Systems and methods for natural language processing-based classification of electronic medical records
Systems and methods of the present disclosure enable improved natural language processing of patient-related medical information for clinical decision support. To do so, a processor receives patient data including a written report, and accesses a dictionary of terminology associated with a disease. The terminology includes descriptors indicative of categories of the disease. The processor inputs the written report into a tokenization function to output tokens by parsing word patterns in the written report, and generating the tokens from the word patterns. The processor determines a presence in the written report of each descriptor based on the tokens and determines a category-specific score associated with each category based on the presence of the descriptors. The processor determines a category recommendation score indicative of a particular category based on the category-specific scores and generates a category recommendation representing the particular category based on the category recommendation score.
1 . A method, comprising:
receiving, by at least one processor associated with at least one electronic health record (EHR) system, patient data comprising at least one written report associated with at least one patient;
accessing, by the at least one processor, a dictionary of terminology associated with mitral regurgitation (MR);
wherein the terminology includes a plurality of descriptors;
wherein each descriptor is indicative of at least one category of a plurality of categories associated with the MR;
parsing, by the at least one processor, at least one word pattern in the at least one written report to produce a set of tokens from the at least one word pattern;
inputting, by the at least one processor, each token in the set of tokens of the plurality of descriptors into a plurality of natural language processing pipelines, each respective natural language processing pipeline being associated with at least one respective detected category of the plurality of categories of the MR and configured to detect, separately and in parallel, at least one respective detected category specific descriptors in the plurality of descriptors, the at least one respective detected category specific descriptors being associated with the at least one respective detected category of the plurality of categories of the MR;
utilizing, by the at least one processor, at least one category-specific rule parameter to determine a category-specific score of a plurality of category-specific scores associated with each category of the plurality of categories;
utilizing, by the at least one processor, at least one recommendation rule parameter to determine at least one recommended category score indicative of at least one particular category of the MR based at least in part on the category-specific score associated with each category;
generating, by the at least one processor, at least one category recommendation representing a particular category of the MR for the at least one patient based at least in part on the at least one recommended category score;
causing to produce, by the at least one processor, graphical user interface (GUI) associated with the EHR, the GUI displaying EHR data for the at least one patient augmented with the at least one recommended category score to present to a user the at least one category recommendation so as to provide clinical decision support for administration of transcatheter edge-to-edge repair (TEER) to effect repair of a mitral valve of the patient in response to the at least one category recommendation for a particular category of the MR of the patient; and
administering the TEER to the at least one patient based at least in part on the at least one category recommendation to repair the mitral valve.
2 . The method of claim 1 , further comprising:
inputting, by the at least one processor, each descriptor of the plurality of descriptors into a tokenization function to output a set of descriptor tokens; and
searching, by the at least one processor, the set of tokens with the set of descriptor tokens to identify occurrences of each descriptor based at least in part on a match of at least one descriptor token of each descriptor to at least one token of the written report.
3 . The method of claim 1 , wherein the plurality of descriptors comprises a plurality of combinations of a plurality of descriptor tokens, wherein each combination of the plurality of combinations represents a particular descriptor of the plurality of descriptors.
4 . The method of claim 3 , wherein the plurality of combinations and the plurality of descriptor tokens are hand-crafted.
5 . The method of claim 1 , wherein the plurality of categories associated with mitral regurgitation comprises:
a primary degenerative mitral regurgitation category indicative of primary degenerative mitral regurgitation, and
a secondary mitral regurgitation category indicative of secondary mitral regurgitation.
6 . The method of claim 5 , further comprising:
determining the at least one category recommendation as mixed mitral regurgitation indicative of a combination of the primary degenerative mitral regurgitation and the secondary mitral regurgitation based at least in part on the at least one set of rules indicating a presence of primary degenerative mitral regurgitation and a presence of secondary mitral regurgitation.
7 . The method of claim 5 , further comprising:
determining the at least one category recommendation as unknown mitral regurgitation based at least in part on the at least one set of rules not indicating a presence of primary degenerative mitral regurgitation and not indicating a presence of secondary mitral regurgitation.
8 . The method of claim 1 , further comprising:
determining, by the at least one processor, the at least one category recommendation indicative of the at least one particular category based at least in part on:
a category-specific score associated with each category, and
at least one medical test result.
9 . The method of claim 1 , further comprising:
determining, by the at least one processor, the at least one category recommendation indicative of the at least one particular category based at least in part on:
a category-specific score associated with each category, and
a set of classification parameters for balancing each category-specific score so as to select a particular category of the plurality of categories.
10 . A system comprising:
at least one processor associated with at least one electronic health record (EHR) system and in communication with at least one non-transitory computer readable medium having software instructions stored thereon, wherein, upon execution of the software instructions, the at least one processor is configured to:
receiving, by the at least one processor, patient data comprising at least one written report associated with at least one patient;
accessing, by the at least one processor, a dictionary of terminology associated with mitral regurgitation (MR);
wherein the terminology includes a plurality of descriptors;
wherein each descriptor is indicative of at least one category of a plurality of categories associated with the MR;
parsing, by the at least one processor, at least one word pattern in the at least one written report to produce a set of tokens from the at least one word pattern;
inputting, by the at least one processor, each token in the set of tokens of the plurality of descriptors into a plurality of natural language processing pipelines, each respective natural language processing pipeline being associated with at least one respective detected category of the plurality of categories of the MR and configured to detect, separately and in parallel, at least one respective detected category specific descriptors in the plurality of descriptors, the at least one respective detected category specific descriptors being associated with the at least one respective detected category of the plurality of categories of the MR;
utilizing, by the at least one processor, at least one category-specific rule parameter to determine a category-specific score of a plurality of category-specific scores associated with each category of the plurality of categories;
utilizing, by the at least one processor, at least one recommendation rule parameter to determine at least one recommended category score indicative of at least one particular category of the MR based at least in part on the category-specific score associated with each category;
generating, by the at least one processor, at least one category recommendation representing a particular category of the MR for the at least one patient based at least in part on the at least one recommended category score; and
causing to produce, by the at least one processor, graphical user interface (GUI) associated with the EHR, the GUI displaying EHR data for the at least one patient augmented with the at least one recommended category score to present to a user the at least one category recommendation so as to provide clinical decision support for administration of transcatheter edge-to-edge repair (TEER) to effect repair of a mitral valve of the patient in response to the at least one category recommendation for a particular category of the MR of the patient; and
administering the TEER to the at least one patient based at least in part on the at least one category recommendation to repair the mitral valve.
11 . The system of claim 10 , wherein, upon execution of the software instructions, the at least one processor is further configured to:
input each descriptor of the plurality of descriptors into a tokenization function to output a set of descriptor tokens; and
search the set of tokens with the set of descriptor tokens to identify occurrences of each descriptor based at least in part on a match of at least one descriptor token of each descriptor to at least one token of the written report.
12 . The system of claim 10 , wherein the plurality of descriptors comprises a plurality of combinations of a plurality of descriptor tokens, wherein each combination of the plurality of combinations represents a particular descriptor of the plurality of descriptors.
13 . The system of claim 12 , wherein the plurality of combinations and the plurality of descriptor tokens are hand-crafted.
14 . The system of claim 10 , wherein the plurality of categories associated with mitral regurgitation comprises:
a primary degenerative mitral regurgitation category indicative of primary degenerative mitral regurgitation, and
a secondary mitral regurgitation category indicative of secondary mitral regurgitation.
15 . The system of claim 14 , wherein, upon execution of the software instructions, the at least one processor is further configured to:
determine the at least one category recommendation as mixed mitral regurgitation indicative of a combination of the primary degenerative mitral regurgitation and the secondary mitral regurgitation based at least in part on at least one set of rules indicating a presence of primary degenerative mitral regurgitation and a presence of secondary mitral regurgitation.
16 . The system of claim 14 , wherein, upon execution of the software instructions, the at least one processor is further configured to:
determine the at least one category recommendation as unknown mitral regurgitation based at least in part on the at least one set of rules not indicating a presence of primary degenerative mitral regurgitation and not indicating a presence of secondary mitral regurgitation.
17 . The system of claim 10 , wherein, upon execution of the software instructions, the at least one processor is further configured to:
determine at least one category recommendation indicative of the at least one particular category based at least in part on:
a category-specific score associated with each category, and
at least one medical test result.
18 . The system of claim 10 , wherein, upon execution of the software instructions, the at least one processor is further configured to:
determine at least one category recommendation score indicative of the at least one particular category based at least in part on:
a category-specific score associated with each category, and
a set of classification parameters for balancing each category-specific score so as to select a particular category of the plurality of categories.