System and method of documenting clinical trials

Patent metadata

Field	Value
Publication number	US-11322229-B2
Application number	US-201816143976-A
Country	US
Kind code	B2
Filing date	Sep 27, 2018
Priority date	Sep 27, 2018
Publication date	May 3, 2022
Grant date	May 3, 2022

Abstract

Disclosed is a system for documenting clinical trials, the system when operated identifies at least one publication related to a clinical trial entry to obtain documented clinical trial. The system comprises: an information repository comprising plurality of publications; a clinical trials registry database comprising a plurality of clinical trial entries; and a server arrangement. The server arrangement is configured to: obtain the plurality of publications; analyze the plurality of publications, using a filtering module, to determine a filtered set of publications; obtain the plurality of clinical trial entries; analyze a context of each of the publications in the filtered set and each of the plurality of clinical trial entries using a mapping module to identify a relationship of each of the publication with at least one of the plurality of clinical trial entries; and associate the plurality of clinical trial entries with publications related thereto to obtain documented clinical trials.

First claim

What is claimed is:

1. A system for documenting clinical trials, the system when operated identifies one or more publications related to a clinical trial entry to obtain a documented clinical trial, wherein the system comprises:
an information repository comprising a plurality of publications;
a clinical trials registry database comprising a plurality of clinical trial entries; and
a server arrangement communicably coupled to the information repository and the clinical trials registry database, wherein the server arrangement is configured to:
train machine learning algorithms using supervised learning using a labelled training dataset;
obtain the plurality of publications from the information repository;
analyze the plurality of publications to determine a filtered set of publications, wherein each of the publications in the filtered set is related to clinical trial research;
obtain the plurality of clinical trial entries from the clinical trials registry database;
analyze a context of each of the publications in the filtered set and each of the plurality of clinical trial entries using the trained machine learning algorithms to identify a relationship of each of the publications in the filtered set with at least one of the plurality of clinical trial entries;
identify associations between each of the publications in the filtered set and the at least one of the plurality of clinical trial entries based on the identified relationship, to obtain documented clinical trials, wherein identifying associations refers to establishing direct and indirect relationships between each of the publications in the filtered set and the at least one of the plurality of clinical trial entries;
determine, whether a given publication of the one or more publications related to the clinical trial entry is a primary publication or a secondary publication for the documented clinical trial based on the identified associations, wherein the primary publication comprises information that is exclusively related to the clinical trial entry and the secondary publication comprises information that is not exclusively related to the clinical trial entry; and
sort the one or more publications related to the clinical trial entry based on the determination of whether the given publication is a primary publication or a secondary publications.

2. The system of claim 1 , wherein the labelled training dataset for the server arrangement comprises labelled publications relating to clinical trial research as a positive training dataset and labelled publications unrelated to clinical trial research as a negative training dataset.

3. The system of claim 1 , wherein the labelled training dataset employed by the server arrangement comprises labelled clinical trial entries and labelled related publications as a positive training dataset, and labelled clinical trial entries and labelled unrelated publications as a negative training dataset.

4. The system of claim 1 , wherein the labelled training dataset for the server arrangement comprises labelled clinical trial entries, and associated related primary and secondary publications.

5. A method of documenting clinical trials, the method comprises identifying one or more publications related to a clinical trial entry to obtain a documented clinical trial, wherein the method is implemented via a system comprising a server arrangement coupled to an information repository and a clinical trials registry database, wherein the method comprises:
training machine learning algorithms using supervised learning using a labelled trainings dataset;
obtaining a plurality of publications from the information repository;
analyzing the plurality of publications to determine a filtered set of publications, wherein each of the publications in the filtered set is related to clinical trial research;
obtaining a plurality of clinical trial entries from a clinical trials registry database;
analyzing a context of each of the publications in the filtered set and each of the plurality of clinical trial entries using the trained machine learning algorithms to identify a relationship of each of the publication in the filtered set with at least one of the plurality of clinical trial entries;
identifying associations between each of the publications in the filtered set and the at least one of the plurality of clinical trial entries to obtain documented clinical trials, wherein identifying associations refer to establishing direct and indirect relationship between each of the publications in the filtered set and the at least one of the plurality of clinical trial entries;
determining, whether a given publication of the one or more publications related to the clinical trial entry is a primary publication or a secondary publication for the documented clinical trial based on the identified associations,
wherein the primary publication comprise information that is exclusively related to the clinical trial entry and secondary publication comprise information that is not exclusively related to the clinical trial entry; and
sorting the one or more publications related to the clinical trial entry based on the determination of whether the given publication is a primary publication or a secondary publication.

6. The method of claim 5 , wherein labelled training dataset for the server arrangement comprises labelled publications relating to clinical trial researches as a positive training dataset and labelled publications unrelated to clinical trial research as a negative training dataset.

7. The method of claim 5 , wherein labelled training dataset for the server arrangement comprises labelled clinical trial entries and labelled related publications as a positive training dataset, and labelled clinical trial entries and labelled unrelated publications as a negative training dataset.

8. The method of claim 5 , wherein the labelled training dataset for the server arrangement comprises labelled clinical trial entries, and associated related primary and secondary publications.

Assignees

• Innoplexus Ag

Inventors

• Pandita Esha
• Verma Vivek
• Patel Sunil
• Agarwal Vatsal

Classifications

• G16H70/00

  ICT specially adapted for the handling or processing of medical references · CPC title

• G16H10/20 Primary

  for electronic clinical trials or questionnaires · CPC title

• G06F16/313

  Selection or weighting of terms for indexing · CPC title

• G16H15/00

  ICT specially adapted for medical reports, e.g. generation or transmission thereof · CPC title

• G06N20/00

  Machine learning · CPC title