System and method to produce and validate weighted relations between drug and adverse drug reactions

1. An apparatus to produce and validate weighted social media relations between drugs and adverse drug reactions (ADRs), the system comprising:

a processor; and

a memory to store instructions, which when executed by the processor cause the processor to perform a process to:

monitor information of a social media system for links between drugs and the ADRs;

extract a social media relation, among social media relations, between a drug among the drugs and an ADR among the ADRs using named entity recognition and to provide a weighted social media relation between the drug and the ADR which is weighted based on a confidence level of a link among the links between the drug and the ADR in the social media system;

store the extracted social media relation between the drug and the ADR in association with the weighted social media relation;

use domain knowledge in an ontology database to apply a refinement process to the named entity recognition for the drug and the ADR, respectively, to refine the weighted social media relation in accordance with one or more ontologies of drug names and/or of ADRs;

quantify the confidence level of the refined weighted social media relation based upon:

drugs and ADR links extracted from research publications and/or from clinical trial reports to quantify a research weight to be included in the refined weighted social media relation, based on a ratio of evidence linking the drug and the ADR to overall mentions of the drug in the research publications and/or the clinical trial reports, and

search results of an Internet search engine from a search of the World Wide Web over the Internet for the drug and the ADR, to obtain a number of search hits returned to quantify an internet weight to be included in the refined weighted social media relation based on the search results; and

compute the quantified confidence level of the refined weighted social media relation based upon information resulting from a user-defined strategy to give a weighting to each of the social media weight, the research weight and the internet weight, and aggregating the social media weight, the research weight and the internet weight to provide information indicating a validation of the quantified confidence level of the refined weighted social media relation.

2. The apparatus according to claim 1 , wherein the weighted social media relation between the drug and the ADR is in form of a triple data set of <drug, ADR, c> where c is the confidence level corresponding to a weight.

3. The apparatus according to claim 1 , wherein to refine the weighted social media relation includes expansion of the weighted social media relation to include equivalent drug names and equivalent ADRs and/or refinement of the weighted social media relation to replace the ADR with a more or less specific ADR.

4. The apparatus according to claim 1 , wherein the confidence level of the weighted social media relation is above a threshold confidence level.

5. The apparatus according to claim 1 , wherein the internet weight is based on a search engine distance between the drug and the ADR.

6. The apparatus according to claim 1 , wherein the process is to,

monitor the social media system for other links between the drugs and other substances;

extract a drug-substance relation between a drug among the drugs and a substance among the other substances using named entity recognition and to provide a weighted social media drug-substance relation between the drug and the substance which is based on a confidence level of a link among the links between the drug and the substance in the social media system;

store the extracted drug-substance relation between the drug and the substance in association with the weighted social media drug-substance relation;

use the ontology database to refine the weighted social media drug-substance relation in accordance with one or more ontologies of drug names and/or of substances; and

quantify the confidence level of the refined weighted social media drug-substance relation based upon,

substance and drug data extracted from research publications and/or from clinical trial reports to quantify a research weight for the weighted social media drug-substance relation, and/or

search results of an internet search engine searching the World Wide Web over the Internet for the drug and the ADR, to obtain a number of search hits returned quantifying an internet weight for the weighted social media drug-substance relation based on the searching.

7. A system to allow a user to assess relations between drugs and adverse drug reactions (ADRs), the system comprising:

an apparatus including a processor and a memory storing instructions, which when executed by the processor cause the processor to perform a process to produce and validate weighted social media relations between the drugs and the ADRs in a knowledge base according to claim 1 ; and

a user interface allowing input of a user query for a social media relation between the drug and the ADR,

wherein the processor is to,

rewrite the query using a domain ontology; and

retrieve an answer to the rewritten query from the knowledge base.

8. The system according to claim 7 , wherein if no relation is found in the knowledge base, the processor is to further monitor the information from the social media system in real time.

9. A method of producing and validating weighted social media relations between drugs and adverse drug reactions (ADRs), the method comprising:

by at least one processor,

monitoring information of a social media system for links between drugs and the ADRs;

extracting a social media relation, among social media relations, between a drug among the drugs and an ADR among the ADRs using named entity recognition and to provide a weighted social media relation between the drug and the ADR which is weighted based on a confidence level of a link among the links between the drug and the ADR in the social media system;

storing the extracted social media relation between the drug and the ADR in association with the weighted social media relation;

using domain knowledge in an ontology database to apply a refinement process to the named entity recognition for the drug and the ADR, respectively, to refine the weighted social media relation in accordance with one or more ontologies of drug names and/or of ADRs;

quantifying the confidence level of the refined weighted social media relation based upon:

drugs and ADR links extracted from research publications and/or from clinical trial reports to quantify a research weight to be included in the refined weighted social media relation, based on a ratio of evidence linking the drug and the ADR to overall mentions of the drug in the research publications and/or the clinical trial reports, and

search results of an Internet search engine from a search of the World Wide Web over the Internet for the drug and the ADR, to obtain a number of search hits returned to quantify an internet weight to be included in the refined weighted social media relation based on the search results; and

providing information indicating a validation of the quantified confidence level of the refined weighted social media relation resulting from a computation of the quantified confidence level of the refined weighted social media relation based upon information resulting from a user-defined strategy to give a weighting to each of the social media weight, the research weight and the internet weight, and aggregating the social media weight, the research weight and the internet weight.

10. A method of allowing a user to query for a link between a drug and an adverse drug reaction (ADR), the method comprising:

by at least one processor,

allowing input of a user query for a social media relation between the drug and the ADR stored in a knowledge base according to claim 9 ;

rewriting the query using a domain ontology; and

retrieving a query answer for the rewritten query from the knowledge base.