Artificial Intelligence (AI) in Adverse Drug Reaction (ADR) collection


As per the regulatory obligations, pharmaceutical and biotech companies need to establish a robust pharmacovigilance process to have a clear watch on the safety profile of medical products during their entire lifecycle to ensure patient safety. The requirement to review the safety of medicines starts from the clinical developmental phase till the product remains on the market (post-marketing). The companies need to observe on a constant basis must the risk-benefit ratio of the medicines and one of the most important steps in this process is collecting ADRs/Adverse Events (AE)/adverse experiences etc.

The patients, relatives and healthcare providers report the ADRs to companies via various channels like letters, e-mails, telephone calls to medical call centers (1-800 numbers), voicemail messages, telemedicine visits, chatbots, and social media, etc.

It has been observed that due to increasing awareness, AE reporting volumes to companies and local authorities have rapidly increased in recent years.

In 2019, the USFDA got notified with around 22 lac reports, a drastic increase in comparison with 5 lac reports reported in 2009 (1). The traditional approach used by the companies for recording and evaluating such numerous AEs in various formats and languages involves manual processing with many steps and is error-prone due to dependence on human resources.

Moreover, Covid-19 vaccine approvals and vaccination programs will cause a surge in ADR reporting and would be an additional strain on the already burdened system. To cope up with this high influx of diverse data, an increasing number of product approvals, and heightened regulatory requirements, the manual method needs scalability to make the process quick, reliable, accurate, and errorless.

The situation is driving the companies to leverage the AI technologies in pharmacovigilance to manage increasing workload and make life easy.

The term AI is not new in today’s era. But its use in PV is indeed accommodating. AI assists companies by automating information capture, yielding quick output and limiting errors. Technologies such as Natural Language Processing (NLP), speech to text conversion and Natural Language Understanding (NLU) are used to collect the ADRs with improved accuracy, speed, scalability and reduced costs.

To identify and collect the potential reports of ADRs / AEs, the companies must screen reported information. The reported information may be in the format of text (literature articles, electronic patient records, patient support program data), graphics (tables, images etc), voicemails and thousands of posts on social media platforms.

Automating data extraction with NLP can help to significantly reduce the burden associated with ADR capturing and further case processing. NLP can not only analyse the data but also extract the relevant information from both structured and unstructured data. NLP evaluates text-based forms as well as emotions (e.g., emoji) and companies can use NLP to read cases and develop a semantic understanding of the data in different languages too. Moreover, ADRs identified with NLP can be considered as a signal helping the companies in the proactive identification of early signals.

AI technologies such as NLP, NLU and machine learning can automate the intake of AEs from various sources and support case creation via multiple measures like redaction and annotation, forms extraction (using OCR technique), AI text mining, machine translation, and automated dictionary coding.

In today’s era, posting on social media platforms like YouTube, Facebook, LinkedIn and Twitter is the new pattern of communicating and sharing concerns with companies and reporting ADRs.

The burgeoning post volumes is making social media a much bigger problem for companies than finding water in the desert. An analysis of almost 10 lac Facebook and Twitter posts found one adverse event discussed on social media before being reported to the FDA Adverse Event Reporting System. (2) Creating a sensitive and specific NLP model to detect a few potential adverse events in a plethora of social media posts is the need of the hour.

Harnessing the power of AI, it takes hardly seconds to complete the ADR collection using semantic search and enrichment methods as compared with conventional ones. This will improve the ability to manage future surges in AE volumes and enhance the productivity of the human resources wherein they can be free to focus on value creation and problem-solving yielding better deliverables.

Due to increasing regulatory expectations, the surge in ADR reporting volumes and requirements of cost reduction initiatives, AI technologies like NLP can become a power booster for companies to screen and collect ADRs. Partnering with technology-enabled service providers will be a cost-effective and scalable model for pharmaceutical companies to efficiently handle the increased influx of data, improve regulatory compliance and enhance patient engagement.

References:

1. Center for Drug Evaluation & Research. FDA Adverse Event Reporting System (FAERS) Public Dashboard. https://www.fda.gov/drugs/questions-and-answers-fdas-adverse-event-reporting-system-faers/fda-adverse-event-reporting-system-faers-public-dashboard (2020).

2. Pierce, C. E. et al. Evaluation of Facebook and Twitter Monitoring to Detect Safety Signals for Medical Products: An Analysis of Recent FDA Safety Alerts. Drug Saf. 40, 317-331 (2017).

Linkedin


Disclaimer

Views expressed above are the author’s own.



END OF ARTICLE



.



Source link

%d bloggers like this: