Skip to main content

Artificial Intelligence Seeing through the lens of Pharmacovigilance

Pharmacovigilance as we know is a science with a set of pre defined functions to collect, analyse, monitor adverse event reports in understanding the safety profile of drug.
The set pre defined functions would include case processing through data entry of adverse event forms into safety database, medical review, aggregate reporting, signal detection, risk evaluation and mitigation strategies.


With patients awareness and regulatory compliance we may have seen a surge of adverse event data over last few years , resulting in the urgent need for the application of automation. Pharmacovigilance is the only discipline where in which timelines and quality data are evaluated on a benchmark of 100 % and a compromise in these two parameters are considered to be a zero tolerance.
Automation of above pre defined function is possible through machine learning, which is an integral components of Artificial Intelligence.

What is Artificial Intelligence ?       
Artificial intelligence may be called as an ability of a computer system to perform task that require human intelligence such as cognition through visual acuity, voice recognition, language translation leading to decision execution of a certain function.
Machine learning is based on reinforced data, where in which when an algorithm is executed to accomplish a specific task.

If it accomplishes the algorithm ends and the entire procedure in auto stored in the program, which means next time one does not need to manually execute the program, it would be auto executed in order to accomplish the task, if presented with the exact same variables as that of the earlier scenario.
In the second case if the task is not accomplished then too the procedure would be stored in the program and next time when the program is auto executed it would not take the same path thus minimizing error
This process self-learning through experience is called machine learning.

For example imagine a scenario where in which you have you have received an email from a patient who has experienced nausea, followed by headache and bleeding from nose on Lisinopril, the patient also mentions that he has a history renal impairment and also that he was a chain smoker for which he took Varenicline to quit smoking

An algorithm created on the principle of machine learning would have the capability to auto recognises and identify the suspect drug from concomitant therapy, adverse event from medical history and not only this through robotic process automation it may integrate the email with safety database, which means not only it identifies the suspect drug and the adverse event, it now also does the auto data entry, prepares auto case narrative and auto sends letters to the patient or physician for further follow up  from the safety database.

This is ‘Artificial Intelligence’, a capability attained through self-learning to process thousands of data within seconds.


With automation employees engaged in manual data entry would be upskilled in the execution of AI process.

Comments

  1. This comment has been removed by a blog administrator.

    ReplyDelete

Post a Comment

Popular posts from this blog

Eritrea ranked top by UMC- WHO in ADR reporting

Eritrea an African country located in the north east region on the red coast sea bordering Ethiopia and Sudan, has been reported as the top country in reporting adverse drug reaction through safety reports submitted to global database by the Uppsala Monitoring Centre (Sweden), the collaborating centre for WHO as part of their international drug monitoring programme.
The Pharmacovigilance programme of Eritrea was carried out at maturity LEVEL THREE as per UMC with the assistance of local strength engaged in providing harmonization. Eritrea has identified 11 safety concerns on their medicinal products, which were also published in international peer reviewed journals. The Eritrean work on Pharmacovigilance has been globally acknowledged with insights on revising the policy of Eritrean Healthcare system.
Recognizing the significant contribution made in pharmacovigilance,Eritrea along with Netherlands and Peru has been invited for the upcoming 40th Anniversary of UMC on May 18 to share the…

Chat Bot- the first artificial intelligence bot in Pharmacovigilance

Merck Sharp & Dohme (MSD) Corp., a subsidiary of Merck & Co., Inc has developed an Artificial Intelligence (AI) chat bot. The MSD Salute chat bot is designed to aid physicians in providing product information and pathology. 
The Physicians who are registered with MSD would use the Chat bot using facebook messenger.  Chat bot is derived from the social media and uses machine learning and implements feedback to develop interactions. 
Currently MSD is monitoring and not filtering the communication between Chat bot and users, as they want to analyse the open ended exchanges between the AI and the users. Chat bot currently in the testing phase, is said augment patient safety through proactive pharmacovigilance
We at Think I are also driven to produce Pharmacovigilance safety database solutions using cloud based technology which are not only cost effective and user friendly but also fully validated for quick, easy import, entry and electronic submissions of adverse event. 
For mor…

What is a ‘backward forward convertor’ ?

Did you know that ICH has a backward forward conversion (BFC) tool which was build using eXtensible Stylesheet Language Transformation (XSLTs) ? This language basically transforms a XML file of a given format into another XML file of a different format. The BFC tool requires a program that understands XSLT language to make conversion of E2B (R2) to E2B (R3) and vice versa. Now considering this, the European medical agency has improvised the BFC tool with additional fields as per the EU implementation guide. The EU version of the amended BFC tool has been published on the EMA website.
The key files in the BFC package includes downgrade-icsr.xsl; which downgrades E2B (R3) file into E2B (R2) and upgrade-icsr.xsl; which upgrades E2B (R2) file into an E2B (R3) file. The acknowledgement convertors include downgrade-ack.xsl and upgrade-ack.xsl What are the technical limitations of BFC tools?
However, the conversion using BFC has some major limitations.
When converting ICSR from E2B (R2) i…