Granted Patent B2

US 12,293,845 · App. 17/809,416 · Granted May 6, 2025

Inventors: Saigeetha Aswathnarayanan Jegannathan (Bangalore, IN); Sridhar Jonnala (Bangalore, IN); V Datta Kamesam Jami (Srikakulam, IN); Chinthalapudi Venkata Sai Vishnu Vardhan (Kandukur, IN); Naman Mathur (Jaipur, IN); Shivangi Tak (Gurgaon, IN); Kartikeya Vats (Dehradun, IN)

Assignee: International Business Machines Corporation

G16H70/40G06F40/289G06F40/40G06N5/022

Abstract

Claims (84)

1. A computer-implemented method comprising:

receiving, by one or more processors, a pair of documents from a user, wherein the pair of documents include a Scientific Drug Label and a Patient Drug Label, further comprising:

converting, by the one or more processors, the pair of documents from a word format to a Portable Document Format (PDF);

extracting, by the one or more processors, content from the PDF of the Scientific Drug Label in a structured format; and

sorting, by the one or more processors, the extracted content into a corresponding section heading or a corresponding section subheading;

converting, by the one or more processors, a complex medical language of the Scientific Drug Label into a simplified patient friendly language;

analyzing, by the one or more processors, the simplified patient friendly language to identify one or more words, one or more phrases, or one or more sentences that have been modified, inserted, or deleted;

responsive to determining the one or more words, the one or more phrases, or the one or more sentences are relevant to a patient, classifying, by the one or more processors, the one or more words, the one or more phrases, or the one or more sentences in one or more categories;

searching, by the one or more processors, for a location in the Patient Drug Label that closely maps to the one or more words, the one or more phrases, or the one or more sentences to the Scientific Drug Label;

incorporating, by the one or more processors, the one or more words, the one or more phrases, or the one or more sentences in a mapped location of the Patient Drug Label;

outputting, by the one or more processors, an updated Patient Drug Label to the user;

subsequent to outputting the updated Patient Drug Label to the user, requesting, by the one or more processors, feedback from the user;

responsive to receiving the feedback from the user, validating, by the one or more processors, the feedback received from the user manually using a confidence score of one or more intermediate outputs; and

annotating, by the one or more processors, the feedback received from the user.

2. The computer-implemented method of claim 1 , wherein the one or more categories include a complete insertion of a sentence in the Scientific Drug Label, a complete deletion of the sentence in the Scientific Drug Label, and an insertion or a deletion of a word or a phrase in the Scientific Drug Label.

3. The computer-implemented method of claim 1 , further comprising:

subsequent to annotating the feedback received from the user, identifying, by the one or more processors, one or more engines to be retrained; and

retraining, by the one or more processors, the one or more engines with the annotated feedback.

4. The computer-implemented method of claim 1 , wherein the feedback received from the user includes an acceptance or a rejection of one or more changes incorporated into the updated Patient Drug Label.

5. The computer-implemented method of claim 1 , further comprising:

subsequent to sorting the extracted content into the corresponding section heading or the corresponding section subheading, extracting, by the one or more processors, one or more keywords and one or more key phrases from the extracted content using a custom-trained Spacey model to understand a concept of each sentence of the extracted content;

mapping, by the one or more processors, one or more relationships between the Scientific Drug Label and the Patient Drug Label using a knowledge graph; and

extracting, by the one or more processors, a confidence score for the one or more relationships mapped between the Scientific Drug Label and the Patient Drug Label.

6. The computer-implemented method of claim 5 , wherein the one or more keywords and the one or more key phrases extracted from structured data includes a name of a drug, a composition of the drug, a shape of the drug, an appearance of the drug, a medical condition treated by the drug, a method to administer the drug, one or more side effects the patient may experience when taking the drug, and one or more precautions the patient should take when using the drug.

7. The computer-implemented method of claim 5 , wherein extracting the one or more keywords and the one or more key phrases from the extracted content using the custom-trained Spacey model to understand the concept of each sentence of the extracted content further comprises:

generating, by the one or more processors, a model for the one or more keywords and the one or more key phrases from the extracted content;

determining, by the one or more processors, a position of the one or more keywords and the one or more key phrases from the extracted content;

annotating, by the one or more processors, the position of the one or more keywords and the one or more key phrases from the extracted content; and

training, by the one or more processors, on the annotated position of the one or more keywords and the one or more key phrases.

8. The computer-implemented method of claim 1 , wherein incorporating the one or more words, the one or more phrases, or the one or more sentences in the mapped location of the Patient Drug Label further comprises:

modifying, by the one or more processors, the one or more words, the one or more phrases, or the one or more sentences in the mapped location of the Patient Drug Label.

9. The computer-implemented method of claim 1 , wherein incorporating the one or more words, the one or more phrases, or the one or more sentences in the mapped location of the Patient Drug Label further comprises:

inserting, by the one or more processors, the one or more words, the one or more phrases, or the one or more sentences in the mapped location of the Patient Drug Label.

10. The computer-implemented method of claim 1 , wherein incorporating the one or more words, the one or more phrases, or the one or more sentences in the mapped location of the Patient Drug Label further comprises:

deleting, by the one or more processors, the one or more words, the one or more phrases, or the one or more sentences in the mapped location of the Patient Drug Label.

11. A computer program product comprising:

one or more computer readable storage media and program instructions stored on the one or more computer readable storage media, the program instructions comprising:

program instructions to receive a pair of documents from a user, wherein the pair of documents include a Scientific Drug Label and a Patient Drug Label, further comprising:

program instructions to convert the pair of documents from a word format to a Portable Document Format (PDF);

program instructions to extract content from the PDF of the Scientific Drug Label in a structured format; and

program instructions to sort the extracted content into a corresponding section heading or a corresponding section subheading;

program instructions to convert a complex medical language of the Scientific Drug Label into a simplified patient friendly language;

program instructions to analyze the simplified patient friendly language to identify one or more words, one or more phrases, or one or more sentences that have been modified, inserted, or deleted;

responsive to determining the one or more words, the one or more phrases, or the one or more sentences are relevant to a patient, program instructions to classify the one or more words, the one or more phrases, or the one or more sentences in one or more categories;

program instructions to search for a location in the Patient Drug Label that closely maps to the one or more words, the one or more phrases, or the one or more sentences to the Scientific Drug Label;

program instructions to incorporate the one or more words, the one or more phrases, or the one or more sentences in a mapped location of the Patient Drug Label;

program instructions to output an updated Patient Drug Label to the user;

responsive to receiving a feedback from the user, program instructions to validate, the feedback received from the user manually using a confidence score of one or more intermediate outputs; and

responsive to annotate the feedback received from the user.

12. The computer program product of claim 11 , further comprising:

subsequent to sorting the extracted content into the corresponding section heading or the corresponding section subheading, program instructions to extract one or more keywords and one or more key phrases from the extracted content using a custom-trained Spacey model to understand a concept of each sentence of the extracted content;

program instructions to map one or more relationships between the Scientific Drug Label and the Patient Drug Label using a knowledge graph; and

program instructions to extract a confidence score for the one or more relationships mapped between the Scientific Drug Label and the Patient Drug Label.

13. The computer program product of claim 12 , wherein extracting the one or more keywords and the one or more key phrases from the extracted content using the custom-trained Spacey model to understand the concept of each sentence of the extracted content further comprises:

program instructions to generate a model for the one or more keywords and the one or more key phrases from the extracted content;

program instructions to determine a position of the one or more keywords and the one or more key phrases from the extracted content;

program instructions to annotate the position of the one or more keywords and the one or more key phrases from the extracted content; and

program instructions to train on the annotated position of the one or more keywords and the one or more key phrases.

14. A computer system comprising:

one or more computer processors;

one or more computer readable storage media;

program instructions collectively stored on the one or more computer readable storage media for execution by at least one of the one or more computer processors, the stored program instructions comprising:

one or more computer readable storage media and program instructions stored on the one or more computer readable storage media, the program instructions comprising:

program instructions to receive a pair of documents from a user, wherein the pair of documents include a Scientific Drug Label and a Patient Drug Label, further comprising:

program instructions to convert the pair of documents from a word format to a Portable Document Format (PDF);

program instructions to extract content from the PDF of the Scientific Drug Label in a structured format; and

program instructions to sort the extracted content into a corresponding section heading or a corresponding section subheading;

program instructions to convert a complex medical language of the Scientific Drug Label into a simplified patient friendly language;

program instructions to analyze the simplified patient friendly language to identify one or more words, one or more phrases, or one or more sentences that have been modified, inserted, or deleted;

responsive to determining the one or more words, the one or more phrases, or the one or more sentences are relevant to a patient, program instructions to classify the one or more words, the one or more phrases, or the one or more sentences in one or more categories;

program instructions to search for a location in the Patient Drug Label that closely maps to the one or more words, the one or more phrases, or the one or more sentences to the Scientific Drug Label;

program instructions to incorporate the one or more words, the one or more phrases, or the one or more sentences in a mapped location of the Patient Drug Label;

program instructions to output an updated Patient Drug Label to the user;

responsive to receiving a feedback from the user, program instructions to validate, the feedback received from the user manually using a confidence score of one or more intermediate outputs; and

responsive to annotate the feedback received from the user.

15. The computer system of claim 14 , further comprising:

subsequent to sorting the extracted content into the corresponding section heading or the corresponding section subheading, program instructions to extract one or more keywords and one or more key phrases from the extracted content using a custom-trained Spacey model to understand a concept of each sentence of the extracted content;

program instructions to map one or more relationships between the Scientific Drug Label and the Patient Drug Label using a knowledge graph; and

program instructions to extract a confidence score for the one or more relationships mapped between the Scientific Drug Label and the Patient Drug Label.

16. The computer system of claim 15 , wherein extracting the one or more keywords and the one or more key phrases from the extracted content using the custom-trained Spacey model to understand the concept of each sentence of the extracted content further comprises:

program instructions to generate a model for the one or more keywords and the one or more key phrases from the extracted content;

program instructions to determine a position of the one or more keywords and the one or more key phrases from the extracted content;

program instructions to annotate the position of the one or more keywords and the one or more key phrases from the extracted content; and

program instructions to train on the annotated position of the one or more keywords and the one or more key phrases.

Assignments (1)

Continuity (1)

References Cited (18)