical record;

identifying clinical indicators and associated contextual information in the documents of the medical record at least using the natural language processor module;

analyzing the clinical indicators and associated contextual information in relation to rule-based scenarios in an information model;

generating markers according to the analysis, wherein the markers identify a medical diagnosis;

identifying one or more expected medical procedures for treating the identified medical diagnosis corresponding to the generated markers;

comparing the received medical procedure data with the identified one or more expected medical procedures;

if the received medical procedure does not match one of the one or more identified expected medical procedures, then

generating a query requesting evidence missing from the medical record for supporting the received medical procedure data, the missing evidence comprising clinical indicators required to generate scenarios and markers that identify a diagnosis treated by the received medical procedure data, and

if the received medical procedure matches one of the one or more identified medical procedures, then

validating the received medical procedure data.

10. The system of claim 9, further comprising assigning the generated markers a confidence level, and wherein the confidence level is used in the step of comparing the received medical procedure data with the identified one or more medical procedures and a relatively high confidence level for one of the generated markers for an identified medical procedure that matches the received medical procedure data results in validating the received medical procedure data.

11. The system of claim 9, wherein if the received medical procedure does not match one of the one or more expected medical procedures, then the computer processor is further configured to perform the step of:

generating a query for a missing medical procedure, the query describing one or more of the generated markers, in combination with evidence describing at least one of the clinical indicators and associated contextual information from which the marker was generated.

12. The system of claim 9, wherein the query requesting evidence missing from the patient medical record for supporting the received diagnosis further comprises identifying one or more documents relating to patient care missing from the medical record that are required to be analyzed to identify clinical indicators and generate scenarios and markers for the received medical procedure data.

13. A computer-based, natural language processing system comprising:

a computer processor operatively coupled to an interactive display;

a natural language processor module executing on the computer processor; and

a database in communication with the computer processor, the database comprising memory for storing a medical record, the medical record comprising documents relating to patient care;

the computer processor configured to perform the steps of;

receiving a diagnosis for a patient associated with the medical record;

generating expected markers using the received medical diagnosis, wherein the markers are generated from an information model and are associated with expected clinical indicators related to the received medical diagnosis;

identifying clinical indicators in the documents of the medical record at least using the natural language processor module;

comparing the expected clinical indicators related to the received medical diagnosis to the identified clinical indicators in the documents;

generating a query requesting evidence missing from the medical record responsive to a determination that the compared clinical indicators do not match, wherein the missing evidence comprises clinical indicators required to generate markers from the medical record for the received diagnosis.

14. The system of claim 13, wherein if the processor determines the compared clinical indicators do not match, the computer processor is further configured to perform the steps of

analyzing the identified clinical indicators in the documents and associated contextual information in relation to rule-based scenarios in an information model;

generating markers according to the analysis, wherein the markers indicate a medical diagnosis; and

generating a query for a missing diagnosis responsive to the determination that the compared clinical indicators do not match, the query describing the generated markers, in combination with evidence describing at least one of the clinical indicators and associated contextual information from which the marker was generated.

15. The system of claim 13, further comprising the step of validating the received diagnosis responsive to a determination that the compared clinical indicators match.

16. The system of claim 13, wherein at least the natural language processor module performs the step of identifying clinical indicators by analyzing the clinical indicators according to a type of document in which the clinical indicator is located and according to one or more contextual factors related to the clinical indicator, the contextual factors including when the clinical indicator was mentioned in a hospital stay, when the clinical indicator was mentioned in a course of treatment, negation data associated with the clinical indicator, whether the clinical indicator was recorded as under consideration or ordered, or how many times the clinical indicator is repeated throughout the medical record.

17. The system of claim 13, wherein the step of receiving the diagnosis involves receiving the diagnosis from the medical record.

18. The system of claim 13, wherein the step of identifying clinical indicators and associated contextual information comprises analyzing the clinical indicators in relation to at least one of identified sections within the document containing the clinical indicators or a type of user creating the document containing the clinical indicators.

19. The system of claim 13, wherein the query requesting evidence missing from the patient medical record for supporting the received diagnosis further comprises identification of one or more documents relating to patient care missing from the medical record that are required to be analyzed to identify clinical indicators for the received diagnosis.

20. The system of claim 13, wherein the step of identifying clinical indicators further comprises synthesizing clinical indicators using information contained within a plurality of documents, wherein the step of synthesizing clinical indicators comprises using the natural language processing module to interpret text into multiple sentences.