Methods and apparatus for extracting facts from a medical text
Cascaded models may be applied to extract facts from a medical text. A first model may be applied to at least a portion of the medical text. The first model extracts at least one first medical fact. The at least one first medical fact is linked to at least first text in the at least a portion of the medical text. A second model may be applied to the first text. The second model extracts at least one second fact that is an attribute of the at least one first medical fact.
1. A method of extracting a plurality of facts from a subject text, the method comprising:
extracting the plurality of facts using at least two cascaded fact extraction models, the at least two cascaded fact extraction models comprising a first fact extraction model and at least one other fact extraction model, wherein the extracting using the at least two cascaded fact extraction models comprises:
applying, using at least one computer processor, the first fact extraction model to first text that is at least a portion of the subject text, to extract at least one first fact identifying a problem, medication, procedure and/or allergy, wherein the at least one first fact is extracted from the first text of the subject text;
selecting, based at least in part on the at least one first fact extracted from the first text and from among the at least one other fact extraction model, at least one second fact extraction model to be applied to the first text for extraction of one or more additional facts related to the at least one first fact, the one or more additional facts related to the at least one first fact comprising an attribute of the at least one first fact;
providing, from the first fact extraction model to the at least one second fact extraction model and as a feature of the at least one second fact extraction model, the first text and/or information identifying a location of the first text in the subject text;
applying the at least one second fact extraction model to the first text using the at least one computer processor; and
in response to extracting, from the applying of the at least one second fact extraction model to the first text, at least one additional fact comprising the attribute of the at least one first fact, storing the at least one first fact and the at least one additional fact.
2. The method of claim 1 , further comprising
providing, from the first fact extraction model to the at least one second fact extraction model, and as a feature of the at least one second fact extraction model, the at least one first fact.
3. The method of claim 1 , wherein
the providing comprises providing the first text.
4. The method of claim 1 , wherein the at least one first fact identifies a medication.
5. The method of claim 4 , wherein the at least one additional fact identifies a dosage and/or frequency of administration of the medication.
6. The method of claim 1 , wherein the at least one first fact identifies a problem, and the at least one additional fact identifies a laterality of the problem.
7. The method of claim 1 , wherein the at least one additional fact identifies a negation of one or more of the at least one first fact.
8. At least one non-transitory computer-readable storage medium encoded with computer-executable instructions that, when executed, perform a method of extracting a plurality of facts from a subject text using cascaded models, the subject text being a first text, the method comprising:
extracting the plurality of facts using at least two cascaded fact extraction models, the at least two cascaded fact extraction models comprising a first fact extraction model and at least one other fact extraction model, wherein the extracting using the at least two cascaded fact extraction models comprises:
applying, using at least one computer processor, the first fact extraction model to first text that is at least a portion of the subject text, to extract at least one first fact identifying a problem, medication, procedure and/or allergy, wherein the at least one first fact is extracted from the first text of the subject text;
selecting, based at least in part on the at least one first fact extracted from the first text and from among the at least one other fact extraction model, at least one second fact extraction model to be applied to the first text for extraction of one or more additional facts related to the at least one first fact, the one or more additional facts related to the at least one first fact comprising an attribute of the at least one first fact;
providing, from the first fact extraction model to the at least one second fact extraction model and as a feature of the at least one second fact extraction model, the first text and/or information identifying a location of the first text in the subject text;
applying the at least one second fact extraction model to the first text using the at least one computer processor; and
in response to extracting, from the applying of the at least one second fact extraction model to the first text, at least one additional fact comprising the attribute of the at least one first fact, storing the at least one first fact and the at least one additional fact.
9. The at least one non-transitory computer-readable storage medium of claim 8 , further comprising
providing, from the first fact extraction model to the at least one second fact extraction model, and as a feature of the at least one second fact extraction model, the at least one first fact.
10. The at least one non-transitory computer-readable storage medium of claim 8 , wherein
the providing comprises providing the first text.
11. The at least one non-transitory computer-readable storage medium of claim 8 , wherein the at least one first fact identifies a medication.
12. The at least one non-transitory computer-readable storage medium of claim 11 , wherein the at least one additional fact identifies a dosage and/or frequency of administration of the medication.
13. The at least one non-transitory computer-readable storage medium of claim 8 , wherein the at least one first fact identifies a problem, and the at least one additional fact identifies a laterality of the problem.
14. The at least one non-transitory computer-readable storage medium of claim 8 , wherein the at least one additional fact identifies a negation of one or more of the at least one first fact.
15. An apparatus comprising:
at least one processor; and
at least one memory storing processor-executable instructions that, when executed by the at least one processor, perform a method of extracting a plurality of facts from a subject text using cascaded models, the subject text being a first text, the method comprising:
extracting the plurality of facts using at least two cascaded fact extraction models, the at least two cascaded fact extraction models comprising a first fact extraction model and at least one other fact extraction model, wherein the extracting using the at least two cascaded fact extraction models comprises:
applying, using at least one computer processor, the first fact extraction model to first text that is at least a portion of the subject text, to extract at least one first fact identifying a problem, medication, procedure and/or allergy, wherein the at least one first fact is extracted from the first text of the subject text;
selecting, based at least in part on the at least one first fact extracted from the first text and from among the at least one other fact extraction model, at least one second fact extraction model to be applied to the first text for extraction of one or more additional facts related to the at least one first fact, the one or more additional facts related to the at least one first fact comprising an attribute of the at least one first fact;
providing, from the first fact extraction model to the at least one second fact extraction model and as a feature of the at least one second fact extraction model, the first text and/or information identifying a location of the first text in the subject text;
applying the at least one second fact extraction model to the first text using the at least one computer processor; and
in response to extracting, from the applying of the at least one second fact extraction model to the first text, at least one additional fact comprising the attribute of the at least one first fact, storing the at least one first fact and the at least one additional fact.
16. The apparatus of claim 15 , further comprising
providing, from the first fact extraction model to the at least one second fact extraction model, and as a feature of the at least one second fact extraction model, the at least one first fact.
17. The apparatus of claim 15 , wherein
the providing comprises providing the first text.